3 Step Guide to Overclock Your i7 / i5 Haswell Platform

Well, we skipped a generation on the front page, because overclocking Sandy Bridge and Ivy Bridge were both relatively similar, and we had this excellent thread on the forum by theocnoob. The overclocking game, however, is in for a bit of a facelift with Haswell.

What if I were to tell you all this super high speed RAM people purchased for Ivy was no longer necessary and/or it can’t be used with Haswell? What if I were to tell you there were people in the industry that think Haswell is going to be so difficult to cool that 24/7 single stage phase change cooling will be more appealing again? Think you could use a guide then? We thought so.

First up though, I have a confession to make. The “3 step” part of the title is not the whole story. There are going to be three steps, but it’ll be more like three categories, with sub-steps.

Step 1: CPU Clock is King

One thing that hasn’t changed with Haswell is that CPU clock is King. Regardless of what you do with the rest of your system, be it RAM clock, BCLK or any other clock, CPU clock is going to be the number one most important thing you can improve; thus, we start there. Before we get there, go into your UEFI, load optimized defaults, and to avoid any potential instability, set your RAM’s voltage to whatever it’s rated at (1.5 V or 1.65 V typically), leaving everything else on “Auto”.

Set your RAM Speed

Set your RAM Speed

Set Your RAM's Voltage

Set Your RAM’s Voltage

The first thing you need to know about Haswell you already found out in our review (read it here if you haven’t already) – it runs hot. You will need a lot more than the stock cooler if you want to overclock. With a very good air cooler or all-in-one water cooler, you’re looking at a heat limited voltage cap of about 1.25 V. At that voltage with air or AIO cooling, you’ll be seeing temperatures in the upper 80′s to lower 90′s (°C) range under normal full processor load.

The easiest way to see how far your particular chip will go is to set the voltage manually at 1.25 V, set all cores to run the same multiplier, then start raising that multiplier until you run into instability.

Set Your CPU Voltage to 1.25 V

Set Your CPU Voltage to 1.25 V

Set Your CPU Multi to 43X or Higher

Set Your CPU Multi to 43X or Higher

A good stress test that stresses more than one instruction set (such as Prime95) is the stability test in AIDA 64 (download that here). However, you should know that will make your CPU run hotter than any normal load you can put on it. It is designed to throw every instruction set at the CPU at once to make sure it can handle the load. In every other scenario, even 24/7 loading programs like Folding@Home or SETI@Home won’t heat it up quite that far; though those will be closer to the AIDA64 temperatures then normal usage scenario like gaming or workstation use.

Run Your Stability Test

Run Your Stability Test

Ideally, you will be able to keep normal usage temperatures under 90° C, just to be safe. Going over that won’t kill the chip, but it won’t do it any favors either. Once over 90~95° C, the chip will begin throttling to reduce temperatures and that will make your overclock pretty useless rather quickly.

Now, I’ll share a secret imparted by the folks at ASUS who gave several reviewers some tips on overclocking the retail stepping Haswell chips: Set Vcore to 1.20 V. Set all cores to 46x (which would be a 4.6 GHz overclock), save & reboot. If the system boots past the UEFI and either begins to load or, ideally, makes it into the OS and is stable, you have a 50th percentile or greater chip on the Haswell overclocking-ability bell curve. If it won’t at least boot there and make it into the UEFI, you probably have less than a 50th percentile chip. You can expect chips in the lower 50th percentile to top out in the 4.4-4.5 GHz range at 1.25 V.

If your chip will boot at 4.6 GHz and 1.25 V, that’s very good. It means you have at least an average chip. If it will boot at 4.6 GHz and is stable there, then you may have an above average chip. The best chips will be able to do 4.8 GHz stable at 1.25 V. Our sample did 4.8 GHz, but at 1.3 V and on a custom water loop. Using 1.3 V will likely put a chip out of the air cooling / AIO water cooling thermal envelope. Temperatures in all of these scenarios, from the dog 4.3 GHz chips up to the good 4.8 GHz chips, will always be in the ~90°C range. That’s just the nature of Haswell. With the VRM on-die, think of Haswell as Ivy Bridge plus 10° C.

Regrettably, according to overclocking legend Andre Yang, there is no way to bin Haswell for extreme overclocking purposes on ambient cooling. You couldn’t do it with Ivy Bridge either. That ship sailed when Sandy Bridge, well, went under the bridge and continued down the river. If you’re binning for extreme purposes (of which the esteemed Mr. Yang has binned over 500 chips as of late April), you simply have to take them cold. However, for most people’s uses, the method above to bin on air is tried and true through the hundred plus they have binned at ASUS headquarters.

So, to sum up,

  • First check to see whether you have a chip on the upper end or the lower end of the Haswell bell curve.
  • There will be a large variation as you can see, from the 4.3 GHz dog up to the ideal 4.8 GHz chip, both at the same voltage and temperature.
  • Once you’ve checked on your chip, you have a general goal and can then set 1.25 V – manually, so it doesn’t overshoot like it would on Progressive voltage.
  • After setting 1.25 V, then raise the multiplier.
  • Test for stability.
  • Raise the multiplier.
  • Test for stability.
  • Etc, etc until you find your chip’s maximum ambient clock.

A Note on Temperatures

Like all chips, you want to try to keep your chips as cool as you can to increase their longevity. Personally, I’d try to keep a Haswell chip below 85 °C for 24/7 operation (definitely below 90°C); not because they can’t take more (they can), but because it’s usually safer to err on the side of caution. Multiple people we talked to as ASUS actually said they think single stage cooling is going to make a comeback (anyone remember the Lian Li cases with SS coolers built-in?). Yes, these run hot; but if you control them, they’ll do well for you.

Speaking of temperatures, surely people will be saying “Haswell has the same problem as Ivy Bridge,” referring to the thermal paste issue. Yes, Haswell has thermal paste, but from Very Authoritative People, the TIM is not the problem. As has been posited across the net when people de-lid Ivy Bridge chips with great results, it’s really the black adhesive that’s the culprit. When you cut out that adhesive, it allows the IHS to sit closer to the CPU die, meaning there is less thermal paste through which the heat has to travel, leading to significantly lower temperatures. Intel’s TIM is really quite good, but the manufacturing process leads to that glue being just a little too thick, which is why you see such temperatures.

Even if you de-lid your Haswell CPU, don’t expect a 60°C chip all of a sudden. These are produced on a very small process, with billions of transistors in an extremely tiny area. Even if you have a perfect die-to-IHS interface, the chip is still going to run pretty hot because there is so little surface area off of which to pull the heat. I guess what I’m saying, to put it succinctly (too late!), is: Don’t worry about temps, within reason. If your chip runs warm, that’s ok. They’re just going to do that. You can’t draw heat away as fast as you used to be able to when you’re dealing with such a small surface area with such a large transistor count. Keep them as cool as you can and be happy.

Step 2: Dial in Your RAM

Now that you have explored your chip for roughly its highest stable overclock, you can move on to other parts of the system. We’ll start by telling you one very important thing: RAM is a different animal than it was on Ivy Bridge. With Ivy, you could get RAM that clocked to the moon. Assuming your chip had a decent IMC (integrated memory controller), you could anticipate an easy DDR3-2133 to DDR3-2400 RAM clock. Good chips would get to DDR3-2800 and even beyond.

Dial In Your RAM

Dial In Your RAM

With Haswell, that has all changed. Memory is a strange beast with Haswell. With your CPU clocked at stock, the Haswell IMC is a beast. Some chips will be able to do DDR3-3000+ without so much as batting an eye. Others not so much, but DDR3-2400 isn’t a stretch at all.

But – and this is a big, bold, italicized, very notable but - overclocking a Haswell CPU at the same time as overclocking the RAM will reduce your IMC’s ability to overclock or reduce your CPU’s ability to overclock. You have to choose one or the other, which is why the CPU came first in this guide. Core speed is king; remember that. If it comes down to choosing which to push farther, the CPU should always win in your calculations.

This will not only be heavily depending on the particular CPU you get, but also on the motherboard you choose. For instance, the Intel motherboard used for the Haswell review couldn’t run greater than DDR3-2133 at 4.8 GHz. However, using the same chip and the same memory, the ASUS Maximus VI Extreme for a future review had no problem running DDR3-2600 at 4.8 GHz.

Speaking of ASUS, let’s start with an example they gave. You might be able to run DDR3-2800 RAM and overclock your i7 4770K to 4.4 GHz. However, drop that down to DDR3-1600 or DDR3-1866 and that same CPU at the same voltage might be able to get to 4.6 GHz now. Everything in Haswell is linked. Reducing RAM speed on poor-to-pretty good chips will almost invariably increase the CPU’s ability to overclock.

Thus, you know all these high speed kits that were selling well with Ivy Bridge? Get ready to see how tight they can tighten their timings at lower speeds. Timings are back folks. No longer can IC manufacturers count on increasing memory clock, timings be damned. If you can get your hands on an older PSC or BBSE based kit with tighter timings (the G.Skill Flare DDR3-2000, 7-9-7-24 kit of yesteryear comes to mind. *Cough* someone killed mine…*cough* IMOG *cough*….), it would behoove you to do so.

Now that you know Haswell’s dirty little secret, you can start to work on your RAM. Memory enthusiasts will want to just see what their chip’s IMC is capable of, or, if you have a high speed memory kit, at least see that your IMC can run your RAM’s rated speed. So that’s where we’ll start. Set your CPU multiplier and voltage back to Auto and then manually set your RAM’s speed and timings. You should already have set its voltage earlier, but if you didn’t, do that now as well.

One tip I will share from my testing – Setting your System Agent (VCCSA) from 1.15 to 1.25 V will help out tremendously. Using those settings I’ve successfully been stable up to DDR3-2933 so far and I’m sure it will climb. So when clocking memory, set those voltages in addition to your RAM’s voltage.

Notice RAM Speed on this DDR3-2600 Kit is Down to DDR3-2133, but With Tighter Timings

Notice RAM Speed on this DDR3-2600 Kit is Down to DDR3-2133, but With Tighter Timings

Boot up into your OS and then send it through a run of HyperPi (download it here) using all available threads and the 32M setting. Warning: This will slow your system – any system – to a crawl while it completes. It’s the best memory stability test I know of other than Memtest or hardware testing. The test should take eight to ten minutes to complete.

Now that you (hopefully) know your memory will run its rated speed and timings, it’s time to push on the CPU again. Set it to the 1.25 V you used earlier and your already known-good multiplier. If you are running DDR3-2000 (or less) memory, chances are you’ll set it and be good to go. If you’re running DDR3-2133 or greater, you are about to find out how good your chip is when the CPU speed is combined with higher speed memory.

If the system fails to boot or is unstable when you go into the OS and test for stability, you are on the wrong end of the bell curve with regard to memory overclocking. Don’t be disheartened though, as long as your CPU is a decent overclocker, CPU clock speed is always king and you will just need to reduce your memory speed. So go back into UEFI and turn your memory to the divider immediately below the one you’re on. Go back into the OS and test for stability. Rinse and repeat until you have a stable system with your known-good CPU clock and your highest stable RAM divider.

If you have had to reduce your RAM speed at this point, don’t be discouraged. With some RAM kits – not all, but many of them – you can now get back some of that lost capability by exchanging reduced speed for tightened timings. With ICs currently on the market, you would do well to reduce timings in the following order, going into the OS and running HyperPi 32M to test stability in between each change.

* For our purposes, we’ll only deal with the first four memory timings (CL – tRCD – tRP – tRAS, i.e. 9-11-9-24) and the command rate.

Reduce Timings in This Order

  • Command rate – Go from 2T to 1T (or 2N to 1N, depending on what your board calls them). This is the easiest change to make and most kits will have no problem with this change.
  • tRP – Try reducing by one or two. This timing is relatively forgiving, but is a little more difficult than tRAS in the next step.
  • tRAS – Try reducing by two to five. Of the four primary timings, this one is the easiest to reduce.
  • tRCD – Try reducing by just one. It gets harder with this timing. You might not be able to reduce it at all.
  • CL – This is the trickiest and is more difficult to tighten than tRCD. Try just one here as well, but expect you won’t be able to reduce it, depending on how far you had to reduce your memory speed. If you can reduce it, this will have the biggest impact on benchmark results.

Once you have your lowest stable timings at your new memory clock, you’re done with memory for now. Not forever as you’ll see shortly, but for most people, this is where you can stop. You have successfully set very close to your most optimized CPU and memory frequencies (and timings for the latter) for your new Haswell build. Congratulations, you’re overclocked!

Of course, we’re overclockers. You knew we wouldn’t just stop there. Now we get to the extra tweaking part. The part that will do very little for you in terms of practical use, but will let you fine tune your system however you want.

Step 3: Fine Tuning With BCLK Strap and BCLK

BCLK has been an interesting beast since Sandy Bridge came out. The days of 100-200 MHz+ bus clock adjustments have been gone, at least in a totally linear fashion. Sandy Bridge and Ivy Bridge were both 100 MHz +/- 5-10%, with some rare samples getting in the ~15% range, but those were definitely the exception.

Sandy Bridge-E introduced a new option for those that enjoy tweaking their systems – BCLK strap, where you could use a 100, 125 or 166 MHz BCLK strap while keeping everything else (i.e. PCIe busses, storage lanes, etc) in spec. That gave you the ability to choose where your BCLK’s starting point was, then adjust the same 5-10% up or down from there.

All was not perfect with Sandy Bridge-E though, in that not all chips could even get to the second BCLK strap (they go 100/125/166 MHz) reliably. While the ones that could do 125 MHz strap weren’t too rare, the ones that made it to the 166 MHz strap were indeed the exception to the rule.

Intel in its infinite wisdom has now granted the mainstream users with the BCLK strap option. Like Sandy Bridge-E, Haswell will have the same three BCLK strap options.

Haswell BCLK Straps

  • 100 MHz (default)
  • 125 MHz
  • 166 MHz
  • 250 MHz (exists on some boards but is not generally usable)
BCLK Strap Options

BCLK Strap Options

Through some “magic” (their word), ASUS has actually gotten a 200 MHz BCLK strap option into their Maximus VI Extreme UEFI and it is indeed usable on chips capable of reaching it. Obviously that is a 20%+ increase over the 166 MHz strap, which is pretty sweet. They would not tell us what they did to get there, simply referring to it as their magic. For obvious reasons (they don’t want everyone doing what they’re doing), they’re reluctant to share how they managed that publicly.

ASUS' Secret Sauce

ASUS’ Secret Sauce

Anyway, the great news about Haswell is that – unlike their SNB-E cousins - most, if not all, CPUs can reach the 166 MHz strap without issue. That gives you a lot more flexibility to tune your RAM and CPU to achieve their absolute maximum stable frequencies. There are far too many variables to go into every single scenario of course, but a couple examples should suffice.

BCLK Strap CPU Multi CPU Freq

RAM Speed
Selection

RAM Freq
 100 43x 4300 MHz DDR3-2133 1066 MHz
125 35x 4375 MHz DDR3- MHz
167 26x 4342 MHz DDR3- MHz

As you can see, this gives you a significant amount of room to toy with. In addition to simply changing the BCLK strap, remember you also have 5-10% (sometimes more) worth of BCLK wiggle room. Say your chip isn’t quite a 4.7 GHz chip, but will do just a hair under it with acceptable temperatures and stability. NO problem! You can shoot for 4680 MHz from any one of those straps.

BCLK Strap BCLK Offset New BCLK CPU Multi CPU Freq
 100 +1.7 101.7 46x 4678 MHz
125 +1.5 126.5 37x 4681 MHz
167 -5.6 161.4 29x 4681 MHz

Notice in the examples, the last one went down instead of up. BCLK is + OR – 5-10%. Say your RAM can’t handle one of the dividers at that particular BCLK but you want to use that particular strap, simply raise your CPU multiplier by one and lower BCLK as necessary. There are myriad possibilities with this sort of configuration. With the new BCLK strap option, you can dial in your system exactly where you want. It’s not FSB clocking of old, but it’s a lot better than the locked-down 100 MHz +/- 5-10% of Sandy Bridge and Ivy Bridge.

When you consider what FSB overclocking of yesteryear was like, this can be considered an improvement. You can select any RAM you want without worrying about clocking it too far (you could never go under a 1:1 ratio and RAM became a very limiting factor on, say, an E8400 at 550+ MHz. You also don’t have to worry about extra cooling needs of pushing a northbridge very far (I had one that burned an impression into a water block).

Now those that enjoy tweaking every single little thing and getting every MHz they can out of both their CPU and RAM have that capability. Speaking of every single MHz…

Tips and Tricks To Get That Extra MHz

Some of these tips will repeat items mentioned earlier, others will be new but weren’t all-encompassing like the rest of the guide and will be very chip-specific.

The number one most helpful tip to make sure you’re getting the most out of your CPU is to reduce the memory speed and try overclocking again. If you followed the guide, you shouldn’t have set your memory to a fast speed to begin with, but some people – myself included – for years have always gone into BIOS and the first thing they did was set memory speed, timings and voltage. It’s so familiar, the action is darn near muscle memory by now. If you did that and have high speed memory (> DDR3-1600 or DDR3-1866), try reducing memory speed and see if that helps your CPU get any further. If you just set the voltage and left everything else alone, you can skip this step.

Next up a short list of quick tips & tricks to get that last MHz out of your CPU.

  • If you manually set it, always keep the Cache/ Ring Bus / Uncore Frequency within 100-300 MHz of your CPU Core speed.
  • There are two new PLL Options – SBPLL and LCPLL.
    • SBPLL is useful for higher BCLK
    • LCPLL is useful for lower BCLK
  • Pay attention to both the package temperature and the core temperatures. The integrated voltage controller has a big influence on temperatures.
  • Always keep temperatures as low as you can.
  • For clocking RAM, good base voltage settings in my experience are:
    • RAM voltage at its rating (usually 1.65 V for high performance RAM)
    • System Agent (VCCSA) voltage: 1.15 V to 1.25 V

This list is not totally exhaustive. We’re happy to revise new tips and tricks in here, so if you find out some great items to overclock your CPU, don’t hesitate to share them!

Stability Testing

How to test for stability has been a very long and drawn out debate. Some people use Prime95, others IntelBurnTest, still others OCCT. I’m here to tell you whatever you want to do is fine, as long as you know their pluses and minuses. For instance, Prime95 only tests the one instruction set on your CPU but not all at once (SSE, AVX, FPU, etc). It’s good for heating up your CPU, absolutely, but it doesn’t necessarily throw the kitchen sink at it.

From the first 3770K and even at present, ASUS advocates using AIDA64′s stability test. The software is free (shareware, with just a couple features blocked; licenses are cheap if you find you need everything) and easy to use. The reason they advocate AIDA64 – with which I agree – is because it tests all instructions at once. It throws everything plus the kitchen sink at your CPU. It’s fast, stable and will also heat up your CPU. Thus, here is my tools list for testing stability.

Alternatives to AIDA64

Super important note: These programs will increase temperatures further than any other program you’ll ever use in day to day PC operations, even Folding@Home. So be sure to watch your temperatures. If you are seeing frequency and voltage throttling because your CPU is north of 95°C (or worse, north of 100°C), you need to dial it back. Even if your chip can clock that far, it’s not good to operate at the temperature of boiling water for prolonged periods.

How long you run for stability testing is totally up to you. After doing this for years, I tend to take the easy route – 10 minutes to an hour of stability testing, then cease that nonsense and use the computer for a week or so. If you want to immediately jump into mission critical work (i.e. writing word documents / editing spreadsheets you can’t lose, programming, etc…anything that you don’t want to potentially crash on you), you might want to do it longer than I. Rational people I know suggest 24 hours of stability testing before they’ll deem their overclocks stable. There are good arguments for doing it that long. Just watch your temperatures and make sure you’re not throttling.

Letting Your CPU Breathe

Now that you have your CPU dialed in, let’s talk about the way you apply voltage to it. So far we’ve only spoken about manually setting voltage and for the purposes of finding your maximum stable overclock, that’s fine and dandy. However, when your CPU isn’t under load, you don’t necessarily want it running at its max stable overclock with a constant voltage applied. So while we’ve covered manual voltage settings, there are two other options that might serve you better.

Offset Voltages

Offset voltages do just what you think it does – offsets the voltage +/- the amount you set. It takes the standard Intel voltage curve and just raises the whole darn thing, top to bottom. While this is better on your CPU than just a manually applied, constant voltage -and it used to be the best option for not running manual voltage- it’s not the best option any more.

Adaptive Voltage

Adaptive voltage is a lot smarter than your average offset voltage. What it does is raise the part of the voltage curve that you need under load, but leaves the lower end of the curve (when the CPU is operating at 800 MHz) alone. Thus, the higher voltage you need is there when the CPU calls for it, but you have no increase in idle voltage at all. This is going to be the best choice for letting your CPU breathe, as it were, when idle. You set the maximum voltage you want the CPU to get and go on your merry way knowing the CPU can draw the voltage it needs when it needs it.

A picture is worth a thousand words, and this slide outlines the voltage override modes quite well.

Various Integrated Voltage Regulator Configurations

Various Integrated Voltage Regulator Configurations – Image Courtesy HWBot

Caveat – big, massive, honking, pay attention to this – caveat! When using adaptive voltage, the top of the curve isn’t necessarily the whole story. ASUS drove this point home when we met and I’m doing so now; we all need to do this as a public service to our users and readers. Even though you set 1.25 V as your maximum voltage, under certain very heavy loading conditions (i.e. stress testing), the voltage can and will exceed the maximum you have set. Let me say that again – if you stress test using adaptive voltage, even with the maximum set to 1.25 V, the CPU will – guaranteed – request and receive more voltage than you have told the motherboard to deliver it. It’s just the way this works.

In the ASUS demonstrations, setting a maximum 1.25 V in this scenario and then running something like AIDA64, Prime95, etc – any stress testing application designed to stress more than normal loads – the CPU requests and gets ~1.36 V. There is nothing you can do about this and it will happen whether you want it to or not. The only way to prevent stress testing programs from pulling extra voltage is to use a manually set voltage, which takes away your CPU’s ability to reduce voltage when idle.

All is not lost though! You probably noticed every time I told you what would happen it went hand in hand with stress testing. That’s because those are the only scenarios that will lead to this behavior. In all other circumstances so far as ASUS can tell, the CPU will cap at the set 1.25 V and never exceed it. Even if you’re Folding@Home it should maintain the 1.25 V cap (though those folding at home are under 100% load all the time and I’d suggest using a manual voltage for those machines just in case). Likewise, video encoding, audio encoding, compression, rendering – any CPU-intensive process, your CPU will maintain the 1.25 V cap.

So, now that you know the very important caveat, set 1.25 V as your maximum voltage with adaptive voltage, enable C1E and EIST (which throttle your multiplier and voltage when not under load) and go on with your merry life knowing your CPU is as relaxed as it can get when idle.

Small Note on Extreme Overclocking

I haven’t really had time to freeze one of these chips yet, but thanks to being able to speak with Andre Yang, we have at least two important notes to share. First off: CB & CBB (cold bug & cold boot bug) are back. In his testing, you’re looking at a CB/CBB of ~ -120 to -130 °C for the worse chips (think the 4.3 GHz dogs) and ~ -130 to -140 °C for the better chips. Thus, if you don’t have a well controlled pot with plenty of mass, it’s time to look for one.

Second tip – give them all the voltage you can with (relative) impunity. That 7 GHz validation you saw floating around before launch with 2.5 V+ applied to the CPU? That’s no bug folks. Earlier ES chips could indeed have 2.5 V applied. There is a lower cap for retail chips; you can’t take them to extra-insane levels like that, but you should still be able to get in the ~1.9 V range similar to Ivy Bridge chips today. Oh, and the 2.5 V+ chip? I’m told it’s still alive and happy. That could very well be the exception to the rule, so don’t think you can apply near two volts to a chip and have them necessarily live to tell the tale.

Lastly, what are you shooting for? After ~500 chips, Andre’s best chip was around 7.3 GHz. It’s OK if you get a less-than-perfect chip though, because a 6.5-6.7 GHz Haswell chip is roughly equivalent to a 6.9-7.1 GHz Ivy Bridge chip.

Until we get to play with these under sub-zero conditions, that’s all we’ve got for the extreme crowd. That’s OK though, because this is a guide geared toward less extreme overclockers anyway.

Final Thoughts

If you’ve made it to this point, congratulations, you’ve done it! Your new Haswell platform is overclocked and tuned to its best, most efficient overclock. If you followed that last section and use adaptive voltage, it is even taking advantage of the power sipping idle states.

All that’s left now is to show everyone else what you’ve done! Comment on this thread with a screenshot of your overclock and fill out one of the forms in the first two posts (for the i7 4770K and i5 4570K, respectively). We’d love to know what you were able to do with your CPU. With your help, we can plot our own results and see just what a crowd-sourced Haswell CPU bell curve looks like.

The only request is that you post stable overclocks. If you want to post your “ZOMG LOOK WHAT FREQUENCY I VALIDATED!” result, we’d love to see that too, but over in the benchmarking section. For the purposes of our – and I use this term loosely – research, let us know what overclock you achieved with stability.

I’ll reserve the first post for you to post your results. Please submit via the embedded Google form and it will automatically add your result to the database and over time it will allow us to graph the results so we can see what sort of average there is as well as whether there is any correlation with batch numbers. Please feel free to also submit a screenshot of your overclock so we can make sure some joker isn’t just filling out the form with bad data.

So, when you post, please share:

  • 24/7 stable CPU overclock (in MHz, i.e. 4625 MHz)
  • Voltage used
  • CPU Batch # (Located on the package)
  • Motherboard
  • Forum username (important for us to easily catch bogus submissions)

Optional, but welcome information:

  • Stress testing temperatures (in °C)
  • CPU Cooler
  • Did you use/like adaptive voltage? Why / why not?
  • RAM clock (in DDR3-xxxx)
  • RAM timings

Happy clocking!

Jeremy Vaughan (hokiealumnus)

Tags: , , , , , , , , , , ,

283 Comments:

'Cuda340's Avatar
Thanks for the guide (And tips) hokie

Much appreciated !!
PolRoger's Avatar
Thanks Hokie!

Alright... some fresh reading material.

Oih Boi... "Thats just the nature of Haswell. With the VRM on-die, think of Haswell as Ivy Bridge plus 10 C."... bummer.
Janus67's Avatar
Aaaaand bookmarked

Nice guide, can't wait to put it to use!
ivanlabrie's Avatar
Really interesting...great guide man!
Can't wait to sell my IB stuff and get Haswell
Boulard83's Avatar
Thanks for the guide !
Itsumo's Avatar
Thanks for the guide, fascinating read!
Reefa_Madness's Avatar
@Hokie, I found that observation of yours about the inverse relationship between CPU clocks and ram clocks to be specially interesting. In that case, all of the really impressive Haswell ram clocks being posted about are pretty much for show, because if you want overall higher performance (like you stated, CPU clocks are what counts), then you aren't likely to be running one of those 2933 kits at their rated clock speed. Granted, there is nothing wrong with showing off the ram at 3000+, but from a practical standpoint that makes absolutely no sense if you have to dial down the CPU to accomplish it. One this is dropping it down a notch if you're right at the edge, but this appears to be way more than that.

Have you had the opportunity to discuss this with Asus or anyone else to confirm that this has been observed by others, or is it possible that it is peculiar to your sample?
ivanlabrie's Avatar
Well, the pifast record is at 6700mhz and uses bbse over 3000mhz with cl5-9-7-18-1t...so I'm not too sure about that.
Janus67's Avatar
The ram that Ney uses is incredible though,
Reefa_Madness's Avatar
@Janus67...sir, in this case, your glowing praise of the ram is simply not sufficient. Dang, those are some serious sticks of ram!

@ivanlabrie, that may very well be true, but Hokie's testing was done at ambient temps so comparisons should probably be as well in order for them to be relevant. You now that what happens at sub-ambient temps doesn't always hold true at ambient temps, much less 24/7 usage.
hokiealumnus's Avatar
Excellent question Reefa. My RAM advice actually come a straight from ASUS. I used it to my advantage when clocking my chip but that is expected to be the new norm. Obviously YMMV with IMC and CPU clocks, and going cold as you've seen on the not throws convention out the window. The guide is for ambient clocking mostly though.
PolRoger's Avatar
@Reefa:

I'm going to be disappointed if Haswell's IMC doesn't run memory overclocks at half bank or even full bank "stable" with higher memory speeds than IB will run in the x44 to x48 range.

I picked up a new combo this morning... So we will see how it goes.
Audioaficionado's Avatar
Let the festivities begin!

Boulard83's Avatar
I just ordered a Gigabyte Z87X-OC and a 4770k, will make use of this guide
Theocnoob's Avatar
Thanks for paying tribute to my Sandybridge overclocking thread.

Your thread is much better. Beautifully written guide. Very long and highly informative. Awesome job man.
Reefa_Madness's Avatar
@hokie, thanks for the reply.

I nearly always bring up the rear when it comes to trying out the new platforms so I really appreciate all the blood, sweat & tears that you guys go through so that users like me don't have to.
PolRoger's Avatar
These chips do seem to run rather toasty... Initial quick (~30 min) Prime large fft at 45x with fixed vcore @1.2v.
hokiealumnus's Avatar
Looks solid so far PolRoger. Don't forget to fill out the form when you're done dialing it in!

EDIT - OMG mdcomp fail. Stand by for form insertion!
hokiealumnus's Avatar
***A separate thread for results was created. These are the same forms and will still work, but to see the graphed results, check out the Intel Haswell i7 4770K & i5 4670K Results Thread!

i7 4770K Result Form



i5 4670K Result Form



I put them in the OP as well.
PolRoger's Avatar
Here is another quick (~30 min.) "new normal toasty" overclock @46x. This makes me wonder just how well the average budget/entry level air cooler will manage with these new chips...
Reefa_Madness's Avatar
@PolRoger,

So how does your ram overclock compare to IB? DDR3-2400 CL9 at 1.625v looks pretty good. Were you able to run those same specs before? Are those one of your GSkill kits?
PolRoger's Avatar
The previous screenshot (2400C9) was with two sticks from a high bin Dom Platinum kit. You have the same one... But perhaps they are still boxed up in a sealed display in your DRAM Memory Museum?

Well of the few IB chips that I've binned... DDR3-2666 seemed to be the max with regards to IMC that I was able to achieve. Still kind of early to say but I hope my sample of Haswell turns out be better than average IMC wise.

I finally got this kit now up to rated speed.
hokiealumnus's Avatar
Looking good PolRoger! What board do you have?

As the second person to actually apply the guide (EarthDog was the first), what do you think of it? Anything missing? Tips that worked...or didn't? Any tips to pass on yourself?
MattNo5ss's Avatar
It says Z87-Deluxe on the sticky note

I should have a chance to test out the guide soon, and I'm really looking forward to the more in-depth OCing.
Audioaficionado's Avatar
How much more performance (benchmarks and real world apps) is Haswell over Sandybridge pushed to the limits?

Also HW LGA 1150 vs SB 2011?

Just trying to get a feeling of how much better, not newer, HW > SB.

I'm most likely going with high-end air. Might consider water if it's a good bang for the buck in performance increases.
xhacker's Avatar
You guys are driving me nuts. My Maximus VI Extreme is back-ordered until July. My 4770k should be here in a week (big deal without the board). I blew my M5F and CPU, too much voltage trying to bench high Maxxmem scores. I'll post stats when my board arrives.

Using wifes computer (sucks).
hokiealumnus's Avatar
So, I have made my first edit to the guide. Previously, I had suggested manually setting Analog and Digital I/O voltages. Don't bother. At very high RAM frequencies (think DDR3-2800+), setting them actually decreases stability. Instead, leave them on auto. All you should need to adjust in addition to RAM timings & voltage is the System Agent voltage (VCCSA) to 1.25V.

To run a G.Skill DDR3-2933 kit we're reviewing (with the CPU at stock, haven't tried clocking it also yet), all I had to do was set XMP (which set all timings and the RAM voltage to 1.65 V), and manually set VCCSA (System Agent Voltage) to 1.25 V. That's it, nothing else. Rather than locking up randomly, as it did with the I/O voltages manually set, it passed SuperPi 32M happily.

Now, I need to run more strenuous RAM stress testing, but that in itself has already improved matters dramatically.

To wit:

Janus67's Avatar
Nice! time for some hyperpi
PolRoger's Avatar
I really like the guide.

I've also found this ASUS Haswell/Z87 overclock guide to also be another good read... Even if you 're not running an ASUS motherboard. Thanks goes to Raja over at XS for the link:

http://kylebennett.com/ASUS/OCingGuidev0.99.zip

I think with this generation that the new cpu cache/uncore/northbridge/ring bus settings and voltages will be important to explore while dialing in a stable overclock.

My ASUS board defaults cpu cache to 39x on auto setting = 3900MHz Northbridge in CPUz. Overclocks are more challenging to stabilize as you increase multi along with both the speed of the cpu cache and the memory ratio. The ideal Haswell sample will most likely need to be strong in all three areas and I suspect that these performance characteristics... multi/uncore/imc.. will vary a good bit within individual wafer sample as well as production batch.

It looks like I've now reached the thermal limits of my particular sample with ~1.26v vcore and a Corsair H110 cooler.
ivanlabrie's Avatar
Nice PolRoger...did you try for higher clocks? And what about ram?
manu2b's Avatar
Hey PolRoger, what do you think of AIDA64 stability test?
DOM.'s Avatar
I know I need to read the guide but im to lazy

should have my mpower max and 4770K friday
harney's Avatar
Hi all ..great guide

Just built up my haswell running on a GA-Z87-D3HP very impressed with the performance of this cheapo board ...so far im at 4.5 2.25... but hitting a wall @4.6 @2.50.. im on air and do not like to push my luck so will retry in the morning ....will post up results later tomorrow...

-----
going off the oc topic here but would like to post this
I have just upgraded from 1st gen i5 so would have loved to gone ivy ....but 1155 board would have been a bad choice and 2011 is a joke price wise so far i am happy with haswell....

overclocking this thing is crazy as its spitting fire but my guess is the on board volt controller adding more heat and the extra amount of cycles per clock its doing too on such a small die but it does not help me clocking it up i have managed to get 4.6 @2.5v and do not want to push it so i am in the process off trying to calm down the heat this thing is spitting out .....

just been benching old to new.....

decoding some media on my oc 1st gen i5 @4.5ghz takes 11.45 mins sec system power pull full watts 256

same file on the new i7 @stock 4.22 mins sec system power pull full watts 193

using the same psu gfx memory ect

so far very happy lots more performance less power usage ...

but ere is the magic my cpu shows sometimes 0.216 volt that's less than a quarter of a volt @ 3.9ghz crazy how does it do that just shows how good its for laptops...
dudleycpa's Avatar
I haven't been on the OC Forums - and this is what I find? How long you you ask (IIRC the USB bug in the Sandy Bridge days).

@hokiealumnus: Thank you for the guide. Is it really this easy? CPU speed first and them Memory speed second? As long as you have good air/AIO water - you don't have to worry about the heat.

I thought I saw it posted somewhere that O/C are:

50% xxxx speed
30% xxxx speed
10% xxxx speed (which I think was 4.7 - 4.8[?])

@PolRoger: You sir must be 1) Very talented 2)Very Lucky and godlike. You are at 4,700 @ 1.26V. I understand that is with your temps @ ~85?!?!?! That is mind blowing!

I know I going to get a for the system that I use everyday (It is worse than you think - it is NOT the rig in my sig). I figure that I have to bite the upgrade bullet at some point and it is either Haswell or Ivy Bridge.

Thanks Again everyone!
noname2020x's Avatar
Awesome guide!

Hopefully I can use it within the next year.
hokiealumnus's Avatar
Welcome back dudleycpa! You've about got the gist of it, it is relatively straightforward.

As to your vaguely remembered tidbit, that was from an ASUS overclocking guide. Here is a quite directly from their guide, which applies to CPUs that can acheive X GHz at 1.25V.

Again, that is at the 1.25V level. I have a 4.8 GHz CPU, but only at 1.3V and on a dedicated custom water loop. I doubt this CPU would be within the preferred thermal envelope on air.
PolRoger's Avatar
I haven't yet pushed beyond 4.8@1.265v... (not fully stable). I'll probably need my custom water setup to dial in a 4.8 overclock. On the memory front... I've only been able to run my 2x4gb (ds) Hynix CFR kit at DDR3-2800 and when I tried the 2933 ratio I was getting a 55 code? I don't currently have a (ss) Hynix MFR kit.

I like it fine and have been using it since some of the manufacturers started recommending it for IB. However I still like to use Prime95 and or LinX/Intelburntest on a somewhat limited basis. I'm not one who like to flog my chip with long, long runs of stress tests.

Here is a Cinebench @47x:
KonaKona's Avatar
I heard somewhere that intel is still using the cheapo paste under the lid like they did with IB. Has anyone tried delidding one of these and either replacing the lid or running it topless and checking the OC improvement, if any?
hokiealumnus's Avatar
Read my review Kona...
Xeon_KI's Avatar
I5's run a lot cooler and a lot "piggier"

Was only able to get to 4.2 @ 1.25v

Running:
GA-Z87X-OC
H110
650D

128495
Xeon_KI's Avatar
prime 95 crashed shortly after I posted that.
Beat my old 8210 @ 4.66 in Cinebench

128496
PetrPan's Avatar
Great guide!
Quick comment, disabling hyper threading made a substantial impact on temps for me.
Asus Z87-Deluxe
650D
H110

Got it to 45x100 @ 1.29v stable
Temps under AIDA64 are on average high 60's with spikes to low 80's.
With hyper enabled spikes were low 90's.
Xeon_KI's Avatar
I'm surprised HT makes that much of a temp difference, that explains my temps
Going to play with it some more tonight, if I can't get the clocks to improve I think i'll end up returning and just getting an 4770k.
Bleh
EarthDog's Avatar
Oh yes... usually its on the order or 7-10C in my experiences. But, not disabling something I paid for, ya know?!
Xeon_KI's Avatar
You said it, would drive me crazy!

One other thing I noticed...
above were my prime 95 temps, both blend and small fft, but when I ran IBT on maximum.
Was clipping along @ 66c, then wham! 99c and it started throttling.

Was not expecting that, as my 8120 was the same temps no matter if it were prime/ibt/occt
EarthDog's Avatar
IBT = INTEL Burn Test. I guess its 'for' Intel CPU's so maybe it can stress them out more it seems.
harney's Avatar
Yes i am too getting some crazy spikes all seems ok then boom it starts spitting fire.....

I now have problems with random bsod on stock having to stop any benching now to find the root of this...

system seems to halt then ok then bsod
debugging the dmp files im getting this

ntoskrnl.exe had this a few times now
&
iusb3xhc.sys which is a new one

Null.SYS Null.SYS+4842f50 fffff880`011bb000 fffff988`011c4000 0x0000010800009000 0x4a5bc109 14/07/2009 00:19:37

MattNo5ss's Avatar
In general, Haswell is better than a SB-E hex if the application uses 4c/8t or less since a SB-E hex using 4c/8t or less is just a SB (SB-E does have more cache than SB though). Haswell's TDP is also 35% less than hexcore SB-E CPUs.
Audioaficionado's Avatar
Thanx. Looks like a 6 core SB-E still beats a 4 core HW.

I assume HW will have a 6 core processor or more on a different socket eventually. 8 cores on a single socket would be even better.
dudleycpa's Avatar
Thanks for the update/links on the percentages and speeds.

I'm thinking that you need really good water or Extreme Air to even think about O/C.

If was going with a gaming only rig a 4670K might be the sweet spot for a "gaming O/C rig". I realize no Hyper-threading but again it is for gaming. Application wise it shouldn't be a "deal-breaker".

Anyone see any 4670K guides anywhere?

Thanks Again!
Janus67's Avatar
I'm sure the rules apply just the same as the 4770k.


One thing I would like to note about the 4c/4t vs 4c/8t is that with the new consoles coming out this fall (both of which are 8 cores), expect to see some games actually take advantage of more than 4 cores/threads.
Eldonko's Avatar
Hell of a lot better than the 2 ES chips I have can do on mem. I can get em stable at 4.9 no prob but only if I drop mem to 2k or less. Really annoying when you have mem rated for 2800.. Basically it comes to how far you push your CPU. I can do 2400 @ 4.8 but have to drop to 2000 at 4.9.

I can run 32M and such at higher mem clocks but cinebench = insta bsod above 2k regardless of the timings and I/O voltages. Going to pick up a retail and see if IMC is better.
dudleycpa's Avatar

Thanks for that - I was wondering about consoles too.

If 4770k and 4670K both work out the same then I'm a little bummed out.

Starting to think the new motherboards might be the best thing about Haswell.
PolRoger's Avatar
Yes... These chips don't seem to like running higher multi/mem/uncore speeds all at the same time. I guess everyone will behave a little different in this regard.

In some ways they remind me a little bit like the old Bloomfields... Where a good clocker that could run well with lower memory/uncore speeds was much more common then one that could overclock well while still running 2000(+) mem/4000(+) uncore.

Pushing mine sample a little higher...
DOM.'s Avatar
need to read up getting mine in tomorrow, but think I need to understand how to oc them before putting it on the ss/ln2

is there certain volts I don't wanna push to hard?
Coolerman1992's Avatar
lol im getting 4.7ghz at 1.225 stable for awhile now, ive only had it for a few hours lol and using hyper212+ and running at 68 max temp
Coolerman1992's Avatar
4.2 @ 1.225 sucks bro srry :/ im using MSI z87-g41 PC Mate
Woomack's Avatar
Yesterday I received Z87X-OC and 4770K. Some quick tests on box cooler so I didn't try max cpu clock.

Team Xtreem 2666 11-13-13 @ 3000 12-14-14 1.75V ( maybe will be lower voltage, I have to check that later )

128530

The only thing that I can say is that all official BIOSes for Z87X-OC are really bad and betas have some issues too. F5 seems best but has some problems with bclk. When I set 100-102 then is fine, when I set 102.5 then all is freezing. The same 125-127 is ok but 127+ is freezing.
Good that GB has now better dual BIOS with switch etc. During ~2h testing BIOS was corrupted 5-6 times and had to make recovery.
EarthDog's Avatar
Woomack, that is the strap mucking with you. When its left on auto, it will auto adjust when you adjust bclk. Try 160 or so. It seems to work in ranges, so the 1.25 strap works from around 110 to 140 while the 1.67 does 155+. Of course, this all depends on the CPU.

As far as bios corruption, tell me about it.... Ugh. I borked both bios, sort of on the mpower max and struggling to get them back. Part of it is me however.
KonaKona's Avatar
I skimmed over that bit.

Still, exactly what kind of difference does messing with it make? Since IB/HW chips seem to be entirely thermally limited, it could make a few hundred mhz of difference...
hokiealumnus's Avatar
Summation: Yes, they're still using paste; it's not cheapo, Intel paste is excellent. It's the gap caused by the sealant (between the CPU PCB and the IHS) that causes the temps to be higher. Yes, delidding will probably help with temps and possibly get some small gains due to raising the thermal ceiling.
EarthDog's Avatar
Plenty of threads around discussing it...5-20c on average... For the most part, its not recommended unless you have voltage headroom. But for some reason people like bricking warranties on $300+ chips for little reason.
PolRoger's Avatar
Gskill TridentX 2800C11 at 2933C11 1.70v:
Reefa_Madness's Avatar
Nice! It would appear that you've got yourself a good HW sample.
Eldonko's Avatar
Dam, nice. Can you tell me how high you can get your mem at 4.9 with cinebench stability? cheers
EarthDog's Avatar
Nice! How high can you clock the cpu with those memory clocks is the question....at stock speeds, that shouldn't be a huge issue... Overclocked, huge issue.
hokiealumnus's Avatar
I'm very close to getting 4.8 and DDR3-2800. It isn't quite stable, the longest I've been able to run AIDA is 9:25. It's getting there though. If I get it stable I'll post my voltages for you guys.
harney's Avatar
Hi All

Right managed to sort out my BSOD at stock default bios settings seems to have been an issue with the usb 3 driver on this gigabyte board fingers crossed...

so i have managed to bench and get 4.6 @1.2 stable but it requires more to hit 4.7 but my heat is an issue and need to get off air i have got water kit coming next week so will be able to push more then....

But i decided to see what i could get at lower volts @ 1.150 i get 4.3 stable will try and go lower later this weekend..

hokiealumnus's Avatar
Solid volts-per-clock there harney; very nice.

What application are you guys using for the little note thingy? I've seen that around - twice now in this thread - but have no idea where it comes from.
Janus67's Avatar
Sticky notes is built into win7
harney's Avatar
thanks hokie

Yep sticky notes i only found it because some one used it 1st page its a great tool...

These chips are great intel have done a very good job with them in the mobile department i know peeps moan about not being a great over clockers & heat ect but i think you have to look at it a different way that being that there efficiency is great in terms of clocks per cycle power usage .....i mean look at this had this

Intel must be using some form of black magic to get 0.216v @ 3891




I must say this gigabyte z87-d3hp with free 35 be quite silent rock pro for 106 was a great deal gotta love these budget boards....so save your money spend less on board spend more on cpu

The DPC latency is great with this board so its thumbs up for me and my audio work usage 63 micro seconds


PolRoger's Avatar
I'll see what I can do... But I only have an H110 right now and I don't have access to my custom water loop.

I agree... Like Hokie mentions in the guide... Running high multi/high memory speed is hard with HS and my sample follows the same pattern.
Woomack's Avatar
No matter what strap I use, max bclk is +2MHz or it won't boot/freeze. Auto setting gives me 100 pcie/1.25 strap when I set 125MHz. I can change it manually to ~127MHz but 127.5 = freeze in windows or it won't boot when I change it in BIOS. The same for 100 / x1. it can go up to ~102MHz but not much higher.
x1.66 strap is not always booting and it also depends from used BIOS so I stick to 1.25 for now.
I still have 2 BIOSes to check so will report later if it helped or not.
Max memory clock so far is ~3070MHz what seem stable but I had no time to use anything else than a box cooler for cpu.
EarthDog's Avatar
Poor cpu in that case?

Leave strap in auto, just change bclk.

I could go 125 but not boot 167. At that point, I borked both bios so... Will test more tonight when I get home and with a different cpu.
MIAHALLEN's Avatar
Way to carry the torch my friend...excellent write-up
PolRoger's Avatar
I had no success running Cinebench @4.9 with my AIO cooler. Freezing/BSOD even at DDR3-1600 speed. Perhaps I've reached the thermal limits of my cooling?

I tested at 48x and was able to complete this:
jamesman32's Avatar
Hey guys, haven't posted on the forums in quite some time, but after reading through the OC guide I decided to give my 4670k a try, and see what she was capable of.

I've gotten 4.688 Ghz stable at 1.188V (set as 1.180 in BIOS). 4.8 Ghz needs almost 1.3V, which I found odd. I've tried increasing my memory speed (Patriot Viper DDR3 1600) but it wont go a lick above stock without boot failure.

I've attached an image of an IBT run I completed. The OC was done on a Gigabyte GA-Z87MX-D3H and was cooled with a Corsair H55. I haven't done much overclocking since the LGA775 days...How did I do?

EDIT: My bad on the image attachment...can't read the text unless you right-click and view the image in another tab.
harney's Avatar
Right had a little more time today to carry on the under volt project ....

So i am getting bios @ 1.140 cpu vore @4.3 ghz full load not 100% stable touch & go but if i do vcore @ 1.142 its stable see pic i know load says 50% load but did not catch screen capture in time what do you think .....



going to try 4.4 ghz n above @ 1.150volts

ta sy
harney's Avatar
@jamesman

4.688 Ghz stable at 1.180 is good

i do like these haswells re.clocks per cycle in terms of power ratio is great
harney's Avatar
nice
Hi What are your temps
dudleycpa's Avatar
4.688 Ghz at 1.188V is amazing from what I have read. I'm confused as to why you have 1.80 set in BIOS. Are we missing something (and I might be).

I just want to make sure that you don't Fry/Bonk/Hurt the Great O/C you have.
jamesman32's Avatar
Oh my, I just now noticed the error in my post! I'm sorry, I meant 1.180, not 1.80. That'd be horrible!
PolRoger's Avatar
Looks good to me...
Do you have a typo? Should be... ("set at 1.180 in BIOS").
Does your motherboard auto set bclk to 99.77? Can you run default 100?


What are you running your memory and uncore at?
harney's Avatar
Hi polroger

its all new this under over clocking so i am a little unsure what you mean by uncore

I know that the memory is set at xmp 1600 9 9 9 24 std volt if that helps

just unsure re uncore if anything its untouched default maybe i am doing summit wrong

ta sy
dudleycpa's Avatar
Thanks for that! - 4689.29Mhz at 1.180V.

Color me jealous/envious/joyous - all at once. If you don't mind me asking where did you get that 4670K from?
jamesman32's Avatar
Yes, auto is 99.77 for some reason. Even when I manually set it to 100, it kinda hovers around 99.8. Is there another feature you know of that might cause this? Also yes, huge typo on my part! What a difference the number 1 makes, huh?

I bought it from Newegg on release day. The batch # is 313B427. I wonder how the chip might do with a real overclocking board and some tighter RAM.
PolRoger's Avatar
My temps for that Cinebench run were okay... less stress/heat than running AIDA stress test at 47x. I have some earlier posts in this thread showing temps.

CPU cache/Uncore/NB speed will show under CPUz Memory tab as NB Frequency... My ASUS board will auto set to 39x or 3900MHz. Running your uncore at 1:1 can increase performance but can also be harder to stabilize as your raise overclocks and or memory speeds. Here I'm running 200Mhz less than 1:1
hokiealumnus's Avatar
In case anyone missed it in the guide, try to keep cache frequency (NB frequency in CPUz) within 100-300 MHz of your CPU frequency.
dudleycpa's Avatar
I was wondering that too!!!! Then again if you can get that frequency stable with at a decent memory setting I wonder if it would make that much difference?

From what I have read so far I don't think it would be that big or that much at all.

Thanks again for the info.
DOM.'s Avatar
How do you get it to 5GHz?

whats the voltage name for it?

been on SStage seems 5.3 is max for 4c8t load
Knufire's Avatar
128622

Picked this up at MC, will see how it clocks tonight on stock. NH-D14 coming in next week.
ivanlabrie's Avatar
Good luck Knufire!
Will give me an idea of what to expect with my Silver Arrow.
hokiealumnus's Avatar
I used 1.375v to bench wprime1024m stable. Cache was at 4.8 GHz at I think 1.3v. Not positive on the cache voltage, that's off the top of my head.

1.4v was good for superpi and wprime 32m at 5.1. Validated 5.2 at 1.425v. Then I stopped.
DOM.'s Avatar

I was talking about which volt for the nb speed XD

Max I got was 5.5 2c 1.725v nb 4500

1.6v 5.3 nb 4500 wprime stable

Need some ln2 should have some sometime this week

Also more gas for the torch since it has a cb
hokiealumnus's Avatar
It's called CPU Cache Voltage in the M6E's UEFI and I set it at 1.3V for 4.8 GHz CPU Cache.
EarthDog's Avatar
I think that is CPU Ring voltage in the MSI.
DOM.'s Avatar
okay thanks ill have to check in the bios using msi mpower max but waiting for the OCF to come out :P
DOM.'s Avatar
I seen not to go over 1.2 on reg cooling I would guess but not sure on colder cooling
Eldonko's Avatar
heh heh nice, that is exactly what Im running except I have uncore at 4800. Same mem speed and timings. Still working on stability though. Can run some benches (cinebench, PI) but not all.

At 4.9 I have to drop uncore and mem speed speed low that it is not worth it. It sucks up all the headroom for IO.
ivanlabrie's Avatar
Anyone got insight on double sided MFR sticks with Haswell?
Talking 2d performance mainly...
dumo's Avatar
Testing 4X4GB MFR on Hero with B0 chip



Uploaded with ImageShack.us
DOM.'s Avatar
new stepping is already out
dudleycpa's Avatar
Thought the B0 stepping was for engineering samples?
DOM.'s Avatar
ones on the bot show C0 for ES
EarthDog's Avatar
I have C0 for retail... 2x of them in fact.
I.M.O.G.'s Avatar
10 yr badge applied.
hokiealumnus's Avatar
CO is the final (retail-ready) ES stepping and carried over into retail chips. B0 came before C0.
Knufire's Avatar
CPU OCing on the stock cooler...sucks.

4GHz, 1.1V. Stable but pushing 100C and AIDA reports throttling.
ivanlabrie's Avatar
that's bad...talk about sucky die/ihs contact.

Is that a 4770k?
Knufire's Avatar
Yup, a 4770K. Didn't try disabling HT...I'll just wait until Thursday when the NH-D14 comes. . Gonna see what I can pump out of these Ballistics now.
DOM.'s Avatar
that makes sense
ansuex's Avatar
i love the overclocking guide, gonne try to up my i5 4670k in clock later today!
Yesterday before reading this guide i got my i5 at 4389.89 mhz

CPU: I5 4670k (c0) batch: ?
Core voltage: 1.236 (still have to finetune)
Temp: 70-75 celcius on Noctua NH-D14 while prime95 blend test
Motherboard: Gigabyte Z87X-OC

Anyone has an idea on how far i could take this cpu?
So far so good having 8 hours runtime on prime95
About to test other stresstests also
Leegit's Avatar
So what's the initial feedback from you guys on Haswell after you've had some time to play? Seems like lots of issues here and there. High OC CPU/MEM proving to be pretty difficult to obtained?
ansuex's Avatar
Tbh, i have not run into any trouble yet, my first attempt on a (dirty) overclock is so far a succes, i will try to get higer clocks with same voltage and see how that goes later today.
EarthDog's Avatar
Depends on what you consider a 'high overclock and memory' really. I was running, which you should see on the front page shortly, 4.9GHz and 1200Mhz(2400) memory fully benchable through our test suite. But pushing further, since all the parts are essentially connected, would likely cause one to lower the memory speed. It just depends on the CPU.
dudleycpa's Avatar
I think once you're here you won't have a problem sourcing parts.

Overclocking? Too early to tell....

Has anyone seen the database results that is/was the frontpage?
hokiealumnus's Avatar
There haven't been enough submissions to even get a feel for it. Three for each CPU so far.
I.M.O.G.'s Avatar
Could probably embed the results spreadsheet.
hokiealumnus's Avatar
I'll do you one better. On MattNo5ss's keen suggestion, I created a dedicated thread since this form isn't getting a lot of attention. It's for the best.

Intel Haswell i7 4770K & i5 4670K Results Thread!
Eldonko's Avatar
Im using ES C0 chips so it doesnt reflect retail but haswell has been good for me. Have 4.9Ghz on both a 4670K and 4770K (high end H20) and dropping em to 4.8 allows me to run 2800 11-13-13 1T on memory. Maxing a chip out totally wreaks havock on what you can run for mem freq and uncore so I recommend finding absolute cpu max and then dropping by 100Mhz to optimize memory and uncore. Your benches will be the same in the end (or very close) with less voltage and heat.
hokiealumnus's Avatar
What did you use to get 2800 stable? I've tried everything I can think of for the 4.8 + 2800 combo and it always crashes after ~10 minutes. 2600 is perfectly fine.
Eldonko's Avatar
Uncore was what was screwing up my stability. I was running 4.8 uncore and stability was sketchy but dropping it to 4.6 did the trick. However, I think the main point is how close you are to maxing the cpu. If you are close to the max cpu clocks, high mem clocks are very tough to get. Like when I run 4.9 I have to drop mem below 2k for stability. My advice is to drop your cpu clock slightly and find the point where you can run 2800 on mem and within ~200 on uncore. It is just the nature of the architecture, you cant have everything at once without dropping cpu speed slightly.

VCCSA and I/O did nothing for me on the mem side to stabilize 2800 because ASUS did a good job with auto adjustments on those. However tonight I am going work on max mem clocks at lower CPU speeds so those voltages may come in handy in that case.
hokiealumnus's Avatar
Cool, thanks; I'll give that a try then. Mine can run 4.9 stable, but the voltage was too high for my liking (prefer to stay at 1.3V).

VCCSA does a good job for me. I set-it-and-forget-it at 1.25V. Both IOs just screw me up if I try to set them manually. When pushing high speed mem (over 3000, which wasn't as easy as I figured it would be), I left digial I/O alone and added a little bit of analog I/O (~1.2V), which did help some.
Woomack's Avatar
I'm still fighting with bclk and straps on Z87X-OC. What is weird I was able to set 106MHz host/pcie clock and I made profile with these settings. It was working for about 5 restarts and later I couldn't make it boot above 103MHz, no matter if I used profile or not.

I don't have much time lately but I was playing some with memory. 1580MHz all on air so far. There was 1600MHz+ but I forgot to save it and later it couldn't boot
http://valid.canardpc.com/show_oc.php?id=2830417
Eldonko's Avatar
Pushed mem to 2933 at 4.8ghz and that is as far it would go regardless of timings and voltages. Going to try with another CPU to see if I can get 3k. Stable with everything from aida64 stability test, pcm8, cinebench, etc.



Problem is I get better BW and bench scores at 2800 and 11-13-13 1T over 2933 and 12-14-14 1T so Ill prob stick with the 2800. Memory is the latest 2800 Trident kit from G.Skill.
Woomack's Avatar
Bw in AIDA64 3.0 is wild. Even at low CPU clocks I see 34GB/s+ read/write and copy. I don't know if there is some bug in new version but in v2.85 all was much lower.
My CPU sucks but I can make it stable @4.33GHz/4330 uncore and 3000+ memory.

While I'm typing this post my memory is running like that:
http://valid.canardpc.com/2830768
Eldonko's Avatar
Noticed that as well, BW numbers are all over the map. I've been relying on latency more than BW because it is more consistent.
Leegit's Avatar
Wow those timings... 13-31-31-45
dumo's Avatar
Imo, its easier easier to clock mfr ss >3000 with 125 strap.

If clocking >145 bclk then probably more stable with 167 strap



dumo's Avatar
Its mfr single sided
ivanlabrie's Avatar
Rather interesting result for 2 sticks...what about 4x4?
hokiealumnus's Avatar
Looking good dumo!
dumo's Avatar
Depending on board and cpu's imc it seems easier to clock 4X4, especially with M6E which has ram presets for base settings.

A little bit harder with board like M6Hero because of more tweaking involved to find specific ram settings manually.

Thanks Hokie

Awesome review
ivanlabrie's Avatar
I'm going for the z87x-oc...hope it doesn't suck too much for that.
I.M.O.G.'s Avatar
Unlikely to find any board in a similar class to the m6e in regards to memory presets... The m5e was quite impressive with its selection.

Not anything special with actual memory overclocking, but the presets are a convenient head start to customize off of.

One thing I find critically important for benching, that isn't compared anywhere I know of... I almost forgot myself... Extreme tweaking setting (3d01 score boost), and superpi optimization (superpi speed boost). I don't know what those options change, but they are a distinct advantage on the mvg and mve. Those still present on the m6 series? Any other board have those? They make a difference for efficiency, helping not a ton, but enough to matter.
dumo's Avatar
I just started to get a hold on mem. clocking and still more to learn with 3D settings and features on M6E. Yes, the extreme tweaking is in there too.

It will be interesting to compare it with the OCF boards from Asrock which seem tuned specific for memory clocking.
ivanlabrie's Avatar
You guys will make me feel bad going with the oc...maybe for a daily driver, and then save for a M6E or OCF if the giga sucks hehe.
hokiealumnus's Avatar
Yes IMOG, the extreme tweak is still on the M6E. Not sure about anything specific to superpi.
I.M.O.G.'s Avatar
On the MVE/MVG, depending on BIOS perhaps, it was called SPI booster... Something to keep an eye out for I suppose. (or ask JJ about... I sent Shamino an email and will update with what I hear back)

It is only presets. The OC is a solid board and I wouldn't be jealous of it for those... the presets on the M5E helped a little with memory, but I still changed everything to tweak things in for SP32M.
ivanlabrie's Avatar
I know...not sure if the spi booster thing makes such a big difference or if I can't tweak my way through to beat guys using it but well, I gotta get a GB board cause of warranty coverage anyway and I happen to like the OC. (only quirk seems to be the debug led spot, and that the buttons need some drivers installed)
Eldonko's Avatar
I guess the key is that the 3000 mem divider doesnt work. That was my issue before, I was trying to use the 3000 divider instead of 2933 + bclk or 125 strap.
hokiealumnus's Avatar
Hah, I fell for that one too. Even emailed G.Skill before facepalming hard when I figured it out and quickly emailing a retraction.
Eldonko's Avatar
Well yeah, how were we supposed to know about a particular divider not working without trial and error haha
ivanlabrie's Avatar


Props to coolhandluke for finding that...
EarthDog's Avatar
That was in the package of my MPower MAX. I was personally hesitant to share that as it is highly unlikely anythng can be run with much more than 1.3v on air. Perhaps Pifast or SPi 1M, but nothing that maxes out all cores/threads. I really believe, at least out of the two 4770K's I have, that the voltage limit for air to be able to run anything on it is around 1.3v. At 1.5v+ you may not even want to be in the bios under air with the small load there...
hokiealumnus's Avatar
+1. I saw that at hwbot this morning and it didn't make it past my personal filter.
ivanlabrie's Avatar
Oh, interesting...then it's nothing like Ivy voltage wise.
Eldonko's Avatar
Got that 3k I was after. Think there is more headroom, still 1.65v.

PolRoger's Avatar
Nice!... DRAM still @1.65v... What kit are you using? MFR ic?

Would you mind sharing your voltage settings for that bench? I'm able to boot into Windows at 2933 but as I increase bclk towards 3000 I start getting 55 code. I'm running a "deluxe" as well... Although it is possible that my chip's IMC is limited to ~2933 or maybe it is my current kit that is holding me back?
Eldonko's Avatar
Using this kit:



Brand new from G.Skill, built for Haswell last month. I cant say the chips but you can prob figure it out...

1.20v on system agent and make sure your trfc isnt too tight.. But it is more about finding a balance between cpu, uncore, mem that works to get stability at high cpu clocks. A lot comes down to IMC as well but I couldnt get 3k for a while so keep working on it and it may come. I found that adding vdimm didnt do a whole lot but I do hope to squeeze a little more with more vdimm when I have some time to work on it.

I also have a 2666 Trident kit that I have to try that came out with IB release.
PolRoger's Avatar
"1400" series should be d/s Hynix CFR ic... Which I thought might be the case since your trfc was lower than the MFR runs that I've seen. Will that kit also run 2800 11-13-13-35 @1.65? or maybe even 2933C11 @1.65(+)?

Yeah... I'm not even trying to push memory with higher multi(s)... Currently testing/running multi @39x to ~44x with cache/uncore running from 1/1 down to auto's default of 39x.
Eldonko's Avatar
Yep sure will and 2800 at 11-13-13 1T and tightened secondaries is actually better perf wise than 3000 12-14-14 2T. Once you get up at that 3k range you have to loosen up too much for it to be worth it.



Drop uncore below 1:1, stability is tough at 1:1. I do 200Mhz less and it is way easier to work with.

hokiealumnus's Avatar
Dropped uncore a good bit (from 4700 to 4500) and finally got the RAM to cooperate at over 2600 with the CPU at 4800. This is with the RAM running at XMP (other than me changing CR to 1T). CPU @ 4.8GHz & 1.3V. I didn't touch anything else - all other voltages are on "Auto", but supplied in the AISuite screenshot for reference.





Now, I don't know what sort of performance sacrifice it makes dropping uncore by 200 MHz. I'll leave that testing to someone with more time. Consider this just a proof of concept, easy to obtain on the new beta UEFI.

I can use the 125 MHz strap, bump VCCSA to 1.23V and boot at 4.75GHz / DDR3-3000, but it's not stable. Yet.
dudleycpa's Avatar
Duly noted kind sir! Appreciate you posting so well with such "fine minutiae"

Dare I ask who is writing the guide (In due time)
ehume's Avatar
I was up in the air between the ASUS something Pro and the Gigabyte GZ-Z87X-UD4. The Microcenter was out of the Asus board, so I got the Gigabyte . . . and now I can't see anything like an Adaptive Voltage setting.

Is there an Adaptive Voltage setting on a Gigabyte board? Or did I make a strategic error in board-buying?
hokiealumnus's Avatar
It is a platform feature, so it should be in there somewhere. You should see three voltage options when setting vcore - manual, offset and then whatever gigabyte calls adaptive.
ehume's Avatar
Can't find it, which is why I asked. Any way to ask Gigabyte? I put in a request on their tech support system. I'll post what I get. But sometime they aren't so helpful.

Seems as if the function could be put in the BIOS if it's not there now.

Another question: my 4770k seems capable of being stable at 4.7GHz. What multiplier should I set it to if I want Turbo Boost to carry it to 4.7?
ivanlabrie's Avatar
Ask Sin0822, or Dinos over at xtremesystems...or OCN.
Lvcoyote's Avatar
MIT/Advanced Voltage Settings/CPU Core Voltage Control.

Check there for adaptive setting.

Once you manually set the CPU core ratio, the turbo boost no longer works. The only time Turbo Boost will work is when the CPU is set to it's default speed of 3.5 GHz, then Turbo boost will kick it up to 3.9 When under load.
Janus67's Avatar
So I found something that may be a discrepancy or maybe an error in the ASRock software.

In the guide you recommend to put CPU Input voltage between 1.15-1.2, the lowest I can set using the ASRock software is 1.2, and it starts/defaults at 1.7v.
PolRoger's Avatar
Default VCCIN should be around ~1.8v... I've tested briefly down to ~1.6v. Some say that bumping it up to 2.0v or 2.2v can help when running higher cpu/cache multi (i.e. sub-ambient). I've been running mostly on (ASUS/auto) ~1.79v.
MattNo5ss's Avatar
I think that's a typo in the guide. Maybe he meant Vcore (for air cooling) instead
hokiealumnus's Avatar
If I put CPU input voltage in there anywhere it was a typo. Please let me know specifically where it's at (direct quote) and I'll search & fix it.
MattNo5ss's Avatar
In the "Tips and Tricks To Get That Extra MHz" section...
That bullet sounds like you meant to talk about CPU input voltage since changing vcore isn't a new option. You just put the wrong values.
PolRoger's Avatar
If you lowered your cpu cache speed even more... 41x/42x/43x you could probably stabilize higher memory speeds while running reduced volts on both CPU Cache and CPU SA. I suppose for benchmarking purposes running higher cache speeds will provide somewhat better scores than lower cache speeds but for those people looking for a nice stable running daily overclock the use of a lower cache multi might provide a better option?

4.6GHz 41x cache DDR3-2800C12 (Team 2666C11 MFR) Rosetta load ~5+ hrs.
4.6GHz 42x cache DDR3-2933C12 (Team 2666C11 MFR) SuperPi 32M
bmwbaxter's Avatar
Great guide! lots of good info in there Hopefully before the end of the weekend I will have my chip under the SS to see how far it can go.
ThorV2's Avatar
Is 1.4v a safe 24/7 voltage for 4.8ghz on an i5 4670K?
hokiealumnus's Avatar
I think it's a bit high personally. I would stick to 1.3 or lower on Haswell. I temd to be pretty conservative for 24/7 voltages though.
dudleycpa's Avatar
Jury may still be out on what voltage is "safe". In general in my very humble opinion I wouldn't go over 1.4v. If I could run say 1.35v @ 4.7Ghz I would certainly do that.

I don't own a 4670K/4770K so I can't test this out myself. If I could I would for where the voltages and temperatures start to rise with little to be gained frequency wise.

I don't know where I saw this but system builders apparently don't want to go above 4.4Ghz. That should tell you something.

One last thing - it doesn't seem that the 4670k is any better or worse than the 4770K. The graphs on the first post show about 100mhz difference with voltages being similar.

Sorry for the long winded post
Janus67's Avatar
Hey Hokie, was reading through the guide again today and in the Bclk section the table with RAM speed calculations is empty for the 2nd two columns for RAM speed.

Going to give this another shot soon with my 4770k
hokiealumnus's Avatar
Whoops, I'll have to fix that. Thanks for the heads-up!
hyperion1980's Avatar
Very nice review!

I have following config:
Intel Core i5 4670K
Gigabyte GA-Z87-D3HP
Scythe Mugen 3 PC Games Hardware Edition
Crucial Ballistix BLS2CP4G3D1609DS1S00CEU 8 Gig

Vcore is on 1.21
4.4Ghz stable
Memory on 1600Mhz (8-9-9-21) stable (V1.5)

I made it up to 4.6Ghz. with 1.25 Vcore (but unstable)
I was able to set my ram on 1866Mhz, but with optimized defaults, when I ran my stable overclock in combination with 1866, blue screens during Hyper-Pi.

one question:
how do I set adaptive voltage on my gigabyte mobo?
I think it is some of the options in Intel Smart Boost, but I don't know which I need to set...

second question:
Is it possible for me to set the voltage of the memory on 1.65 and retry setting 1866mhz? Also the system agent on a gigabyte mobo isn't a fix value, but only plus or minus a value. Is it safe to set it +0.1 (the difference between 1.15 and 1.25)?
hokiealumnus's Avatar
First question, sorry, I have no idea. Hopefully someone with a gigabyte board will stop by.

For the second question, absolutely give 1.65v a try. I'm not surprised it didn't work, RAM is binned pretty precisely these days. You shouldn't need any system agent voltage at 1866. For the most part that only needs experimenting above 2400.
lamarw's Avatar
Hi all!
Great Guide!
E have a question and maybe someone here can help...
I use extensively excel with macros that took a long time, (With excel 2003 single thread... it works faster than the newers).
I build a pc with... i7-4770k, 8gb Corsair 2130Mhs 1.65V, Asus Z87-C...windows7 (tomorow arrive the Coler Master Seidam 120M)
As the excel 2003 is single tread i dissabled the hypertreading to release a core completly to the tread of excel. I belive its vantageous.

Someone know what is the maximum overclook that it possible on 1 Core only? (I nees the other cores at infra stock speed or 2 dissableds ...)

Thanks,
L
ARFacchini's Avatar
Hard job...

My build is: ASUS Maximus Hero, i7-4770k, Corsair H80i, Corsair XMS3 2000MHz at 9-10-9-27 2T.

Very difficult to stabilize at 4600. Last stable result was:
- CPU 4500 / cache 4200;
- CPU 1.287V / cache 1.2V; (going to try lowering my CPU voltage)
- DRAM XMP 2000MHz at 1.65V;
- Temps at 88~92C with AIDA64 after 15min.

Tried:
- 4600 / 4400 with AUTO Voltage (1.37V / Cache 1.287V) = no sucess and 96C...
- 4600 / 4200 with 1.35V / cache 1.2V = 92~96C on AIDA64, AIDA apparently stable (15min), but crashed Prime95 after 8min.
- 4700 .... not even imagined...

Here in Brazil, beginning Spring season, ambient temps at 25C and rising... think my new H80i is little for the job...
hokiealumnus's Avatar
Getting hot there! I'd say a stronger water lop is in order, but it will cost a fair bit and won't gain you all that much. I'd be happy with 4500 for the summer, then see what the winter brings after six months.
ARFacchini's Avatar
News from the "Hard Job":

- new DRAM: now is 2x8GB G.Skill 2400MHz C10. XMP Profile working perfect at 1.65V;
- got the CPU running at 4500 with cache at 4300, with manual Vcore in 1.345V.

On AIDA "sensors" page, my processor temperature was 68~70C, while my cores were at 80~90C.

Question: TCASE of i7-4770k is 73C, correct? Is this the AIDA "processor" temp? In this case, what is the max core temp considered "safe"?

Thanks,
EarthDog's Avatar
Keep those temps under 90C is my suggestion. You are as hot as you want to be. I assume those temperatures are while running the AIDA stress test? I would go by what Realtemp says or use the cores value for temps.
ARFacchini's Avatar
Yes, during AIDA Stress, with CPU/TPU/Cache/Memory stress.
hokiealumnus's Avatar
Yep, keep temps under 90C for 24/7 use IMHO. I prefer below 85C, but I have water loops to keep it that cool. When I've got an air cooler installed I tend to be more flexible up to the 90C range. The CPU is perfectly safe there. It will throttle to save itself before it gets too hot anyway, but you don't want it ramming up against that limit all the time.
DrDrummer's Avatar
What kind of RAM would you recommend for a 4770k with Maximus Hero? I was planning on getting TridentX 2400 because it's on sale. Should I opt for 1600 and try to overclock instead of getting 2400 and decreasing?
jiccman1965's Avatar
Ok guys I followed the guide, Using SS for cooling CPU set at 5.2GHZ cache at 4.8. Intel stepping off, I can get it to run 3dmark11. When I look at CPUz it said 3389 then when stressed 5.2 keeps jumping around. Any Ideals what the issue is.
greg2601's Avatar
I have had my 4770K w/ GAZ87-UD3H motherboard for about 3 months and I have been to every forum or read 100's of pages of Intel literature to get where I am. Stable at CPU 4.4/ Uncore 3.9/ Base multiplier 36. If I go less than 36 the clock never exceeds 35. My memory is only DDR3-12800 running XMP 1 at 1600 10/10/10/27. I have no Turbo boost enabled & all the C-states turned off. Speed-step is ON. My vcore is 1.2 and my Vrin is 1.8. I have learned that no matter I do, once I hit the TDP of 84watts, things start to go south. I have plenty of cooling, and I plan on playing with some faster memory as soon as it arrives to see if that is part of the problem. Will keep you posted.
hall1k's Avatar
I may be coming in a bit late, but I've found that, contrary to ASUS's claims, I routinely run into the "+0.1V" core voltage problem, causing thermal shutdowns in the middle of long-running calculations. I've had to set a fixed voltage of 1.1V (@4GHz).

Admittedly, the calculations I've been running are possibly even more stressing than Floating@Home. In short, I'm generating phylogenetic trees for massive data sets using the maximum likelihood method. This is, I believe, an almost entirely AVX floating-point workload.
hall1k's Avatar
Also, I've been searching forever for an answer to this, and nobody seems to mention it: What is the actual, stock voltage for these chips? 1.1V?
Sebolatius's Avatar
Hello,

maybe someone here can help me with this problem:
I plan to replace my current AMD Phenom II 965 System with a Haswell System (i5 4670k). My plan is not only to overclock the CPU but also increase the frequencies which are important for the memory bandwidth (Ring bus frequency, memory frequency, etc.) as far as possible.
On my AMD Platform I used Kingston Hyper X CL8 DDR3-2000 RAM which used Elpida Hyper MGH-E/MNH-E chips and could run very nice timings + good frequencies. I have read that this RAM does not work so good on todays Intel Platforms (Sandy-Bridge, Haswell, etc.). Furthermore I only have 2x2 GB. Therefore I plan to get new RAM as well.
Which RAM would you recommend if one not only plans to overclock the CPU but also the memory?
I usually also keep an eye on a good price/performance-ratio but I anyway would like to have a good replacement for the memory which is in place in my current system since the games which I usually play require a high memory bandwidth (Arma 2/DayZ, Arma 3).
Do you have any advice regarding memory with a good price/performance ratio for haswell and which is overclockable?
In another board they recommend the following memory kit's:
Teamgroup TXD38G2400HC10QDC01 (DDR3-2400 CL10, 2x4GB)
and
G.Skill F3-2400C10D-8GTX (DDR3-2400 CL10, 2x4GB)

Would you recommend the same RAM or different RAM?

Thanks in Advance!

best regards,
Sebolatius
Mandrake4565's Avatar
Your best bet is to start a new thread in the Intel Cpu section or Memory section, you'll probably get more responses there.
EarthDog's Avatar
Performance gains on the RING bus and memory are not really to there in every day clocking. In benchmarks, different story. Personally, I would stick with DDR3 1866 CL9 1.5v and not a Mhz higher, that is unless you plan on benchmarking.

I see you list games. Are you sure these actually respond to higher memory speeds and bandwidth? Most games do not...at least at 1080p resolutions and above. If you have some links to some results, I would love to check those out and learn.

Sorry for the late reply, but ot will vary. I have seen some as low a .99 and as high as 1.15.
Sebolatius's Avatar
Yeah, Arma definitely requires a lot of memory bandwidth due to all the objects and textures of the whole map being loaded. It's also one of the few games which benefit a lot from SSD's.
Here you have some sources:
http://www.pcgameshardware.com/aid,6...marks/Reviews/

http://www.realitymod.com/forum/f329...arma-2-oa.html

The Arma-Games/Engine definitely benefits from a higher bandwidth. This is also the reason why I overclocked the NB and HT-Frequency as high as possible in my current configuration. I would choose memory with a lower voltage if it would help me in overclocking the CPU or the ring bus frequency higher but if the memory bandwidth is in the end a more important factor when it comes to performance i probably really choose the recommended DDR3-2400 Mhz sticks.
Or do you guys have any special recommendation (even cheapter but better overclockable RAM)?

Thanks in Advance!

Best regards,
Sebolatius
EarthDog's Avatar
Excellent links, thank you! I had to dig to find the results but did find them. It seems in that testing there are several peaks of high IOPS in that game but overall use is low. I can see why that may get choppy with a slower HDD. Though, 2 SSD's can reach 120K IOPS, even back then...

I didn't see anything on the ram bandwidth though.

Snag some 2400Mhz sticks though and see how she goes.
Sebolatius's Avatar
Regarding the RAM bandwidth you have to check the first link. So I will probably take the 2400 Mhz Sticks then. Everything above (2600 Mhz+) probably doesn't make a lot of sense from a prife/performance ratio perspective or does it? And it is not even guaranteed that the CPU is able to handle everything above 1600 Mhz...
EarthDog's Avatar
I didn't see where anything over 1533mhz ram made a difference in the first link. Conventional wisdom is there are negligible if any improvements over 2133 and even from 1600 to 2133 there are nominal gains at most. 1333 to 1600 does show the most gains.

I'd research a bit more and confirm that something over 1866/2133 really makes a difference. That price difference is significant between 2400 and 1866 sticks.

I see they theorize about the bandwidth thing and triple channel...

Anyway.. you may want to start your own thread on which ram to get as you will likey get more responses.
Sebolatius's Avatar
Alright, I have opened a thread for this topic now:
http://www.overclockers.com/forums/s....php?p=7595767

The point "Triple Channel" you are talking about is an interesting point. Until some weeks ago I also thought that Triple Channel and Quad Channel memory configurations probably deliver better framerates in Arma 2/3 than Dual Channel configurations. But since I have investigated this issue further (I tried to find as much information as possible and compared several values. Unfortunately this was quite difficult because it's almost impossible to find all the specific information in the Internet) Nevertheless, I do not think anymore that it is really a big benefit. As far as I understand the problematic Dual Channel/Triple Channel/Quad Channel mostly is a great benefit if an application is Multithreaded. Single Threaded applications do not benefit a lot from Multi-Channel-memory configurations. Arma was always one of the few games that required a high performance per Core. It has always delivered better framerates on configurations with CPU's which have a small number of high-performing cores in comparison with configurations with CPU's which have a big number of cores which are mostly optimized for multi-threaded applications but have a worse performance per Core. It is a more single-thread optimized game and therefore I do not really expect it to run better on a Triple Channel or Quad Channel configuration.
sotanaht's Avatar
Ok I'm completely new to overclocking, I've never tried it before but I got an i5 4670k and a Noctua NH-C14 cooler so here I am.

At the moment, I'm thinking I must be doing something wrong because the guide here mentions that at a 1.25 voltage and 46x clock the temperatures should be just about maxing out at 90c. Well I've set the core voltage and clock to that and it seems to be working, but the temperature is only averaging 50c during the stability test (core 1 is pushing 60c, the 0, 2 and 3 are around 45-48). I just figure I must not be configuring it right because there's no way it can run that cool if I'm actually overclocking it properly right?

So if anyone can tell me how to verify that everything is working up to this point or else point out what I am missing I would appreciate it.
Itsumo's Avatar
I'm new to this myself, but I've heard of 4770K's hitting over 1.39v at tolerable temps even during the stress test. Heck, I've done it, no core over 79C. I've seen another who put a stable OC at 1.35V. Some chips have excellent heat dissipation. Meanwhile, I OC'd my neighbor's 4770k, and it could go faster with less voltage, but overheated at anything over 1.25v. Literally, it was throttling. at 1.26v

Now to compare.. at 1.29v, my 4770k sometimes spikes to 71 C. At 1.25, I still reach 65 C although the averages are lower. I thought I might have a chip from a batch with particularly good heat dissipation qualities, or it was my water block and 240mm radiator. Could be some chips are even better at heat dissipation without delidding.

I would encourage you to check out the 4670k/4770k Results Thread and compare, results do vary. Post a screenshot too. Also curious if you noticed what batch your CPU was manufactured in.
sotanaht's Avatar
I think I've found a stable voltage for 48x at 1.42V and it only seems to be running 70-80C during a stability test. Ok it doesn't seem to be perfectly 100% stable, it can usually run a stability test for 10+ minutes just fine, but it can crash if I try to do several other things while the stability test is running.

Here's a screenshot of what I'm currently getting on the stability test. The batch# is 3319C231.
mysticode's Avatar
I can't even run at 1.25v on 43x for 10 minutes in Prime95.

Running the 4670k, and an Asus VI Hero motherboard. Any help is greatly appreciated!

Thank you

Edit: CPU-Z is reporting a voltage of 1.264, Prime95 Small is running for 15+ minutes now, along with an HD Youtube video in Chrome (100% load for sure). 43x with the same set 1.25v. RealTemp is reporting 80c average temp. No BSOD yet! We shall see how it goes.
blacksheep242's Avatar
1st of all, excellent article! My i5-4670K, GA-Z87x-OC, and H100i will be here on Tuesday and I will be following this article to a "T." However, I do have a quick question that's been on my mind for a few weeks now. How is it possible that folks are hitting 80, 90, 95+ temps with their (Haswell) CPUs?

Take the i5-4670K for example. Per http://ark.intel.com/products/75048/ this CPU has a Tcase of 72.72C as the max temp. I thought I understood Tcase and Tjunction, but apparently I'm way off. I figured the CPU would be throttling once the Tcase temp is reached. Perhaps throttling IS happening but the CPU continues to heat up to the 80+ temps many folks are reporting?

Clearly I'm missing something here so please don't hesitate to set me straight on this. In fact, I would greatly appreciate it.
mysticode's Avatar

I'm running a 4670k and an h100i as well, but having a ton of problems dialing in a voltage that works. Hopefully you'll have an easier time than me.
sotanaht's Avatar
My understanding about that from asking around elsewhere is that Tcase temp isn't the same as core temp and tends to be much lower, from what I've heard about 20 degrees lower which would explain the throttling around 95.
mysticode's Avatar
Ran some gpu and cpu mining, temps on cpu hit 84c but just now I got a bsod with error 101. Putting up voltage to 1.26v and turning off LCC (read that it doesn't matter for haswell). Running OCCT with avx enabled.

Edit: 1.26v caused another bsod error 101 running OCCT with AVX enabled. Going to 1.265v in bios and seeing if OCCT will do minimum 30 minutes with avx. Two minutes in and OCCT has failed due to temps getting to 86c. I feel if I knock out of avx mode, it won't be an accurate test of the 24/7 load that I'm throwing at this PC as I am doing cpu based crypto mining.

Edit 2: Running at 1.265v, cpu miner is stressing the cpu at 100% load for about 15 min now, max temp is 75c which is pretty good, starting opening Chrome and a few tabs in, bsod error 101. Have to hard reset my PC. At this point I don't think it's a basic vcore voltage issue. Any one have any advanced steps I can try out here?
Johan45's Avatar
Use P95 blend test for your initial OC stability and the blue screens will go away. I found that using anything else for stability,I was failing F@H all the time. So I stabilized with P95 blend and no problems since.
mysticode's Avatar
I was starting off doing P95 Small, and would for sure get a BSOD.

I'm running just a CPU miner, called minerd.exe, and it's BSOD with error 101, so this has me worried.
Johan45's Avatar
I don't know anything about minings but if you get your cpu stable with P95 blend torture test, I'm pretty certain you error will go away. Just sounds like your CPU is unstable to me. The small ffts gets it hot but doesn't test your IMC/Memory very much. That's where the blend test comes in. I see you're using the Hero. Try setting defaults then usinf the CPU UP option and set it for 4.4, run P95 with some type of monitoring software, I use HWMonitor free. This will show you a voltage that the board will set automatically. I tried this with mine for 4.6 and it ran stable with P95blend. Keep an eye on your temps though, it may be more than your cooler will handle.
mysticode's Avatar
That's a really good idea! Is Cpu_Up the max boost setting? Do you know exactly where I can find it in the BIOS?

Thanks again, this is a great idea.
Johan45's Avatar
It's about midway down the tweaker menu. And I think it's just called CPU UP. When you select it you'll have auto 4.4 , 4.6 can't remember if there was a 4.2 option. Like I said this is likely to set the voltage higher so you really should monitor the temps for a while.
mysticode's Avatar
P95 Small get things really hot, like low 80s. OCCT with AVX got things to around 86c and auto stopped, a bit too toasty for my liking. I'd like to keep things around 80c as close as possible.

Should I bother setting anything on my memory? I have it at 1600mhz for a low base, to make sure it's not the issue. Also what's your take on LLC? Use it or no.
Johan45's Avatar
Just give it a shot and find out. With my straight OC at 4.5 G I have the memory at 2400, the cache at 4200 and LLC set to 3. When I did the Auto I had 1866 memor in the machine and that's where it set it cache stayed at 3900 didn't check LLC. What I liked the most was all the power saving stuff is still enabled as well which made it really convenient. It was literally the easiest "stable" OC I have ever done. Just set it and go. There's no guarentee it'll work for you but it did for me.
blacksheep242's Avatar
Ahh, great info thanks!
mysticode's Avatar
Tried stock bios settings, manual tweaking level (no xmp), cpu level up at 4.6, ran p95 blend, and in five mins or so my PC reset. Temps never got over 65c though.

Doing 4.4 now.
mysticode's Avatar
Nothing happened with stock and 4.4. Prime95 small and the cpu stayed at 3.8ghz. So confused.
Johan45's Avatar
Ya don't know what to say. I did check and ther's a 4.2 setting. Try just setting that and leaving everything else alone. I did 4.6 CPU Level UP last night and ran P95 blend torture test for 4 hours. Like I said what works for me may not work for everyone. Either wayTry the 4.2 setting and pay attention to the voltages in your monitoring software to see what your BIOS thinks that cpu needs. Initially you were trying to get 4.3 and couldn't so 4.4 may not be realistic. 4.2 is a good starting point IMO and maybe you can work up from there.
EDIT: I don't see anywhere what ram you're using but I have put some 1866 in for this test. The Hazwell needs more volts to run higher speed ram , at least mine does. When you use the Level up it set the ram to the XMP profile so what ever that is in your case may be throwing it off?????
mysticode's Avatar
Not sure what's happening, but I went ahead and did F5 for defaults, and then set sync all clocks to 42x.

I am running Aida64 with the FPU test currently. It was the only test I could get some heat out of, so I assumed it was the most intense. FPU option only, no CPU or Memory selected.

Been running this at 42x for the past 8 hours, hottest one of my cores got was 83c, full stability thus far. Voltage is at auto, and memory is hard set at 1600mhz to keep it low to rule out as a problem.

How would you recommend I proceed?

Thanks
Johan45's Avatar
At this point it's probably about time youstarted your own thread and put up a signature so people know what you have for a system. What I was saying earlier was that if you want to run your miner you should get things stable with P95 blend. I suggested using the cpu level up to get an idea of the voltage you'd need. Heck if it runs stable then just use it. What is your memorys rated frequency?? You psu there are so many things treally that can cause instability aside from CPU voltage.
mysticode's Avatar
Cheers Johan, thanks for the help. From one guy in Canada to another

Memory is 1866mhz rated. PSU is a new Corsair 760 Platinum.

Edit: New thread: http://www.overclockers.com/forums/s...24#post7612924
TBob's Avatar
I am trying to follow the guide above and either I have a really bad chip or I am doing something seriously wrong. I cannot get a stable boot when I try to OC even to 43 or 44 at 1.20 volts.


Here is my setup:
OS- WIN 8.1 64 bit
CPU- i5 4670k Haswell
CPU Cooler: Noctua NH-U14S
MB: Gigabyte GA-Z87MX-DH3
RAM: Crucial BLT2KIT8G3D1608DT1TX0 (Ballistix Tactical 8-8-8-24- DDR3 1600- 2 x 8GB)
PSU- Seasonic X660 Gold


What I have tried:
If I use EZ Tune I can get 4.6 stable so I am hopeful I am missing something when I delete EZ tune and try to manually set the BIOS

I set RAM Voltage to 1.5v
I set CPU Voltage to 1.2v
I raise Multiplier.

I can get into windows but I get BSOD whea_uncorrectable_error at anything over 40.


The best I have come up with is 42 at 1.250v This is not stable with Prime95 though. Settings for this OC were VRIN- Extreme, CPU VCORE- 1.250, DRAM V- 1.50, CPU Mult- 42. All else unchanged.

Any thoughts as to what else I could or should adjust?
I have seen other guides adjust the Ring Bus or the Uncore or the VCCIN.

Thanks in advance.
EarthDog's Avatar
43/44 at 1.2v may be low. Is you memory rated for 1.5v? What speed and timings did youbset them to? Did you record the settings at 4.6 and try to.work down from there?

It may be best to create your own thread and get focused help there.
TBob's Avatar
Memory is rated 1.5v. I set for Auto first which was 1333 but the XMP is actually 1600 8-8-8-24. I plan to repeat the EZ Tune clock and try to see what it sets at and work from there. Just very surprised at how un-clockable this is using the normal 1.2v setup.
EarthDog's Avatar
Wait, 3.6 to 4.2 is sizeable for a .1v increase. That said, you may have a chip on the bad side of the curve.. too early to tell though.
TBob's Avatar
Ok. So I used EZ tune and it set 4.6 with 1.543V. To get it to run 4.2 seems to take 1.410V. Bad chip?
EarthDog's Avatar
Horrible chip if tbat is actually true. Seems way too high though.

I'd start another thread in the proprer section so we can talk this with a bit more detail... Post the link here...
TBob's Avatar
New to these forums. Where would the correct location for a new thread be?
EarthDog's Avatar
Same section, your own thread though. Im afraid this may be a lot of posts going back and forth. A thread of your own with your specific issues will draw help to your specific problem is why I suggest creating a new one.
davideinmd's Avatar
I can't seem to get past 3.9 and be stable. I've read this article and everyone seems to get past 4.0 easy. Here's my setup:
Gigabyte GA-B85M-D3H motherboard
16GB 1600 RAM
480GB SSD drive
I5-4670K
Old Thermatake liquid cooling that cools only the CPU. Highest temps at 3.9 GHz while running AIDA peaked at 65c after running it for several minutes. Seems like I should be OK in the cooling department.

I have tried all the steps in the post here and some other's but no matter what I do, if I run AIDA when at 4.0GHz, I crash about 3-5 seconds in.

I have tried voltages at 1.2, 1.25, and 3.0. Runs fine at any of those voltages until I increase the multiplier past 39.

Am I doing something wrong?
TBob's Avatar
I have the same issue. See my thread link above in this thread. I can get to 4.0 but only if I set my cores at 40, 40, 39, and 38. It was suggested to me that I change these settings: 1.change the voltage to RAM from stock 1.5 to 1.55? 2. Change system agent voltage(VCCSA) and CPU IO Analog and Digital to .15v offset. These settings did not help me but might help you.
davideinmd's Avatar
OK, I'll give that a whirl. Thanks!

...update

Made those changes on my Windows (7 Pro 64) computer runs fine after. Then I change the multiplier across the board to 41 (4.1 GHz) and it boots the the "Starting Windows" screen but stops before the flag animation gets going and freezes. Same with the previous voltage settings. I don't know why I wanna go faster. At 4.0 I'm pretty well blazing. Just want to!
EarthDog's Avatar
Your motherboard is not exactly made for overclocking. While the motherboard doesn't play a huge role with these modern processors, understand your bought a BUSINESS class (see the B75 versus the Z77 which is for enthusiasts) level board. I am GUESSING its a board issue.
greg2601's Avatar
When I ran into these problems, I reflashed my BIOS back to stock, set everything to AUTO and loaded programs that would monitor everything(voltages, speeds and temps). I started turning up the multiplier and watched what the system was turning up. This gave me a good idea as to how to set voltages, wattage, etc. I used a program called Throttle Stop. You can find it doing a search and it's FREE. It allows you to turn up the clock multiplier while you have the machine running.

Good Luck,
solidstate's Avatar
Batch 3404B341 and I can't even get 4.3 stable on 1.25v. Just to get 4.4 stable its 1.4v+ ! Even the auto 4.4 setting crashes in Prime95 in a few minutes.

Sure it stays really cool but that's because I can't OC it much.

I have the Gigabyte G1 Sniper 5 as well. I haven't even tried to OC the memory yet.

One thing though, I was able to get to the bios settings with 4.6 and 1.20v so I don't think that's a really reliable for the upper 50%

I just got it at newegg btw.

edit: I don't suppose anyone knows of a vendor selling these pre-binned?
hokiealumnus's Avatar
Unfortunately not; you pretty much have to get lucky watching classified sections in forums. You'll pay a pretty penny for pre binned chips too. Well, binned on cold. Maybe not so much on ambient.
ilovejuice's Avatar
I have trouble with OC i5-4670k stability on ASRock Z87 Extreme4 mobo.

My recent settings:

DRAM Voltage: 1,5V
DRAM Frequency: 1333MHz
CPU Clock (All core): x46
VCore Voltage: 1,3V

I didn't touch any other settings, Cache frequency/voltage left on auto too for now (seems to be running x38 at 1,3V). Trying to figure out maximum Clock multiplier.

4.7Ghz was unstable at 1,35V and I didn't feel confident with my temps occasionally jumping over 85 degrees.

So I thought I could settle on 4.6. On 1,29V it crashed in Intel Extreme Tuning Utility Stress Test quite quickly. Raised it to 1,3V - 13 hours stability test passed. I was so happy to finally see this result.

Decided to test gaming performance. Played Borderlands 2 for 1-2 hours, then PC crashed at loading screen. Checked Windows Error log - Kernel-Power critical error.
Tried Battfield 4 - crashed (with a really weird sound that kept going until I hit Reset button) after about 10 minutes, again Kernel-Power error.

Then I raised VCore Voltage to 1,31. BF4 crashed after a little longer this time (coincidence?), 20-30 minutes, maybe more, I'm not sure exactly. 3rd Kernel-Power error.

On the other hand Dota 2 worked fine, I guess it wasn't stressful enough for the CPU.

Could anyone please suggest what should I do? Is there any point in increasing VCore Voltage further?

I hope I didn't damage my chip, as I had tons of crashes/BSODs over the last few days. I've always paid close attention to temperatures which almost never gone above 80 degrees.


Anyway, last night I tried to work with 4,5GHz / 1,25V
Played bf for about 30-40 minutes, didn't crash. But I experienced 2 really bad frame rate drops, each of them for a few second, like 3-5s - 20FPS or less, extremely choppy (never happened before). I checked temps after that game - GPU (asus gtx 780 dc2oc) reached 68 degrees, I've never seen anything above 63, usually 60 max in BF4. CPU temps were normal. Is it possible that it's somehow connected with CPU? I didn't even try to OC my GPU yet... I'm worried that I messed something up. I've also noticed that programs that used to load pretty much instantly sometimes launch with 5s+ delay, I click on icon, nothing happens for 5-10s and then it starts (for example HWMonitor, Origin platform).

Sorry for such a long post, thanks in advance for taking time to go through this. I really don't know what to do. I'll be grateful for your help!
EarthDog's Avatar
If you want that clockspeed, certainly as you need stability.

Not sure about the long pause at 4.5Ghz/1.25v but logic says to raise the voltage and see if it goes away. That said, there is always a slight pause for those applications to come up, no?
ilovejuice's Avatar
Yea, but 5+ seconds is not that short. I'm not really sure if it has anything to do with this OC.
Is it possible that all these freezes/BSODs/resets I had somehow affect system speed?
I forgot to mention that my current input voltage is 1,9V. Could rising this one help? I've read somewhere that I should only touch it if I run into trouble with rising Cache multiplier -but I'm not there yet.
Going to play more after I get home and see if these FPS drops occur again, hopefully it was a one-time thing.
Isn't that weird that I can run 4,5/1,25V, but can't run 4,6 with 1,31? This seems like a big step in VCore voltage.
EarthDog's Avatar
It shouldn't. You can try resetting things back to stock and see if it still happens. That is another way to tell if it was the o/c.

Not weird at all. At some point, you hit a voltage wall.
ilovejuice's Avatar
Could you please refer to rising CPU Input Voltage and trying 4.6 again? Or no need for more than 1,9 and it's only matter of VCore? I will get back here after doing some more tests.
EarthDog's Avatar
You should not need to touch VCCIN at your clocks. It is a matter of vcore and cooling ability. Personally, I would focus on getting 4.5Ghz stable.
ilovejuice's Avatar
I managed to get 4,5GHz/1,24V core and 4,1GHz/1,24V Cache, VCCIN 1,9V. Seems stable. Although guide says I should keep Cache within 100-300Mhz from core - does it mean this OC is bad?

Kept getting crashes with x42+, I am really disappointed with this chip.

I guess I could get better cache multiplier by increasing V, but another tutorial said cache voltage should never exceed VCore V. So maybe I should increase both (but then again, VCore voltage was supposed to be as low as possible...)? It's pretty confusing.
EarthDog's Avatar
No reason at all to touch cache and vccin....
Luke1978's Avatar
A lot of BSOD's can start to affect your install of windows.
solidstate's Avatar
My processor with batch 3404B341 has little or no OC ability so I ordered another one. I used a different vendor so I wouldn't get the same crappy batch.

What are the odds? I order another i7-4770k CPU from a different vendor and get another cpu from the SAME DAMN BATCH.

Now I have two crappy CPUs! fffffffff... Why spend over $300 on a cpu "K" series if it's so likely I'll end up not being able to overclock? I can't even keep stable 4.3Ghz without over 1.3 volts!
EarthDog's Avatar
Batches mean almost nothing these days... Sorry about that bad luck.

Are you sure its the CPU that is the problem? Perhaps start a thread with your system specs in it and we can try helping out...
solidstate's Avatar
I have the Gigabyte G1.Sniper 5 which is how I found this article. I put the current bios which mentioned better OC on 4th gen. I haven't even tried to do anything with the memory yet with the i7-4770k so it's just stock non-profile default settings on the memory.

The new memory is G.SKILL Ripjaws Z Series 32GB (4 x 8GB) 240-Pin DDR3 SDRAM DDR3 2133 (PC3 17000) Desktop Memory Model F3-2133C9Q-32GZH.

I was using 32GB before with 2 16gb sets of Crucial Ballistix 16GB (2 x 8GB) 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800) Desktop Memory Model BLT2KIT8G3D1608DT1TX0. I wasn't really going for OC memory when I got that.

My previous CPU was an i7-3770k which was fine. I compared my PC mark 7 scores and it went from 6716 with the i7-3770k down to 5440. Same SSD, GPU, and disk controller. I only ran the 3770k at 4.5ghz to keep temps down but could have gone higher.

The cooler is a Thermaltake Water 2.0 Extreme which works fine. I replace all of my case fans with Noctua with high pressure versions on the radiator.

I couldn't stay stable with a 43 multiplier at 1.275 volts fixed, even getting BSODs (code 124) a couple of times in 24 hours which watching videos, listening to music, etc. So I tried normal/ +0.199 offset. Prime95 seemed ok until it suddenly shot up to 90C+ and my system froze. So, I went down to .1 which still ran too hot for my comfort when spiking periodically. I'm trying it at offset +0.08 which seems ok for now. I haven't tried anything particularly complicated given the cpu performance so far. I'm just seeing what I can get stable for everyday use. My top voltage is usually 1.329. I have C1E and EIST in active use now as well although when testing originally I ran at 100% fixed voltage full time.

I'll try the new cpu this evening and see if anything is different. I have a Maximus V Gene I was planning to sell off so maybe I'll try the first i7-4770k in that if it's ok for 4th gen with current bios just to make sure it isn't the mobo.

I'll try the new CPU just to see how close it is but if it stinks I'm going to put my i7-3770k back in. :-(
EarthDog's Avatar
This isn't a new thread, LOL! (its going to get confusing as hell helping you and another guy in the same thread... I will see if I can start one from your existing post.

That said, you never mentioned anything you tried as far as overclocking yet outside of you not touching the memory.
EDIT: I see where you mentioned it a few of days ago in this thread.

Let me try to make your post above a thread...
solidstate's Avatar
That's ok, I was mostly just bitching about getting the same batch again after the first CPU was performing so poorly. I don't think there's much anyone can do since I haven't even gotten past the first steps in evaluating the processor(s) and just didn't like the results. I put the voltages on there I'm using but it's just bad luck and there's not much help for that. :-)
EarthDog's Avatar
Try using a fixed voltage instead of the offset...


Good luck!
NoobOCerz's Avatar
HI overclocker folks..

i got a problem when i started overclocking my cpu. My problem is when i boot up up my pc it boots up 2 times and then it continues to its normal way..Is this normal when overclocking?
Audioaficionado's Avatar
This could be normal. What motherboard/processor/etc are you running?
greg2601's Avatar
I have had the same thing happen. When I change the RAM timings it always boots twice. I have not had it happen when I just change the multipliers or voltages. I also flashed my BIOS back to F3. For some reason it seems to let me overclock more. There is also windows trying to keep track of what is going on. HAL is windows Hardware Abstraction Layer. WHEA is Windows Hardware Error Architecture. These are trying to keep track of your system. They can cause a reboot.
NoobOCerz's Avatar
Mobo:Maximus VI hero
CPU: I5 4670k(overclocked to 4.5)
GPU: Sapphire r9 280x toxic edition
Cpu cooler: corsair h100i
PSU: NZXT 850

This is my specs
Audioaficionado's Avatar
No worries. This is normal for your board after you've changed settings.
Octavia's Avatar
Thanks for the great guide.
I have followed all steps and have my Haswell i5 4670k running at 4.5Ghz at 1.25v
I used AIDA64 fro nine hours and it showed no problems. Max CPU temp was 84. Is this too hot? Average was 78.
Cores one and two had a mas of 100.
But it was running fine with no issues
The bottom graph showed CPU throttling at 14% max too but i could not see any spikes on the graph. The red line seemed to be flat the whole time. Could this just have been a random spike?
CPU i5 Haswell 4670k
Motherboard z87pro
Ram 8gig Hyperbeast 2400
Coolong Corsair h80i

Everything seems fine and running watchdogs for a few hours with no issues.
Would appreciate any advice
EarthDog's Avatar
Welcome !

84C is fine...but you then say "mas of 100".. .what does that mean? Max?

Random spike would be my guess.
Octavia's Avatar
Sorry for the Typo that was Max of 100 on core no1
Also should i be concerned abou the cpu throttling 14% notice?

Im just running AIDA64 again after dropping my RAM frequency from 2400 to 1600 and the results seem much better so far.

Thanks for the response! I forgot to screenshot last time but I will on next completion
greg2601's Avatar
Haswell overclocks better with lower RAM speeds in most cases.
Octavia's Avatar
Sorry for the Typo that was Max of 100 on core no1
Also should i be concerned abou the cpu throttling 14% notice?

Im just running AIDA64 again after dropping my RAM frequency from 2400 to 1600 and the results seem much better so far.

Thanks for the response! I forgot to screenshot last time but I will on next completion
greg2601's Avatar
Leave your RAM at 1600MHz dial up your CPU until you reach a MAX stable. Then dial up your RAM until the system becomes unstable.
EarthDog's Avatar
ram shouldn't be causing throttling.

Ram speeds on Haswell RARELY come into effect at your clockspeed and memory speeds. Its not until you get sub-ambient cooling clockspeeds that you need to lower ram.
Octavia's Avatar
Thanks Greg ill try that, it seems to be making a difference. Should i be concerned about AIDA64 reporting "CPU Throttling (max: 2%) - Overheating detected
It only happens once or twice in the whole test and the line is at zero. Below are screenshots from my latest test Id be grateful for any input.

Thanks again




For some reason those images wont display but just copy the address into your browzer
EarthDog's Avatar
Register to the site properly (clicking on link in your email - not saying you didn't but you need to do that to post pics) and host internally...

I cant see 3rd party hosted pictures at the office, and most get deleted after they are not accessed.

greg2601's Avatar
The CPU is getting hot. That is why it throttles. You would not be able to see a 2% throttle on the graph in AIDA. Just believe that it happened. Are you still at 1.25 vcore? If so see if you can dial it back. That will lower temps.
EarthDog's Avatar
His pictures show into the 100C range? That is where that CPU throttles...
Octavia's Avatar
Hi there, some troubles with registration. Below are the most current results after a 3hr 20min stress test with AIDA64 with DDR3 RAM at 1600Mhz
CPU Voltage at 1.25
Thanks for the help everyone. Im fairly new to Overclocking so dont wanna fry anything
greg2601's Avatar
Your core 1+2 show hotter temps than 3+4. You might want to re-mount your cooler. The TIM might not be spread properly. There should not be that much discrepancy between the core temps.
Octavia's Avatar
Ok ill have a check
Octavia's Avatar
On that note,what would be the best TIM you would recommend? Its a copper heatsing on the H80i and i just used the stock. I may clean it and use something else if its better
EarthDog's Avatar
Just grab some mx-2 and call it a day. Paying double or more for 1 degree difference isn't worth it.
Octavia's Avatar
Again thanks for all of the help guys, its been very informative and helpful.
So I checked the mounting and tightened everything and now much better temperatures with no throttling.
But there is still a difference in between cores (see screenshot). is this an acceptable difference or would it be best to remount the whole thing with new compound?
greg2601's Avatar
I don't have a clue.. It sounds almost too good to be true. I might like to try it though, because the price is not out of site. Would sure make things easier when mounting CPU water blocks or Fan coolers.
Octavia's Avatar
So how do you feel about the more current core temps and variations after checking heatsink placement and screw tension?
greg2601's Avatar
Looking better. You do not want the high temps between the to vary more than about 5C if possible. There are going to be times it happens, but on a stress test they try to stress all cores equally. I would leave it there.
Octavia's Avatar
So no need to refit with new TIM? These variations are ok?
Just running AIDA64 again and set the RAM back to 2400 and im gettig lower temps again
Sorry fro taking up so much of your valuable and appreciated advice but im not sure if the sore temps should be uniform.
Octavia's Avatar
Sorry forgot screenshot
Octavia's Avatar
I know im probably a pain right now but ill jump back on tomorrow to ask a few questions about whether to set voltage to adaptive once im happy with a stable overclock
Id really like to get 46 and id be over the moon with 47.
You guys have been so much help, and i really do appreciate it

ThankYou ThankYou ThakYou
greg2601's Avatar
If you are going raise your multiplier above 45 you may need to raise the uncore or ring frequency. I would also look at your CPU ring voltage; it may need to be raised. Also the system agent; I/O Analog; I/O Digital voltages may need to go up. I would recommend against using adaptive voltages on the CPU. Adaptive tends to over-volt the CPU.
Leave a Comment