• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Help with FX-6300 Overclock

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

CerebralEdge

Registered
Joined
Jul 22, 2013
Location
Canada
Hey everyone, I've been tinkering around with my second build for a while now, but I'm looking to squeeze some more out of my CPU.

Here's my rig:

CPU: AMD FX-6300
COOLER: OEM Corsair H60 w/ Push/Pull
MOBO: GIGABYTE 990FX-UD3
GPUs: HIS Radeon 7950 and MSI Radeon 7950 Twin Frozr III CrossfireX
P/S: Corsair HX750
RAM: 8GB Corsair Vengeance
SSD: 60GB OCZ Agility 3
HDD: 2TB Seagate Barracuda
CASE: Fractal Design MIDI R2

Currently, I'm running my FX-6300 at 4.39Ghz, but judging by others' results I think I can go higher.

My Vcore is 1.416V Idle, and 1.464V under load with my LLC settings.

Currently this overclock is 24/7 Prime95 Stable. I also installed the 60mm fan that's included with the stock heatsink over the VRMs on my Motherboard to aid in cooling.

Here is a picture of my Temperatures after 30 mins of Prime95.

PRIME95TEMPSjpg.jpg

I'm a little confused however, is TMPIN2 my CPU Socket/VRM temp? Or is it my CoreTemp?

When playing Crysis 2 or Battlefield 3 TMPIN2 stays around 45c, so if that's my CPU CoreTemp, I think I can risk running it hot to do a stability test.

If TMP1 is my CPU CoreTemp, I think I have at least 5 degrees more headroom to go up.

If I adjust my LLC and offsets, I can hit the 1.524-1.536v area if needed.


What do you think? Can I push her even higher?? 4.7Ghz?
 
I'm amazed how much better temps you have at that clock than I do. I had the UD3 and it ran pretty hot. Can you possibly show us some pics of your set up? And what is your ambient temp?
 
Capturing the full screen of a large monitor makes reading/viewing the captured image pretty tough.

Generally speaking to suggest changes it is easier if we know certain things. Of course beginning with a setting that is 'known' to pass P95 Blend mode makes good sense, because a failed P95 Blend test is not going to give a baseline of a configuration that 'does' work.

CPU Tab in CPUz from CPUID com
attachment.php


Memory Tab in CPUz from CPUID com
attachment.php


SPD Tab in CPUz from CPUID com
attachment.php


And this is screen capture of HWMonitor (free version) from CPUID com
HWMonitor has been scrolled enough and large enough to show Min/Max of Voltages and includes the CPU Core Temps which is nOw called "Package" Temps fully visible. Most of us that assist in this forum section have been seeing users posts of captures of HWMonitor for at least 2 years and are very familiar with any foibiles that might occur. That is why we stick to one we are very familiar with and not have to go and learn a new piece of software. And we like free software mostly or that is what most use.

This capture is made of HWMonitor after it has been open on the desktop logging Min/Max temps and voltages while Prime 95 was running Blend Mode test on all cores for at least 20 mins and then the capture of HWMonitor was made and it shows the Min/Max temps and voltages before P95 Blend was started and while running P95 Blend mode and gives much greater insight into how the system is performing without guessing.

attachment.php


In order to attach screenshots of INDIVIDUAL images as suggested, first crop and capture the images with Snipping Tool found in Windows Accessories or equivalent. Then click on Go Advanced, a button at the bottom of every new post window. Then click on the little paperclip tool at the top of the Advanced post window when it opens. Clicking on the paperclip tool brings up the file browser/upload tool and the rest is fairly obvious.
 
Hey, thanks for getting back to me so quickly!

I'll take some pictures of my cooling setup for you Gilbert.

RGone, as for the CPU Z validation/HW monitor shots, I'll run some tests tomorrow and upload the results.
 
We don't want a Cpuz validation but the individual captures of CPUz Cpu Tab, Memory Tab and the SPD tab. Also individual HWMonitor capture.
 
First up, I got a picture of my cooling setup:

WP_20130917_002.jpg

1x140mm intake on front (upper position)
1x120mm intake on front (lower position)
2x140mm exhaust on top
2x120mm push/pull on radiator exhaust out the back

Ok, so here's the results after running prime95 for just over 20mins using a custom blend setting using 75% of my available RAM.

EDIT: Ambient Temperature during this test was probably around 20c... I don't have a thermometer.

CPU-Z:

CPUZIDLE.jpg

Memory:

CPUZMEM.jpg

SPD:

CPUZSPD.jpg

HWMonitor: (started during idle state, then ran prime95 for 20 mins)

HWMIDLELOAD.jpg


I tried the BCLK method of overclocking to increase my RAM speed while keeping timings tight. The RAM used is 8GB Corsair Vengeance 2133mhz RAM rated at 1.5V. I nudged the voltage up to 1.536V to set timings to 9-9-9-24. Passed MemTest86.

The GIGABYTE 990FX-UD3 limits me to a RAM multiplier of 9.33.

I'm willing to start my overclock from scratch starting at stock if it means I can reach stable frequencies at a lower voltage.
 
CerebralEdge, do you have the Cpu V set manually or is it on auto? I usually don't like when my LLC bumps my Cpu V core that much higher then where I set it. It tends to overshoot how much I really need to be stable. I haven't worked with your board specifically, though you may want to try lowering the Cpu LLC and then re run your stress test at those clocks. You may need to raise the set Cpu V but it will probably be lower then what it's going up to due to the Cpu LLC.

Furthermore, your temps are warm you have around 10c on the socket and maybe 3-5 c headroom on the cores, depending on how hard you want to safely push it.
 
CerebralEdge, do you have the Cpu V set manually or is it on auto? I usually don't like when my LLC bumps my Cpu V core that much higher then where I set it. It tends to overshoot how much I really need to be stable. I haven't worked with your board specifically, though you may want to try lowering the Cpu LLC and then re run your stress test at those clocks. You may need to raise the set Cpu V but it will probably be lower then what it's going up to due to the Cpu LLC.

Furthermore, your temps are warm you have around 10c on the socket and maybe 3-5 c headroom on the cores, depending on how hard you want to safely push it.

I'll double check my BIOS settings when I get home, but I'm using a manual voltage offset and very high LLC.

The problem I had when stress testing is that cores 5 and 6 would give rounding errors in prime95 at lower voltages.

I used a higher LLC to try and lower my base Vcore when running idle.

So to clarify, my socket temps are the ones hitting 59c, and core temps are 49c?

I'm comfortable with pushing core temps around the 55c area if needed to hit 4.6-4.7ghz.
 
You might want to have a look at your ram too. Those timings don't seem right to me. I'd try something like 10-11-10-31-41 1T se if that helps your stability. You're frequency is a bit high for the current timings IMO You mught want to bump up the NB frequency as well to say 2400 with 1.25v on the CPU_NB
 
You might want to have a look at your ram too. Those timings don't seem right to me. I'd try something like 10-11-10-31-41 1T se if that helps your stability. You're frequency is a bit high for the current timings IMO You mught want to bump up the NB frequency as well to say 2400 with 1.25v on the CPU_NB

Johan, if I set my timing back to auto will that suffice? The spec for my ram is 2133mhz 11-11-11-28.

My 4.39ghz OC is 12 hour stable with the RAM at 1950mhz 9-9-9-24.

I'll revert to stock settings and let you guys know what everything sits at. I wanna try some new settings. I see FX chips getting high frequencies around 1.4V
 
IOk you can try leaving them the way they are but I'd still increase the NB frequency a bit the faster the CPU goes I find the faster the IMC needs to run to keep up in intensive apps like Prime95 when you're running tighter ram timings. You don't need to go to defaults.
 
Just checked bios settings:

Stock vcore 1.344v

I had a +0.050v offset with extreme LLC for 4.39ghz
 
Ok, so I tried again to get past my 4.39ghz clock but I ran into issues.

I tried from stock settings without adjusting voltage and was able to boot into windows up until 4.4ghz which was where I started to crash.

I started adding voltage in 0.025v increments, but in order to boot at 4.6ghz it was taking .100-.125v in order to boot.

Now I'm getting lockups and freezes... Should I adjust my ram speeds and timings? I'd like to hit 4.6 GHz but I'll settle for 4.5.
 
Ok, so I tried again to get past my 4.39ghz clock but I ran into issues.

I tried from stock settings without adjusting voltage and was able to boot into windows up until 4.4ghz which was where I started to crash.

I started adding voltage in 0.025v increments, but in order to boot at 4.6ghz it was taking .100-.125v in order to boot.

Now I'm getting lockups and freezes... Should I adjust my ram speeds and timings? I'd like to hit 4.6 GHz but I'll settle for 4.5.

Sounds like your hitting the inevitable wall. Only Rgone can save you now.
 
CerebralEdge, In you last SS you were using a combination of FSB and multiplier to get to the 4386 Mhz clock speed. When you tried to push over 4.4 did you up the multiplier or the Fsb or both? What I feel you should is try to eliminate as many variables as possible that could be causing your instability. If you were using the FSB to get over 4.4 it could be any of the following causing it, the Ram, Ht Link or the NB Frequency, though it being set at 2.1k it's probably not the issue.

What I would do is this, I do agree with Johan, you should give the Cpu Nb V a bump to at least 1.3 or 1.35. I would also drop the Dram Mhz for now to below 1866. Reason being is there is no guarantee that your chip can go much above that speed. I know that I have a hell of a time trying to run mine at anything north of 2k even though my ram is rated for 2133.

Another thing you could try at your own risk, because we do not know what the long term affects are. You can up the raise the Cpu Pll voltage, a while back there was a discussion on this and it seems that if you raise the Cpu Pll on the Gigabyte boards it can help stabilize the Cpu at slightly lower Cpu voltage. Here is the article if you want to read it.

You may be hitting an area where your chip requires a lot more voltage to get to the next X Mhz. It will be up to you whether or not you feel it is worth while to use that much more voltage to get to that clock speed. I would also slow down, if you are having trouble trying to get stable at 4.4 trying for 4.6 will be very frustrating. It is always better to go slow, otherwise you may find your computer flying out the window. :)

Set it back to your known stable settings, drop the dram and ht link multiplier and then try to get it stable at a slightly higher clock. When you find stability, then I would try raising the Dram back and test it again, to see if it's still stable.
 
Mandrake, thanks for the input, I'll definitely try lowering my ram speeds. I upped the NB to 1.25v. But I'll play around to see if it affects stability.

When I pushed it to 4.4ghz the second time around I was using 200 fsb with no bclk adjustment.

I'll try adjusting some settings in my current overclock, and I also have a few semi-stable profiles at 4.52ghz I can tinker with.
 
Maybe...

Looking at your 4396 Mhz run there are some items to see and interpret.

1. WHAT is the Revision of your mobo? It is usually etched into the mobo on the lower left corner. PCB REV X.X.

2. CPU Vcore normally running at 1.464V with a low of 1.416V and a high of 1.476V. = I had access to a mediocre FX-6300 from another forum member and it shook me up with how much cpu voltage it needed to run in the 4.7Ghz region. However it only needed 1.38Vcore to run in the 4.4Ghz range. I recorded all of its' cpu voltages from 4.1 to 5.5Ghz using my very good water setup. Nearly but not quite 0.1 more cpu volts to get 4.4Ghz stable makes one wonder if you got the surely enuff bowwow of a cpu or just what is up. Something seems not so right.

3. From your 4398 Mhz captures you have the HT Frequency running over 2700Mhz which I doubt the validity of. Such high HT Freq may induce an issue. I would rather see it closer to the standard 2600Mhz. Or less than 2600 for the sake of just knowing.

4. Then I see ram that is supposed to have a speed of DDR3-2133Mhz with timings of 11, 12, 12 28, 54 and CMD of 2T running at DDR3-1950 and timings of 9, 9, 9, 24, 50 with invisible CMD rate. That ram speed and much tighter timings than the ram is rated for at its' real speed of 2133 maybe just too dang tightly timed for DDR3-1950.

5. One thing that always ticks me off when I try and look at temps for a Giga board is the fact it is too dang hard to tell which is the CPU_NB or the CPU Socket temp. It seems depending on model of Giga board and maybe even the bios version, no temp monitoring program can stay sorted as to which of the TMPlNx is which. I see that your TMPlN1 is maxing at 49c and hope that might be CPU_NB and then I see TMPlN2 at just 59c. Couple that TMPlN2 at 59c and the Package/Core temp at 59c also is confusing. Generally there is a delta/difference between the Cpu Socket temp and the Cpu Core/Package temp of at least 4c to 6c. There is none. The Socket temp should be that 4c to 6c higher. All of these oddities are "making" me believe you may have a PCB Rev 1.1 mobo with its latest bios. They were all over the place and took maybe 30 times the effort to get dialed in. Anyway I asked what REV your board is in #1 above.

6. I think we are spinning our wheels here. I would drop back to 4.0Ghz with C1E, C6, TurboCore, Cool N Quiet and APM all disabled in the bios. I would manually adjust Vcore to 1.38V and adjust LLC so you do not get a huge jump when running Prime95 Blend mode at 4.0Ghz. Also go into Windows Power Manager and set to Performanc Mode. So winders does not throttling.

I would take a look at the ram timings and DDR3-xxx speed. I would prefer to see 9, 9, 9, 24, 41, and 2T command rate at DDR3-1600 with the ram voltage set to 1.55V or close. That should take the ram out of a possible problem creator. HT Freq 2400Mhz or close to. CPU_NB set to about 2200Mhz or so with CPU_NB at 1.35Volts.

Now run at least 20 mins of Prime95 Blend mode. Capture and post CPUz > CPU Tab, Memory Tab, SPD Tab and the capture of HWMonitor that was running during the P85 Blend testing. Maybe then we can have a more level playing field displayed.

The 990 UD3 came in 5 Revisions:
1.0 > superseded by the 1.1 as the 1.0 was pure crap for FX processors.
1.1 > which a couple of g00d tweakers in here had and did pretty good with. However 1 moved on to the Saberttooth 990FX board and the other actually sold and configured Giga boards and told us all that where the CHV might take 5 runs to adjust stabily, the 990 UD3 might take 50 adjustment runs to reach stability. That is a hale of a difference.
1.2 > No idea since many do not report their Revision.
3.0 > gave most people real problems and was never really sorted without a beta bios to help some.
4.0 > which is the latest and have no news or reports about. It has digital PWM instead of Analog VRM circuit. I have seen Reply emails from users of Giga boards that said some cpu throttling was as designed by AMD and blah blah which makes me pretty certain that Giga at least in their entry level boards would never be my top 5 choice of mobos for FX processors. Maybe a UD5 if Giga wanted to send me one to put thru the paces but currently I have no intention to give them my money when there are better and easier overclocking FX processor boards for running in the +5.0Ghz range with my water setup.

That covers as much of the history and outcomes that we have seen here in this section where we have dealt with 1,000s of "how do I overclock my shett?" threads.
RGone...ster.
 
Rgone, you seem to be very experienced with the over clocking game, I appreciate the post!

So I added a signature to make it easier to identify what hardware I'm running, and I looked at my board... Its a rev 3.0, which is why I put a 60mm cooling fan onto my VRM heat sink. When initially purchased the board, I was unaware of the issues surrounding GB boards. I saw dual x16 lanes for Xfire and a decent price... Haha

Looks like it'll be Asus for me when i upgrade to Excavator down the road.

I'm gonna take the advice I've been given and make some changes to my settings. I'll see if i can find a stable 4.0ghz at a low voltage, upload some screen captures, and take it from there.

I'll be thrilled if i can get a stable 4.5 GHz.
 
Thanks for the updated signature, it helps a lot. CerebralEdge, sometimes when really trying for extreme overclocks it gets easy to lose our way. Finding a good stable starting point and having it to reference from is always helpful. A suggestion is to also take very detailed notes about changes you make.

I had made a suggestion in my last post about dropping your Dram multiplier but I didn't get into the specifics of why I mentioned it. I have issues with my setup running "stable" over 1866. I still haven't pin pointed exactly what the issue is but I did find one of my issues was a bad flash on my motherboard. Anyhow, keeping the ram below 1866 or even dropping it to 1600 with the timings suggested by the Gonester, may help you find a stable overclock easier. This is because it will remove, in theory one factor that could be the issue when pushing your overclock. Once a satisfactory overclock is found, you can always try to raise the Dram and then if there is an issue you know what is causing it.
 
Back