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

Intel 5820K - First Overclock...guidance????

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

JrClocker

AKA: JrMiyagi
Joined
Sep 25, 2015
Hey guys - I have been reading these forums for about a week. Very impressive knowledge base - thank you for all of your posts!

Here is my rig:

Motherboard: MSI X99S Gaming 7
Processor: 5820K @ 3.3 GHz
CPU Cooler: Asetek 550LC AIO
Memory: 4x4GB ADATA XPG Z1
Graphics: EVGA GTX 970 SSC (2nd one on the way for 2 way SLI)
Power Supply: 800 W
SSD: SanDisk 256 GB
HDD: ST 3 TB

I am new to this whole overclock thing...but find it strangely addictive! :bang head:bang head:bang head

I have been able to achieve the following overclock:

CPU: 4.1 GHz @ 1.01V on the core, everything else on auto
Memory: 2666 MHz, 1T, everything else on auto (failed to post on XMP for memory)

Temps:
AIDA64: 72 C after 30 minutes (116 watt draw on processor)
Intel Extreme Tuning Utility: 69 C after 30 minutes (116 watt draw on processor)
Prime 95 (28.5): 91 C after 30 minutes (175 watt draw on processor)

I was conservative on the Vcore as I had Prime 95 give errors at 1.07...so I bumped to 1.10 and have no errors.

Now - here is the question:

What is the realistic max temperature to run at? I have heard people say not to use version 28.5 of Prime 95 as it "draws too much power". But I have had overclock voltages that were stable with the Intel utility and AIDA64 that failed on this version of Prime 95. So, there is definite merit to using this as a stress test. I am just afraid to push the overclock more as the temp on 1 core hit 91 C.

Thank you in advance for your feedback!
 
I got the cache frequency up to 3.2 GHz.

I cannot get the memory beyond 2666 at 1T. Every time I try a 125 MHz clock the mother board will not post. Does anybody have any ideas?
 
Try all at auto but:
CPU ratio x34
Cache ratio x27
CPU voltage 1.20V ( if won't work then 1.25V )
Memory ratio x21.33 ( if you want 2666 )
Memory voltage 1.35V ( so you can play with tighter timings )
BCLK 125MHz should work now at above settings

Later check power saving options ( C states etc ) as you don't need all and when you disable them then it's easier to keep CPU stability after OC.
 
Hey Woomack...thanks for the reply.

I am assuming that I can start from these to get the mother board to post/boot into windows, and then teak down from there (particularly the CPU voltage) for stability?

I am at work now, but will give this a try when I get home.

Thanks!
 
Not sure...my paperwork says "ATNG 800 WATT 80 PLUS POWER SUPPLY". I will have to check it when I get home.

I don't think the power supply is limiting anything though.

I have my system plugged into a UPS, and this UPS has a readout of the amount of power it is sourcing. When I am pushing the PC for stability testing, I have only peaked out at about 380W of power draw from the UPS. At 80% efficiency, the power supply is only running at 304 Watts (or about 38% of rated max power).
 
That PSU is either an Antec, or it is an ATNG. If that is an ATNG, I have not been able to locate any reviews for them. The only thing that I can tur up about them is the they are one of the manufacturers for Rosewill power supplies. At this point you might want to look at a better quality psu.
 
mimart7 - thanks for the feedback.

At this point I am hesitant to suspect the power supply as I have not had any stability issues. Just the issue with getting the BCLK to boot at 125 MHz.


On an additional note - do the temperatures I posted seem high? The reasons I am asking are:

1. This is a custom built PC from CyberpowerPC
2. They put in the components I selected, but the build quality has been very poor. So far I have had to:
--- A. Repair the hard drive mounting: the hard drive was not secured properly, and it fell out during shipping, breaking the power connector on the SSD drive (my heart stopped when I took the computer out of it's packaging and heard stuff rattling around inside the case!)
--- B. Replace the graphics card: the graphics card that shipped with the PC was FUBAR. I bought a new one, and sent the old one back on RMA. On the bright side, I will be SLI when the new card comes back from RMA (about 6 months ahead of schedule) :clap:
3. Re-seat all memory, cables, etc...everything was loose when I first got it...even the screws holding in the components!
4. I ordered a AEROCOOL TOUCH 2000 to monitor internal temperatures/fan speeds
--- A. I am not sure if this is hooked up properly (None of the case fans are sensed by BIOS)
--- B. The 550LC liquid cooler is sensed by BIOS, but it shows constant RPM
--- C. There are a total of 4 fans in the case (not including graphics)
--------- 2 on the radiator of the 550 LC (air flow is net out of the case...not sure if the push/pull was hooked up properly)
--------- 1 on the top of the case near the CPI (air flow is out of the case)
--------- 1 at the bottom front (air flow is into the case)

The temps on my CPU max out at 69 C with the Intel Extreme Tuning Utility at 4.1 GHz overclock with about 118 watts on CPU. The spec for the 550LC says I should have a thermal resistance of 0.15 to 0.1 C/W based on air flow through the radiator. At the high end number (low air flow), this equates to a 18 C temperature rise at 118 watts...I am seeing about 36 C or so above the ambient temperature at this power draw...or about 2x...equating to 0.30 C/W of thermal resistance.
 
I've been using RealTemp as the temperature feedback...it agrees with the Intel tool... and AIDA64 to monitor CPU power.

The thermocouples on the fan controller provide a relative temperature. The temperature rises more slowly (as each is connected to a thermal mass that must be heated up by the adjacent component). Once steady state is achieved, there is an offset to the actual temperatures. Unfortunately, the unit does not allow you to calibrate this offset. It does let you set alarm temps though.

I originally planned on this unit, as the machine has 3 monitors set up in NVIDIA surround...and I didn't have enough space on my desk for a 4th "status" monitor.

I can increase and decrease fan speeds with the controller, and the fans are spinning.

I will check the fan connections tonight. I just have avoided wanting to check the wiring of the internal parts, as the assembler is supposed to know how to do it, and quality control is supposed to check it.
 
OK - dug around:

1. The case intake fan is connected directly to the power supply
2. All other 3 fans are connected directly to the fan controller and not to the motherboard
3. I have all 3 fans running at maximum speed (1700 rpm) and they are really quiet (my keyboard typing is louder)
4. The 2 fans on the chiller radiator suck are in from the case and push it out the back
5. The last of the 3 fans is connected to the top of the case and blows hot air out

The power supply is made by ATNG Power Company (bronze 80 plus rating). The part number is ATA-800FB-B
 
4.1 ghz is pretty good. I am at 4.3 ghz for my cpu.

I am concerned about that psu. I am unable to find any reviews for it. Getting a decent quality psu should be a priority. I have my doubts as tho whether that psu can handle the wattage required for sli.
 
I got the 2 GTX 970 cards (both EVGA) working last night...no issues getting them to run (even got a 100 MHz overclock on both cards)...

I don't understand all the whining I have read on the forums about SLI setup and all that. It was pretty straight forward. Cards run their 100 MHz overclock, stay under 100% power draw (even thought I let it go up to 110%), and stay under 75 C. Performance of both cards is about 1.9X of the single card according to the Heaven 4.0 benchmark. (Shows 65 FPS average at 5760x1080, max everything, no AA).

The heaviest load game I have right now pulls 550 Watts from my UPS. At 80% efficiency of the power supply, it's running at 440W or so, or about 55% of max power). No stability issues.

Other than that it's not a common brand, what do you think could be an issue issue with the PSU?

I think the fans that are on the radiator don't move enough air. I ordered some new fans for my radiator (Cooler Master JetFlo 120 - Red LED...muhuhuhuhuhuhahahaha). The reviews say they move 95+ CFM at max RPM, but with a bit of noise. When they come in next week I will strike the balance between noise and overclock performance.
 
I got the 2 GTX 970 cards (both EVGA) working last night...no issues getting them to run (even got a 100 MHz overclock on both cards)...

I don't understand all the whining I have read on the forums about SLI setup and all that. It was pretty straight forward. Cards run their 100 MHz overclock, stay under 100% power draw (even thought I let it go up to 110%), and stay under 75 C. Performance of both cards is about 1.9X of the single card according to the Heaven 4.0 benchmark. (Shows 65 FPS average at 5760x1080, max everything, no AA).

The heaviest load game I have right now pulls 550 Watts from my UPS. At 80% efficiency of the power supply, it's running at 440W or so, or about 55% of max power). No stability issues.

Other than that it's not a common brand, what do you think could be an issue issue with the PSU?

I think the fans that are on the radiator don't move enough air. I ordered some new fans for my radiator (Cooler Master JetFlo 120 - Red LED...muhuhuhuhuhuhahahaha). The reviews say they move 95+ CFM at max RPM, but with a bit of noise. When they come in next week I will strike the balance between noise and overclock performance.

With a no name brand psu, you run the risk of taking out components when it fails. It also has the potential to be a fire hazard.
 
The PSU has a UL mark, CE mark, and bronze 80 plus rating. If 3 different independent test labs certified it as good...it's good enough for me. (However, if it ends up being an issue, I'll let you say "I told you so" :D

My next focus is to get better cooling on the CPU...

I'm going to tweak around more with the memory tonight.

(Dang this stuff is addicting!)
 
Just put in the new Jetflo 120 fans from Cooler Master. CPU temp dropped 7 C!

These things look cool and move A LOT of air!
 
OK - got the processor overclock to 4.2 GHz (3.0 GHz cache) with a Vcore of 1.15 V. This is with the new fans that I mentioned above.

The temps are:
- Intel Extreme Tuning Utility: 72 C
- Prime95 ver 28.5: 95 C

I use the Prime95 version as it will cause a crash for an overclock that the Intel utility will not cause a crash for. When running, this version draws 200 W of power through the CPU...and my 120 mm cooler barely keeps up. The Intel utility maxes the processor power at about 125 W.

Here is the question:
- I manually adjust Vcore to hit a stable overclock, everything else on auto.
- The processor has more room to overclock, but my current cooler is at its max (maybe an upgrade in the future)
- Power = voltage * current
- Current is proportional to the clock frequency of the chip
- Which voltages should I experiment to adjust DOWN on the CPU to lower the power dissipation?

Thanks guys...love these forums!
 
Which voltages should I experiment to adjust DOWN on the CPU to lower the power dissipation?
You may be able to lower VccIN, but outside of that, vcore. Otherwise, welcome to overclocking a 125W processor... lots of heat when overclocked!
 
4.2 GHz at 1.13V on the core, the rest on auto.

Processor draws 122 W with Intel Extreme tuning facility @ 71 C max core
Processor draws 189 W with Prime95 v28.5 @ 93 C max core
Processor draws 127 W with AIDA64 @ 73C max core

I think I'm good here until I decide to upgrade my cooling! :)
 
Back