Overclock.net › Forums › Intel › Intel CPUs › == Haswell-E Overclock Leaderboard & Owners Club ==
New Posts  All Forums:Forum Nav:

== Haswell-E Overclock Leaderboard & Owners Club == - Page 1660

post #16591 of 21934
Quote:
Originally Posted by michael-ocn View Post

p95 was my stess test tool the old first gen i7... but no mas... now i use occt for that final pass of testing, i like verifying with a blend of testing
I don't think it's as cut-n-dry as you propose it to be.

p95 would overwhelm my rig's cooling system, it would either crash (most likely) or slow to a thermally throttled crawl. It would be ugly, I don't ever want to put it thru that. However, I believe the video encoders use the avx (and familily) instructions, yet my system is capable of running the x265 and x264 bench/stress tests indefinitely. Ditto occt which also incorporates avx (and familty) instructions. Those stress tests run fine for as long as i've been willing to let them run.

In my view, p95's usage pattern of specific heat inducing instructions creates an adversarial thermal environment that is not necessary or appropriate for general stability testing. My system can execute avx2 instructions just fine, but not when its at 100c.

I started out with an H240X and indeed p95 did overwhelm the cooler after some time. I was on the bandwagon with 'dont run p95 on this platform' and ran 4.4GHz for months at 1.32v using h265 as my primary stabiltity test along side realbench. I decided to build a custom water cooling solution because I suffered a pump failure from the H240. One day, I crashed during game play so I decided to try p95 again to replicate the crash. Sure enough, after 10 minutes I could, repeatedly.

So I decided to bump down to 4.3, lower vCore and VCCIN slightly, and decided I wanted to be able to run 1344k and 448k FFTs for 2 hours before I would call it stable. I also ran a blend test to completion to verify. Then I encode h265 videos for 12 hours using handbrake. I've also started folding just for the hell of it. So far, no problems.

This however doesn't guarantee me a 100% stable OC. I don't believe there is such thing, but it should help get me reasonably close.

Yes, p95 can produce a ton of heat - if you run 8-768k FFTs. The larger the FFT the less heat is produced. The same behavior can be replicated in other stability tools such as OCCT. I don't believe p95 is taboo on this platform so long as you know how to use it.
Edited by hemirunner426 - 1/26/16 at 6:34am
post #16592 of 21934
Hello

While stress testing at stock settings if the temperatures increase any noticeable amount over a period of time the ambient temperature is high or the water cooling system is crap. Testing at stock settings will give no indication of overclocking potential.
post #16593 of 21934
Quote:
Originally Posted by GreedyMuffin View Post

The voltages.

Well the water was warm-ish, been folding for the last week 24/7. Just turned off to switch CPU. Will report back on the final temps. smile.gif

EDIT: Gone 15 min. Same temp.

1.04v at 4.0GHz? Perhaps. There is only one way to find out. Each chip scales differently. You won't know until you start turning up the clocks.

If your on custom water, start at 1.3v manual and 100x45. Stress test it. If you feel comfortable its stable you have a decision to make... you can either up the clocks or start working the voltages down.
Keep in mind you'll probably need to tinker with VCCIN when you start upping vCore. You can start at 1.92-1.95 and keep in mind LLC.
post #16594 of 21934
That was my old CPUs overclock. smile.gif

Will test later, have a test on Friday so will tinker with it tonight i guess. smile.gif
post #16595 of 21934
Quote:
Originally Posted by GreedyMuffin View Post

That was my old CPUs overclock. smile.gif

Will test later, have a test on Friday so will tinker with it tonight i guess. smile.gif

Good luck! thumb.gif
post #16596 of 21934
Quote:
Originally Posted by hemirunner426 View Post

I started out with an H240X and indeed p95 did overwhelm the cooler after some time. I was on the bandwagon with 'dont run p95 on this platform' and ran 4.4GHz for months at 1.32v using h265 as my primary stabiltity test along side realbench. I decided to build a custom water cooling solution because I suffered a pump failure from the H240. One day, I crashed during game play so I decided to try p95 again to replicate the crash. Sure enough, after 10 minutes I could, repeatedly.

So I decided to bump down to 4.3, lower vCore and VCCIN slightly, and decided I wanted to be able to run 1344k and 448k FFTs for 2 hours before I would call it stable. I also ran a blend test to completion to verify. Then I encode h265 videos for 12 hours using handbrake. I've also started folding just for the hell of it. So far, no problems.

This however doesn't guarantee me a 100% stable OC. I don't believe there is such thing, but it should help get me reasonably close.

Yes, p95 can produce a ton of heat - if you run 8-768k FFTs. The larger the FFT the less heat is produced. The same behavior can be replicated in other stability tools such as OCCT. I don't believe p95 is taboo on this platform so long as you know how to use it.

This is what I did also.
I have 2 overclocks :
One at 4,7GHz at 1,23Vcore which is stable Aida64 12 hours and Realbench 8 hours.
i tested it with P95 v28.7 small FFT's....Of course, too much heat and crash after 2 or 3 minutes....

Then I also defined a second overclokc, which I called rock solid and I wanted it to pass 12 hours of P95 v28.7 Small FFT's.
I did it at 4,3GHz and Vcore=1,07V.
Core Temperature max in P95 v28.7 small FFT's was 77°C at 22°C ambient => 12 hours P95 v28.7 Small FFT's without any errors or warnings.
I consider this overclock as rock stable due to this Prim95 test.

Of course, for daily use (Mainly gaming), I use 4,7GHz at Vcore=1,23V.
My war machine
(21 items)
 
  
CPUMotherboardGraphicsRAM
INTEL Core i7-5930K @ Core=4,6GHz / Cache=4,5GHz ASUS X99-Deluxe II (BIOS 1401) GIGABYTE GeForce GTX 1080 Ti Gaming OC 11G with... G.SKILL Trident Z DDR4-3200 32GB=8GBx4 CL14 @ 3... 
Hard DriveHard DriveOptical DriveCooling
SSD SAMSUNG 850 PRO 1TB (System) HDD WESTERN DIGITAL Caviar Blue 1TB SATA3-600 6... LG BH16NS40 NOCTUA NH-D15 with 2 fans NOCTUA NF-A15 PWM pu... 
OSMonitorKeyboardPower
MICROSOFT Windows 10 Professional 64-bit ASUS ROG Swift PG278Q LOGITECH G810 Orion Spectrum COOLER MASTER Silent Pro M2 1000W 
CaseMouseMouse PadAudio
ANTEC Nine Hundred Two V3 (Opened side panel) LOGITECH G402 Hyperion Fury RAZER Goliathus Control (Small) SOUND CARD REALTEK ALC1150 on-board 
AudioAudioOtherOther
SPEAKERS CREATIVE Inspire T6160 HEADSET ACER Predator Gaming WEBCAM LOGITECH C170 PRINTER EPSON DX8450 
Other
FTTH INTERNET CONNECTION : 300Mbps DL / 110Mbps UL 
  hide details  
Reply
My war machine
(21 items)
 
  
CPUMotherboardGraphicsRAM
INTEL Core i7-5930K @ Core=4,6GHz / Cache=4,5GHz ASUS X99-Deluxe II (BIOS 1401) GIGABYTE GeForce GTX 1080 Ti Gaming OC 11G with... G.SKILL Trident Z DDR4-3200 32GB=8GBx4 CL14 @ 3... 
Hard DriveHard DriveOptical DriveCooling
SSD SAMSUNG 850 PRO 1TB (System) HDD WESTERN DIGITAL Caviar Blue 1TB SATA3-600 6... LG BH16NS40 NOCTUA NH-D15 with 2 fans NOCTUA NF-A15 PWM pu... 
OSMonitorKeyboardPower
MICROSOFT Windows 10 Professional 64-bit ASUS ROG Swift PG278Q LOGITECH G810 Orion Spectrum COOLER MASTER Silent Pro M2 1000W 
CaseMouseMouse PadAudio
ANTEC Nine Hundred Two V3 (Opened side panel) LOGITECH G402 Hyperion Fury RAZER Goliathus Control (Small) SOUND CARD REALTEK ALC1150 on-board 
AudioAudioOtherOther
SPEAKERS CREATIVE Inspire T6160 HEADSET ACER Predator Gaming WEBCAM LOGITECH C170 PRINTER EPSON DX8450 
Other
FTTH INTERNET CONNECTION : 300Mbps DL / 110Mbps UL 
  hide details  
Reply
post #16597 of 21934
Quote:
Originally Posted by GRABibus View Post

This is what I did also.
I have 2 overclocks :
One at 4,7GHz at 1,23Vcore which is stable Aida64 12 hours and Realbench 8 hours.
i tested it with P95 v28.7 small FFT's....Of course, too much heat and crash after 2 or 3 minutes....

Then I also defined a second overclokc, which I called rock solid and I wanted it to pass 12 hours of P95 v28.7 Small FFT's.
I did it at 4,3GHz and Vcore=1,07V.
Core Temperature max in P95 v28.7 small FFT's was 77°C at 22°C ambient => 12 hours P95 v28.7 Small FFT's without any errors or warnings.
I consider this overclock as rock stable due to this Prim95 test.

Of course, for daily use (Mainly gaming), I use 4,7GHz at Vcore=1,23V.

That's nice silicon you have there! biggrin.gif
post #16598 of 21934
Tested quick 4500 1.252V. Cache stock, vccin 1.920 llc 6.

Will try lower. Then i`ll try to find cache OC.

Can anyone recommend me some good timings for 2400mhz?

EDIT: tried CB15 with 1.203 100x45, tried three rounds. It gives me a slight indication. Will try to lover it even further due temps.

EDIT: Crashed on me when i tried 1.165V for 100x45. ^^

I think i`ll manage 45x100 under 1.200. :-D

Testing cache 1.050.
Edited by GreedyMuffin - 1/26/16 at 9:17am
post #16599 of 21934
Quote:
Originally Posted by hemirunner426 View Post

That's nice silicon you have there! biggrin.gif

Yep, I was lucky when the company who built my computer choose the CPU thumb.gif
My war machine
(21 items)
 
  
CPUMotherboardGraphicsRAM
INTEL Core i7-5930K @ Core=4,6GHz / Cache=4,5GHz ASUS X99-Deluxe II (BIOS 1401) GIGABYTE GeForce GTX 1080 Ti Gaming OC 11G with... G.SKILL Trident Z DDR4-3200 32GB=8GBx4 CL14 @ 3... 
Hard DriveHard DriveOptical DriveCooling
SSD SAMSUNG 850 PRO 1TB (System) HDD WESTERN DIGITAL Caviar Blue 1TB SATA3-600 6... LG BH16NS40 NOCTUA NH-D15 with 2 fans NOCTUA NF-A15 PWM pu... 
OSMonitorKeyboardPower
MICROSOFT Windows 10 Professional 64-bit ASUS ROG Swift PG278Q LOGITECH G810 Orion Spectrum COOLER MASTER Silent Pro M2 1000W 
CaseMouseMouse PadAudio
ANTEC Nine Hundred Two V3 (Opened side panel) LOGITECH G402 Hyperion Fury RAZER Goliathus Control (Small) SOUND CARD REALTEK ALC1150 on-board 
AudioAudioOtherOther
SPEAKERS CREATIVE Inspire T6160 HEADSET ACER Predator Gaming WEBCAM LOGITECH C170 PRINTER EPSON DX8450 
Other
FTTH INTERNET CONNECTION : 300Mbps DL / 110Mbps UL 
  hide details  
Reply
My war machine
(21 items)
 
  
CPUMotherboardGraphicsRAM
INTEL Core i7-5930K @ Core=4,6GHz / Cache=4,5GHz ASUS X99-Deluxe II (BIOS 1401) GIGABYTE GeForce GTX 1080 Ti Gaming OC 11G with... G.SKILL Trident Z DDR4-3200 32GB=8GBx4 CL14 @ 3... 
Hard DriveHard DriveOptical DriveCooling
SSD SAMSUNG 850 PRO 1TB (System) HDD WESTERN DIGITAL Caviar Blue 1TB SATA3-600 6... LG BH16NS40 NOCTUA NH-D15 with 2 fans NOCTUA NF-A15 PWM pu... 
OSMonitorKeyboardPower
MICROSOFT Windows 10 Professional 64-bit ASUS ROG Swift PG278Q LOGITECH G810 Orion Spectrum COOLER MASTER Silent Pro M2 1000W 
CaseMouseMouse PadAudio
ANTEC Nine Hundred Two V3 (Opened side panel) LOGITECH G402 Hyperion Fury RAZER Goliathus Control (Small) SOUND CARD REALTEK ALC1150 on-board 
AudioAudioOtherOther
SPEAKERS CREATIVE Inspire T6160 HEADSET ACER Predator Gaming WEBCAM LOGITECH C170 PRINTER EPSON DX8450 
Other
FTTH INTERNET CONNECTION : 300Mbps DL / 110Mbps UL 
  hide details  
Reply
post #16600 of 21934
Quote:
Originally Posted by michael-ocn View Post

GPGPU bench runs a lot hotter for me too compared to the standalone Mandel test, maybe because its all warmed up from from running the whole suite instead of just the one?


Same temps and bench as I'm getting at 4.4ghz with a 1.28v. smile.gif
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
Overclock.net › Forums › Intel › Intel CPUs › == Haswell-E Overclock Leaderboard & Owners Club ==