What are people using to stress test

[||Console||]

Last year around this time I rebuilt my haswell system that had the MB die , right after rebuilding it I was unable to get the same clocks on the new Gigabyte z97 that I got on my asus z87 . I didnt really have time to play with it . I was at 4.7 ghz with the asus and couldnt get that on the new mb . So I just did a easy 4.5 @ 1.3v and left it .

So tonight I finally wanted to see if I can get back to the 4.7 I had before I have made cooling upgrades and figured it would be easy . I have gone from a OLD RDX water blocjk to a new Heatkiller and I have added in 2x120 more rads for a total of 5x120 no gpu on this loop .

Playing with aida 4.7 didnt seem bad CPU only checked 55 deg FPU checked 75-82 , ok lets open old faithfull p95 small ffts BAM 95-100 deg c WOW I wasn't expecting that
Im still playing with voltages but I was not expecting that .

 

[Lochekey]

What version of P95 are you using? The newer versions should not be used on a haswell chip. You need to use a version before 27.9 if I remember correctly.

 

[||Console||]

28.7 fresh download

 

[Lochekey]

Get an older version and try again.

 

[EarthDog]

Realbench and AIDA64, both FPU and the normal test.

 

[wingman99]

I use Pime95 Blend V 28.10. If you don't like the temperature with Prime95, run it with another program at the same time.

 

[trents]

I use Pime95 Blend V 28.10. If you don't like the temperature with Prime95, run it with another program at the same time.

Huh?

- - - Updated - - -

A lot of people have moved away from stress testing with Prime95 toward AIDA64 and/or XTU which don't drive temps up so high but if you run one of them overnight and pass it seems to adequately demonstrate stability of the overclock. Those latter two create a more real life stress test environment than does Prime95, the kind of you would encounter when gaming or rendering AV files.

 

[wingman99]

No there are a lot of people like me who still use prime95. Have you tried running Prime95 with another program the CPU runs cool.

Over the years, Prime95 has become extremely popular among PC enthusiasts and overclockers as a stability testing utility. It includes a "Torture Test" mode designed specifically for testing PC subsystems for errors in order to help ensure the correct operation of Prime95 on that system. This is important because each iteration of the Lucas-Lehmer depends on the previous one; if one iteration is incorrect, so will be the entire primality test.

The stress-test feature in Prime95 can be configured to better test various components of the computer by changing the fast fourier transform (FFT) size. Three pre-set configurations are available: Small FFTs and In-place FFTs, and Blend. Small and In-place modes primarily test the FPU and the caches of the CPU, whereas the Blend mode tests everything, including the memory.

By selecting Custom, the user can gain further control of the configuration. For example, by selecting 8-8 kB as the FFT size, the program stresses primarily the CPU. By selecting 2048-4096 kB and unchecking the "Run FFTs in-place" checkbox, providing the maximum amount of RAM free in the system, the program tests the memory and the chipset. If the amount of memory to use option is set too high, then the system will start using the paging file and the test will not stress the memory.

On an absolutely stable system, Prime95 would run indefinitely. If an error occurs, at which point the stress test would terminate, this would indicate that the system may be unstable. There is an ongoing debate about terms "stable" and "Prime-stable", as Prime95 often fails before the system becomes unstable or crashes in any other application. This is because Prime95 is designed to subject the CPU to an incredibly intense workload, and to halt when it encounters even one minor error, whereas most normal applications do not stress the CPU anywhere near as much, and will continue to operate unless they encounter a fatal error. https://en.wikipedia.org/wiki/Prime95

 

[||Console||]

I'm running real bench stress right now getting 57-63 def @4.8ghz 1.385v
Real test 15min test passed going to run Aida64 with CPU and FPU for 15 min if that passes going to go game for a bit before I run overnight .

 

[trents]

No there are a lot of people like me who still use prime95. Have you tried running Prime95 with another program the CPU runs cool.

But why does running another program with Prime95 make it run cooler? Is it because the CPU must divide its attention between the two? And what kind of "other" programs would you suggest? Never heard this before.

 

[EarthDog]

I don't understand why you would run another CPU program with it... it just makes what P95 is doing less effective and doesn't show anything more? What is the point?

 

[wingman99]

Prime95 is just as effective when folks are running applications with it, the only difference is CPU cycles alternate with the programs. The point is if you want the CPU to run cooler and test multitasking.

 

[||Console||]

So real bench is a hard test to pass I passed the stress test last night then I opened aiad64 with CPU and FPU checked for the night . That ran fine , then I tryed the realbench bench and it crashed =( looks like more tweaking is needed . I also added my 950gtx into my Wc loop last night after doing so I found I needed to add more vcore to my cpu . Im getting closer to stable but not there yet .



-edit After a fresh reboot it ran realbench bench fine , maybe it was tired after 7 hours of aida64 =)

going to call this semi stable for now and see if I get any crashes in my games bf1 ,cs go , hitman absolution , metro last light redoux

 

[Rainwater]

Personally I can not comment on running a stress test with another open application so the Cpu cycles the programs. The main and pretty much only reason I can think why that would cool down the Cpu some is because it is simply not under as much stress some of the time because it is cycling. BUT I do not know for sure so I am not going to argue anything Wingman said. Version 27.7 and up use the AVX instructions which will raise your heat a lot. Version 26.6 will keep your temps about 12-15c cooler. But keep in mind at least for me and my setup, I was able to run version 26.6 for hours, Aida64 for hours, Realbench for hours. Then for my final test I used Prime 95 version 28.10 and couldn't cross the 90 minute mark with the settings I had and needed to raise my vcore .010v to maintain stability. So it seems to me that out of the handful of stress testing utilities I've used p95 version 28.10 was the most demanding and for ME I can't call my 2400 dollar pc stable if a free piece of software called Prime 95 causes it to buckle. Real world use who knows if you run an all nighter with Aida or Realbench you might never crash down the road. But after the 3 days of testing and tuning and going for hours and hours and feeling pretty darn stable just to it crash at the end like that for my own piece of mind I have to be able to overcome 28.10 as well for me to call my pc stable. So it's all about preference, just exactly how stable do you want it and how high of an overclock do you want. And for overall system stability blend test is what you want to run for hours. That may be another reason your temps got so high. Small fft causes a little more heat. My max temp on p95 version 28.10 blend is 83c and I'm just an aio cooler.

 

[EarthDog]

Prime95 is just as effective when folks are running applications with it, the only difference is CPU cycles alternate with the programs. The point is if you want the CPU to run cooler and test multitasking.

You are dumbing down P95s effectiveness... but ok.

 

[Kenrou]

Prime95 28.10 for video editing. Realbench gives me BSoD/freeze instead of dropped threads, Prime95 26.6 only uses SSE and Aida64 doesn't seem to test for AVX2/FMA3 (handbrake crashes straight away with both 2h stable).

 

[wingman99]

So real bench is a hard test to pass I passed the stress test last night then I opened aiad64 with CPU and FPU checked for the night . That ran fine , then I tryed the realbench bench and it crashed =( looks like more tweaking is needed . I also added my 950gtx into my Wc loop last night after doing so I found I needed to add more vcore to my cpu . Im getting closer to stable but not there yet .


-edit After a fresh reboot it ran realbench bench fine , maybe it was tired after 7 hours of aida64 =)

going to call this semi stable for now and see if I get any crashes in my games bf1 ,cs go , hitman absolution , metro last light redoux

Post back how the games run, 4.8Ghz looks like your doing great.

You are dumbing down P95s effectiveness... but ok.

Explain how multitasking with running prime95 reduces the effectiveness, Prime95 still has to complete the tasks?

 

[EarthDog]

I can't get technical, but from the outside looking in... if the system runs cooler using two different tests than one, seems like it isn't working as hard/something is different? If its running and doing the exact same thing, it seems to reason that it should reach the same/similar temps, right? After all, it still 'has to complete the (same) tasks'. In other words, it goes through the same p95 iteration that roasts the CPU whether there is another program running or not. If its sharing/splitting cycles, it isn't exclusively working on that iteration with its full capacity, and therefore reducing its effectiveness...

... at least, that is my logic entrenched theory. What is your take on why that would happen?

 

[||Console||]

Post back how the games run, 4.8Ghz looks like your doing great.

So far games have been fine passes 50 runs of the metro bench @ the settings I play at , 61c was the max temp on the core .
Hitman Absolution Benchmark max temp was 54 c .
I plan on testing Bf1 tonight . (quicjk testing was ok 64c temp on 64 player conquest )

I ran cinebench 5x in a row I can get that CPU a bit higher 737ish if I raise my uncore but thats for tweajking after =)

 

[wingman99]

So far games have been fine passes 50 runs of the metro bench @ the settings I play at , 61c was the max temp on the core .
Hitman Absolution Benchmark max temp was 54 c .
I plan on testing Bf1 tonight . (quicjk testing was ok 64c temp on 64 player conquest )

I ran cinebench 5x in a row I can get that CPU a bit higher 737ish if I raise my uncore but thats for tweajking after =)

Looks like you are doing good. BF1 is a tough game to run stable with overclocking.

I can't get technical, but from the outside looking in... if the system runs cooler using two different tests than one, seems like it isn't working as hard/something is different? If its running and doing the exact same thing, it seems to reason that it should reach the same/similar temps, right? After all, it still 'has to complete the (same) tasks'. In other words, it goes through the same p95 iteration that roasts the CPU whether there is another program running or not. If its sharing/splitting cycles, it isn't exclusively working on that iteration with its full capacity, and therefore reducing its effectiveness...

... at least, that is my logic entrenched theory. What is your take on why that would happen?

I Have the same understanding, it just makes prime95 run slower with shared resources. Some folks don't want to use the new versions of prime95 with AVX and FMA3 because of the heat, so you can slow the process down with multitasking with the new versions prime95 with the added benefit of AVX 2 or FMA 3 running cooler.