CPU stress software tested

S

SvenBent

2[H]4U
Joined
Sep 13, 2008
Messages
3,323
I decided to redo my old CPU stress test and just wanted to share some of the early results

All software that could test multiple items was set to CPU and memory except for Asus RealBench that didn't provide customization for the stress test.

Code: 
Core I7 2700K 4.4     Status  temp    Power

Asus RealBench           OK   79-84   318w*
Aida64 Extreme           OK   74-81   181w

Folding@Home             OK   73-78   182w
HeavyLoad                OK   72-76   184w
Handbrake 1.0.2          OK   71-76   165-177w

LinPack V11.2.2.010      OK   78-90   191-210w
LinPack x2 V11.2.2.010   OK   84-96   220w

OCCT 4.4.3 (Own/large)   OK   78-88   211w
OCCT 4.4.3 (Linpack90%)  OK   82-93   207-214w

PRime95 28.5 (8-8-30)    OK   84-93   217w
Prime95 28.5 (small)     OK   79-89   212w
Prime95 28.5 (Blend      OK   82-92   215w

Passmark BurnInTest 8.1  OK   67-72   104-180w
XTU (CPU)                OK   73-79   176-190w

Winrar                   OK   60-67   151w
7-zip                    OK   68-71   124-173w


* This uses OpenCL to put load on GPU as well.
Not direct comparable against other candidates

- Status indicated whatever the system passed 4hours of testing
- Peak temp are the range of the cores different peak temps
- Power is the powerdraw measured from the wall (+/-2watt)

The high temps is partly due to using a semi defective cooler on purpose.


Note that linpack was only running 4 threads. were as linpack x2 was running two instance of 4 threads and affinity controller so they would each get access to all physical cores.
Prime (8-8-30) is a custome settings of Prime95 that use size 8 for max and min fft size and running each size 30 minutters at a time (not in place)
 
Last edited:
SvenBent said:
I decided to redo my old CPU stress test and just wanted to share some of the early results

All software that could test multiple items was set to CPU and memory except for Asus RealBench that didn't provide customization for the stress test.
Click to expand...

This is excellent.
Thank you for this. I was very curious about this. So, Linpack continues to be the most taxing.
Great work with the affinity.

About Asus' realbench - are there any limitations as to what OS and GPU can be used? Any openCL card on Win7+ ?
 
What part of the cooler is "semi defective"?

Also,where do you download Linpack?
 
Ocellaris said:
What part of the cooler is "semi defective"?

Also,where do you download Linpack?
Click to expand...

Its an old corsiar AiO watercooler
- the radioed is mounted upside down so any air rises to the where the "hose" connects.
- The pump is having some constant mechanical noise.

Its use on purpose for simulating a bad cooler environments.

I just lowered the voltage a bit and started 2n iteration of the test.. im going to repeat that until all software errors out withing the 4hours allotted time and thereby can see which software detects error the fasts both in time and on instability level

Also please note due to the nature of the test running on differente times of the day there is a small inaccuracy comparing in temp due to varians however the AC as been adjusted to keep a constant temperature in the room. I would assume anything within 2c should be considered the same.
 
michalrz said:
One way to get Linpack is downloading IntelBurnTest. It has a GUI for linpack itself.
Click to expand...
Linpack in IBT is old and does not have AVX which is why it falls behinde prime95. ( it did on any AVX capabile CPU in my previous testing and also why its not part of this test ( i might add it in to illustrate the old IBT issue)

Linpack is freely avaialbe from Intel
https://software.intel.com/en-us/articles/intel-mkl-benchmarks-suite

Its also in my CPU stress test suit if you want a GUI
www.techcenter.dk/Phoenix.exe
 
You must log in or register to reply here.
Back
Top