Overclock.net banner
Cancel
Create thread New Forums

New Linpack Stress Test Released

Jump to Latest
1 - 20 of 38 Posts

Regeneration

· Registered
Joined
·
67 Posts
Discussion starter · #1 · (Edited)
Image

Linpack Xtreme is a console front-end with the latest build of Linpack (Intel Math Kernel Library Benchmarks 2018.3.011). Linpack is a benchmark and the most aggressive stress testing software available today. Best used to test stability of overclocked PCs. Linpack tends to crash unstable PCs in a shorter period of time compared to other stress testing applications.

Linpack solves a dense (real*8) system of linear equations (Ax=b), measures the amount of time it takes to factor and solve the system, converts that time into a performance rate, and tests the results for accuracy. The generalization is in the number of equations (N) it can solve, which is not limited to 1000. Linpack uses partial pivoting to assure the accuracy of the results.

Linpack Xtreme was created because Prime95 is no longer effective like it used to be. LinX, IntelBurnTest, OCCT use outdated Linpack binaries from 2012. Modern hardware requires modern stress testing methodology with support for the latest instructions sets.

Linpack Xtreme is available for Windows, Linux, and as a bootable media. The bootable version is considered to be the most superior as the Linux SMP kernel is a lot more sensitive to hardware instabilities than Microsoft Windows. Watch this video for a short comparison of Prime95 vs. Linpack Xtreme.

Make sure to keep an eye on the temperatures as Linpack generates excessive amount of stress like never seen before.

Changes (v1.1.4):
  • Fixed a crash on AMD Ryzen processors.
  • Updated CPUID HWMonitor to version 1.43.

Downloads:
Linpack Xtreme for Windows | Mirror #1 | Mirror #2
Linpack Xtreme for Linux | Mirror #1 | Mirror #2
Linpack Xtreme Bootable Media
 
Discussion starter · #2 ·
I've updated this release with support for AMD CPUs. In addition, Now you can have unlimited runs, disable Windows' sleep mode (always annoyed me when stress testing), more RAM selection, and included (optional) CPUID HWMonitor for temp monitoring.
 
#3 ·
You could really use some GUI. How much does it matter to have the binaries up to date? It still is same stress test, isn't it?
 
Save
Reply Quote Rep+ Rep+
#4 ·
I like P95 for the same reason you mentioned, it crashes an OC fast and without AVX it doesn't really run as hot (for me) as people claim. I won't be able to use it soon but I'll give it a try at some point. Thank you.
Is the updated linpack harder to pass?
 
Save
Reply Quote Rep+ Rep+
Discussion starter · #5 ·
BroadPwns said:
You could really use some GUI. How much does it matter to have the binaries up to date? It still is same stress test, isn't it?
Click to expand...
The new MKL binaries are optimized for recent CPUs (AVX-512). LinX, IBT use Linpack binaries from 2012.

There's a good reason for the missing GUI. Users can stress test remotely via telnet/SSH, or monitor remotely. RDP is extremely slow when stress testing.

white owl said:
Is the updated linpack harder to pass?
Click to expand...
Yes... especially with the highest problem size (9.5GB of RAM).
 
#6 ·
Console in fine by me.
 
Save
Reply Quote Rep+ Rep+
#8 ·
Thanks for this. Will try it out while squeezing the last drop of performance from my CPU now that my RAM is 100% stable.

:)
 
Save
Reply Quote Rep+ Rep+
#11 · (Edited)
I'm a bit confused:

I have a 8700k (delidded, custom IHS), running at 4.7GHz, 1.22v and this tops out at 62c when stressing 8gb, it will pass 100 trials no problem, however, I only get ~213gflops.

Someone on the techpowerup forums gets 310-320 gflops with a 4.9GHz 8700k. What's up with that?

*edit*
Also, according to HWMonitor, this app only pulls 120w from my CPU, hence the low temps. P95 w/ AVX pulls 170w+ and reaches 75c
 
Save
Reply Quote Rep+ Rep+
#12 ·
Or you could just download linpack from Intel website, same thing, if you need a GUI then use something like LinX etc. with this latest from Intel linpack. Plus this only stresses heavily a part of the CPU unlike other tests, as such it tends to overheat even stock CPUs. It's like the furmark of CPUs.
 
Discussion starter · #13 · (Edited)
Cyph3r said:
I'm a bit confused:

I have a 8700k (delidded, custom IHS), running at 4.7GHz, 1.22v and this tops out at 62c when stressing 8gb, it will pass 100 trials no problem, however, I only get ~213gflops.

Someone on the techpowerup forums gets 310-320 gflops with a 4.9GHz 8700k. What's up with that?

*edit*
Also, according to HWMonitor, this app only pulls 120w from my CPU, hence the low temps. P95 w/ AVX pulls 170w+ and reaches 75c
Click to expand...
Are you using 'all available threads' with HT on? The clocks are normal? Did you try different RAM size? check the task manager, maybe there's something else using the CPU (like Windows 10 automatic app updater).
 
#14 · (Edited)
Cyph3r said:
I'm a bit confused:

I have a 8700k (delidded, custom IHS), running at 4.7GHz, 1.22v and this tops out at 62c when stressing 8gb, it will pass 100 trials no problem, however, I only get ~213gflops.

Someone on the techpowerup forums gets 310-320 gflops with a 4.9GHz 8700k. What's up with that?
Click to expand...

Possibly because it's running the amd version instead of intel. For me 0.9.3 stress test picks the right exe but the benchmark uses amd for two Intel CPU's Ive tried it on. Or maybe it's supposed to be like that. You can confirm with Task Manager.
Image

Note that running the "linpack_intel64.exe" under binaries\x64 manually with same problem size of 14200 and only one thread per core (dual core processor at 3.6GHz) I get 95GFLOPS vs less than 60 in the bench run via linpack_x64.exe (Linpack Xtreme).

Note RAM (Bandwidth) makes a big difference with this bench as well as problem size and to a small extent setting the LDA optimally. Best not to use Intel HTT or at least keep it to one thread per core.

I would hope you should be able to hit 80% (360GFLOP) or more of max possible.
 
Discussion starter · #15 ·
ucode said:
Possibly because it's running the amd version instead of intel. For me 0.9.3 stress test picks the right exe but the benchmark uses amd for two Intel CPU's Ive tried it on. Or maybe it's supposed to be like that. You can confirm with Task Manager.

Note that running the "linpack_intel64.exe" under binaries\x64 manually with same problem size of 14200 and only one thread per core (dual core processor at 3.6GHz) I get 95GFLOPS vs less than 60 in the bench run via linpack_x64.exe (Linpack Xtreme).

Note RAM (Bandwidth) makes a big difference with this bench as well as problem size and to a small extent setting the LDA optimally. Best not to use Intel HTT or at least keep it to one thread per core.

I would hope you should be able to hit 80% (360GFLOP) or more of max possible.
Click to expand...
The compatible binary is used intentionally for benchmarking to deliver more accurate results.
 
Discussion starter · #16 ·
JackCY said:
Or you could just download linpack from Intel website, same thing, if you need a GUI then use something like LinX etc. with this latest from Intel linpack. Plus this only stresses heavily a part of the CPU unlike other tests, as such it tends to overheat even stock CPUs. It's like the furmark of CPUs.
Click to expand...
Linpack vanilla doesn't run on non-Intel CPUs and is optimized for plain benchmarking. LinX and IBT won't work with the new binaries.
 
#17 ·
Cyph3r said:
I'm a bit confused:

I have a 8700k (delidded, custom IHS), running at 4.7GHz, 1.22v and this tops out at 62c when stressing 8gb, it will pass 100 trials no problem, however, I only get ~213gflops.

Someone on the techpowerup forums gets 310-320 gflops with a 4.9GHz 8700k. What's up with that?

*edit*
Also, according to HWMonitor, this app only pulls 120w from my CPU, hence the low temps. P95 w/ AVX pulls 170w+ and reaches 75c
Click to expand...
Any AVX offsets set in bios?


My bench result
 

Attachments

#18 ·
LINPACK is extremely sensitive to memory subsystem performance.

Here is ~310 GFLOPS in the default test on a 4.25GHz core (4.125GHz uncore) 5820K, with DDR4-2667 @ CL12 (my 24/7 stable summer settings on my Primary system):

Image


Cyph3r said:
I'm a bit confused:

I have a 8700k (delidded, custom IHS), running at 4.7GHz, 1.22v and this tops out at 62c when stressing 8gb, it will pass 100 trials no problem, however, I only get ~213gflops.

Someone on the techpowerup forums gets 310-320 gflops with a 4.9GHz 8700k. What's up with that?

*edit*
Also, according to HWMonitor, this app only pulls 120w from my CPU, hence the low temps. P95 w/ AVX pulls 170w+ and reaches 75c
Click to expand...
Probably using a major AVX offset, with a low uncore clock and/or slack memory timings.
 
Save
Reply Quote Rep+ Rep+
#19 · (Edited)
Nope, no AVX offset. 4.7GHz is held solid throughout the test. Selected to use all threads of course.

Uncore/cache is running at 4.4GHz. Again held solid throughout the test.

I have 3600MHz ram @ 16-16-16-36.

Every other stress test runs as expected, this one throws a weird low gflops result and the actual power draw is lower than expected.

*EDIT*

Even though I was getting passes on all tests I tried bumping my vcore from 1.22v to 1.3v just in case. No difference.
 
Save
Reply Quote Rep+ Rep+
#20 ·
What's the GFLOPS reported in the stress test without using HT (only 6 threads)?
 
Save
Reply Quote Rep+ Rep+
1 - 20 of 38 Posts
About this Discussion
37 Replies
12 Participants
Last post:
Regeneration
Overclock.net
posts
28M
members
562K
Since
2004
The premier forum for overclocking experts and enthusiasts. Discuss hardware optimization, custom builds, benchmarking, cooling solutions, and pushing the boundaries of computing performance. From beginner guides to extreme overclocking, join our technical community to master system tuning.
Full Forum Listing
Explore Our Forums
Intel CPUs Hardware News NVIDIA PC Gaming Water Cooling