New Linpack Stress Test Released

[Hawk-NGO]

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 LinX, IntelBurnTest, OCCT use outdated Linpack binaries from 2012.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

[vario]

Thx, gonna give it a try.I assume it uses avx/avx2.

[Hawk-NGO]

Depends on supported instruction set by the CPU. AVX-512 too.

[mongoled]

I get some FAIL's when running on my setup.

Only ran the short test with max memory.

Not the settings in my sig, got XFR turned on so CPU frequency is control by AMD software.

Memory is at 3200 mhz | 1.36v | 14-14-14-14-28-42-267

All core boost is between 3850-3900mhz.

Load temps are on par with prime95 blend.

Rig is prime95 stable for over 6 hours, but just running a quick 1 hr loop to confirm its still stable .....

** EDIT **
So I let prime95 (29.4.7.0) run for 1:30 hrs and it didnt have any errors....

[vario]

Depends on supported instruction set by the CPU. AVX-512 too.

I like the software, i have a question tho, i have 3 "tiers" of stability, however for gaming i sometimes use a more relaxed version that does not heat up cpu as much, its pretty much a cooling capacity thing ,not stability .Is there a way to run this with some cmd line option to turn off avx1/2 (i dont have 512) ?
At 4.5ghz my cooling is just not enough to keep cpu beneath 100C with all cores going full tilt with avx2 :P

[Hawk-NGO]

I like the software, i have a question tho, i have 3 "tiers" of stability, however for gaming i sometimes use a more relaxed version that does not heat up cpu as much, its pretty much a cooling capacity thing ,not stability .Is there a way to run this with some cmd line option to turn off avx1/2 (i dont have 512) ?
At 4.5ghz my cooling is just not enough to keep cpu beneath 100C with all cores going full tilt with avx2 :P

No way to turn off AVX. Just stress with the case open, or even better, reduce the OC to a level where cooling is sufficient. Some games (and their DRMs) also stress the CPU.

[xpower]

thanks. always eager to test my system to make sure its stability.

EDIT:
@mongoled....might be memory related

[Hawk-NGO]

Linpack is a lot more stressful than Prime95.

[xpower]

@mongoled
retest with latest ver0.7 and see how it goes

[mongoled]

Linpack is a lot more stressful than Prime95.

Yes, this I understand.

Power at wall from prime95 max is around 233-238w

Power at wall from your stress test app (thanks BTW) max is around 248-253w

I upped CPU-NB voltage to 1.050 from 1.025 and now all the test pass (did the 2nd option for run time), although the "Linpack Stress Test" now passes my video driver kept crashing and then recovering (about 5 times when doing the shortest run).

@mongoled
retest with latest ver0.7 and see how it goes

Thanks for the newer version, will do and will report back!

[Hawk-NGO]

I upped CPU-NB voltage to 1.050 from 1.025 and now all the test pass (did the 2nd option for run time), although the "Linpack Stress Test" now passes my video driver kept crashing and then recovering (about 5 times when doing the shortest run).

Glad to hear you got it sorted.

Systems crash with the new Linpack 2.5x times faster than Prime95.

There's an option somwhere in the registry regarding display driver timeouts.

[mongoled]

Glad to hear you got it sorted.

Systems crash with the new Linpack 2.5x times faster than Prime95.

There's an option somwhere in the registry regarding display driver timeouts.

Thanks for the tip, will look it up.

Has something changed with version 0.7/0.8 in how the results are shown?

Now looks like the screenshot below, is this normal?



Re the speed in finding errors, this is very helpful as electricity is very expensive in Cyprus.

Just from stress testing the CPU for a period of three weeks increased my electricity bill by over 30%!

[Hawk-NGO]

Yes, at least for AMD CPUs.

[Hawk-NGO]

v0.9.2
- Added several optimizations for AMD CPUs.
- Improved multithreading efficiency for the benchmark.
- Fixed insufficient memory error on 32-bit systems.
- Updated CPUID HWMonitor to version 1.36.
- Some minor changes.

[mongoled]

v0.9.2
- Added several optimizations for AMD CPUs.
- Improved multithreading efficiency for the benchmark.
- Fixed insufficient memory error on 32-bit systems.
- Updated CPUID HWMonitor to version 1.36.
- Some minor changes.

Hi, someone over at ananadtech forum (.vodka)

was asking if your version uses AVX/2 on Zen based architectures.

Looking at their results it does not look this way.

Will you be implementing this ?

Whats strange for me compared to the results at ananadtech is that when I use either the version you have posted or the Korean version the power draw of my system is identical (248-250w) but the GFLOP are around 70-74 for your version and 151-156 for the Korean version.

Where as the other peeps although also seeing higher GFLOPS were also seeing higher power usage.

** EDIT **
OK a correction to the above

PBO wasnt working correctly, now I have amended it (PPT/141 | TDC/95 | EDC/140 | PBO Scaler/Manual x 10) im seeing the following

Your version of LinX, total system power draw 260-268W, max load temp 68C, GLOPS 72
Korean version of LinX, , total system power draw 278-286W, max load temp 71C, GFLOPS 161

[Hawk-NGO]

Latest version *should* use AVX on Zen (don't have the hardware to verify).

The GFlop measurement is based on the problem size (RAM used).

Try running it with all available threads, or per real core.

[mongoled]

Latest version *should* use AVX on Zen (don't have the hardware to verify).

The GFlop measurement is based on the problem size (RAM used).

Try running it with all available threads, or per real core.

OK, the beta your linked in anandtech forum is acting closer to the Korean version.

Im seeing power draw of 272-278W and temps just over 70C.

So it looks like the beta is acting more like the Korean version.

Am waiting for the stress test to finish and will then update the post with the GFLOPS

** EDIT **
GFLOPS 156, so less than the Korean version which was hitting +160

[Hawk-NGO]

With version 0.9.3, AMD Ryzen users can now bake their CPUs like never before. Thanks to .vodka from AnandTech forums for beta testing.

[Hawk-NGO]

The video above demonstrates the superiority of Linpack Xtreme versus the latest version of Prime95. The overclocked PC passed nearly 2 hours of Prime95's small FFTs torture test and yet completely crashed within less than a minute and a half (77 seconds to be precise) with the bootable version of Linpack Xtreme integrated on Porteus Linux.

Computer specifications: Intel Xeon W3680 @ 4.17 GHz (144x29), ASUS P6T (vanilla), 3x 4GB G.Skill Ares F3 @ 2304 MHz (10-12-12-31 CR1), and Nvidia GeForce GTX 970.

[Hawk-NGO]

Final version (1.0.0) is out now... brings a lot of improvements and fixes.

[mongoled]

Just a heads up,

tried this with my 3600x

Fails at stock settings almost immediately

Prime95 latest versions does not crash after 10 minutes with AVX2 work loads enabled.

Just got my CPU ………….

[FlanK3r]

it can be issue of software, notcpu real error

[mongoled]

Hopefully the OP will fix this issue

[mongoled]

Hopefully the OP will fix this issue

OP has fixed the issue

You can find version 1.1.2 works as expected

[mongoled]

Hi peeps!

Can someone confirm if this is the case on their Ryzen CPU?

Before Linpack Extreme used to generate heat that was equivalent to what Prime95 small FFTs would do.

The latest version 1.1.3 is not doing that anymore.

Prime95 small FFTs would generate maximum CPU temps of 94-95C.

Latest version of Linpack Xtreme is maxing out at temps of 63-64C.

Using the stress test tool GFlops are at 151-152.

While using the benchmark tool GFlops are at 294-297.

This is using the settings in my sig

Cheers