Stress testing
Running an overclocked PC is fine as long as it is stable and that the temperature of its components do not exceed their acceptable range. There are several programs available to assess system stability through stress testing the system and thereby the overclock level. The steps of overclocking a PC are beyond the scope of this article, but there is pretty inclusive guide written by graysky on the topic: Overclocking guide.
Stress testing software
This section lists stress testing software and classifies it by electrical power as high power or medium power.
Power | Program | Description |
---|---|---|
Medium | ||
Cc/Gcc | Both cc/gcc compilation is a great method of stress testing. Both are available in the base-devel group. | |
HandBrake-cli | handbrake-cli can be used to encode using high quality settings. | |
Systester | systesterAUR Systester is a multithreaded piece of software capable of deriving values of pi out to 128,000,000 decimal places. It has built in check for system stability. | |
Stressful Application Test | stressapptestAUR is a memory interface test. | |
High | stress | stress is a simple CPU, memory, I/O, and disk workload generator implemented in C. |
mprime | mprime-binAUR factors large numbers and is an excellent way to stress CPU and memory. | |
linpack | linpackAUR - Linpack makes use of the BLAS (Basic Linear Algebra Subprograms) libraries for performing basic vector and matrix operations. and is an excellent way to stress CPUs for stability. |
It is recommended to use programs in both categories to assess the overall system stability. It can happen that a system is more sensitive to a test from the medium than from the high demand category. Higher demand voltage programs require the most CPU core voltage (VCORE) due to intense hardware usage to perform their tasks. Medium demand voltage programs do not always call for the highest VCORE when running and as such can be more prone to throwing errors for systems that are undervolted relative to the clock speed requested.
Example on an overclocked i7-3770K (4.50 GHz); VCORE is +0.020 V in offset mode with all power-saving features enabled. This machine running with a VCORE of +0.005 in offset mode remains stable with both #MPrime and #Linpack for hours, but throws errors under both x264 and gcc after only several minutes:
Idle: 0.7440 V - 0.8320 V (varies). Mprime small FFTs: 1.2880 V (steady). Mprime large FFTs: 1.3040 V (steady). Mprime blend: 1.2960 V (steady). Linpack: 1.2320 V - 1.2720 V (varies). x264 encoding: 1.2320 V - 1.2720 V (varies). gcc compiling: 1.2720 V (steady).
Stressing CPU and Memory
stress
stress performs a loop that calculates the square root of a random number in order to stress the CPU. It can run simultaneously several workers to load all the cores of a CPU for example. It can also generate memory, I/O or disk workload depending on the parameters passed. The FAQ[dead link 2020-04-03 ⓘ] provides examples and explanations.
To spawn 4 workers spinning on sqrt(), use the command:
$ stress --cpu 4
MPrime
MPrime (also known as Prime95 in its Windows and MacOS implementation) is recognized universally as one defacto measure of system stability. MPrime under torture test mode will perform a series of very CPU intensive calculations and compare the values it gets to known good values.
The Linux implementation is called mprimeAUR and is available in the AUR.
To run mprime, simply open a shell and type "mprime"
$ mprime
When the software loads, simply answer 'N' to the first question to begin the torture testing:
Main Menu
1. Test/Primenet 2. Test/Worker threads 3. Test/Status 4. Test/Continue 5. Test/Exit 6. Advanced/Test 7. Advanced/Time 8. Advanced/P-1 9. Advanced/ECM 10. Advanced/Manual Communication 11. Advanced/Unreserve Exponent 12. Advanced/Quit Gimps 13. Options/CPU 14. Options/Preferences 15. Options/Torture Test 16. Options/Benchmark 17. Help/About 18. Help/About PrimeNet Server
There are several options for the torture test (menu option 15).
- Small FFTs (option 1) to stress the CPU
- In-place large FFTs (option 2) to test the CPU and memory controller
- Blend (option 3) is the default and constitutes a hybrid mode which stresses the CPU and RAM.
Errors will be reported should they occur both to stdout and to ~/results.txt
for review later. Many do not consider a system as 'stable' unless it can run the Large FFTs for a 24 hour period.
Example ~/results.txt
; note that the two runs from 26-June indicate a hardware failure. In this case, due to insufficient vcore to the CPU:
[Sun Jun 26 20:10:35 2011] FATAL ERROR: Rounding was 0.5, expected less than 0.4 Hardware failure detected, consult stress.txt file. FATAL ERROR: Rounding was 0.5, expected less than 0.4 Hardware failure detected, consult stress.txt file. [Sat Aug 20 10:50:45 2011] Self-test 480K passed! Self-test 480K passed! [Sat Aug 20 11:06:02 2011] Self-test 128K passed! Self-test 128K passed! [Sat Aug 20 11:22:10 2011] Self-test 560K passed! Self-test 560K passed! ...
Linpack
linpackAUR makes use of the BLAS (Basic Linear Algebra Subprograms) libraries for performing basic vector and matrix operations. It is an excellent way to stress CPUs for stability (only Intel CPUs are supported). After installation, users should copy /usr/share/linpack/linpack.conf
to ~/.config/linpack.conf
and adjust it according to the amount of memory on the system.
Systester (AKA SuperPi for Windows)
SystesterAUR is available in the AUR in both cli and gui version. It tests system stability by calculating up to 128 millions of Pi digits and includes error checking. Note that one can select from two different calculation algorithms: Quadratic Convergence of Borwein and Gauss-Legendre. The latter being the same method that the popular SuperPi for Windows uses.
A cli example using 8 threads is given:
$ systester-cli -gausslg 64M -threads 8
Intel Processor Diagnostic Tool
The Intel Processor Diagnostic Tool is a tool that verifies the functionality of an Intel Microprocessor by stress testing the CPU. A Fedora Linux LiveUSB ISO images are available[dead link 2020-04-03 ⓘ]. The LiveUSB image allows you to stress test your machine without using your main operating system; such method might be useful in extreme cases especially when dealing with cold reboots/crashes.
Burn the image to a USB stick by using dd or Gnome Disks and then boot the Live CD. Once booted, open the terminal and type the following command to install Intel Processor Diagnostic Tool for 64-bit machines:
$ install64
Once it is installed, you can run the Diagnostic Tool by clicking on the IPDT Icon that is located on the desktop.
Stressing memory
Use MemTest86 (proprietary) or Memtest86+ (GPL) to test your memory (RAM). There are "new" and "old" testers:
- "New" versions do not support BIOS. For a new version, use a proprietary MemTest86 version greater or equal to 8. Install it as memtest86-efiAUR or boot the Arch Linux install image.
- "Old" versions do not support UEFI nor DDR4. Old versions are available as GPL memtest86+ (development discontinued). It is roughly equal to proprietary MemTest86 version 4. After installation, update GRUB: it will auto-detect the package and allow users to boot directly to it.
- A reliable source of the version history is the history[dead link 2020-04-03 ⓘ] page in memtest86.com, in particular the section "MemTest86 and MemTest86+" and the following paragraph. Notice the proprietary MemTest86 from version 5 through 7 claims to support both BIOS and UEFI, but they simply bundle old and new versions.
- Allowing tests to run for at least 10 cycles without errors is usually sufficient.
Discovering Errors
Some stressing applications like #MPrime or #Linpack have built in consistency checks to discover errors due to non-matching results. A more general and simple method for measuring hardware instabilities can be found in the kernel itself. To use it, simply watch the output from the kernel ring buffer by this command:
# cat /proc/kmsg
The key error to watch for looks like this:
[Hardware Error]: Machine check events logged
The kernel can throw these errors while the stressing application is running, before it ends the calculation and reports the error, thus providing a very sensitive method to assess stability.