matszpk / clgpustress Goto Github PK
View Code? Open in Web Editor NEWHeavy OpenCL GPU stress tester
Home Page: http://clgpustress.nativeboinc.org
License: GNU General Public License v2.0
Heavy OpenCL GPU stress tester
Home Page: http://clgpustress.nativeboinc.org
License: GNU General Public License v2.0
Linux machine on amd64.
g++ --version
g++ (Debian 9.3.0-13) 9.3.0
$ make
g++ -Wall -std=gnu++11 -Os -fexpensive-optimizations -Wall -I`fltk-config --includedir` -c -o gpustress-gui.o gpustress-gui.cpp
In file included from gpustress-gui.cpp:52:
/usr/include/CL/cl.hpp:143:110: note: #pragma message: This version of the OpenCL Host API C++ bindings is deprecated, please use cl2.hpp instead.
143 | #pragma message("This version of the OpenCL Host API C++ bindings is deprecated, please use cl2.hpp instead.")
| ^
In file included from /usr/include/CL/cl.h:20,
from /usr/include/CL/opencl.h:24,
from /usr/include/CL/cl.hpp:167,
from gpustress-gui.cpp:52:
/usr/include/CL/cl_version.h:22:104: note: #pragma message: cl_version.h: CL_TARGET_OPENCL_VERSION is not defined. Defaulting to 220 (OpenCL 2.2)
22 | #pragma message("cl_version.h: CL_TARGET_OPENCL_VERSION is not defined. Defaulting to 220 (OpenCL 2.2)")
| ^
gpustress-gui.cpp: In static member function ‘static void TestConfigsGroup::selectedDeviceChanged(Fl_Widget*, void*)’:
gpustress-gui.cpp:897:72: error: cannot convert ‘__gnu_cxx::__alloc_traits<std::allocator<_cl_device_id*>, _cl_device_id*>::value_type’ {aka ‘_cl_device_id*’} to ‘const cl::Device&’
897 | t->allConfigsMap.find(t->choosenClDeviceIDs[index])->second);
| ^
gpustress-gui.cpp:766:57: note: initializing argument 1 of ‘void SingleTestConfigGroup::setConfig(const cl::Device&, const GPUStressConfig&)’
766 | void SingleTestConfigGroup::setConfig(const cl::Device& _clDevice,
| ~~~~~~~~~~~~~~~~~~^~~~~~~~~
make: *** [Makefile:28: gpustress-gui.o] Error 1
Hello,
I'm using gpustress-cli on two different servers with the same GPUs (RX 6700XT). On the first server everything work as expected, but on the second one I always get a segfault when interrupting gpustress-cli:
root@rescue:~# gpustress-cli -L 0:0 -w -S 3200
CLGPUStress CLI 0.0.9.4 by Mateusz Szpakowski ([email protected])
Program is distributed under terms of the GPLv2.
Website: http://clgpustress.nativeboinc.org.
Sources available at https://github.com/matszpk/clgpustress.
Binaries available at http://files.nativeboinc.org/offtopic/clgpustress/.
WARNING: THIS PROGRAM CAN OVERHEAT OR DAMAGE YOUR GRAPHICS CARD FASTER
(AND BETTER) THAN ANY FURMARK STRESS TEST. PLEASE USE THIS PROGRAM
VERY CAREFULLY!!!
WE RECOMMEND TO RUN THIS PROGRAM ON THE STOCK PARAMETERS OF THE DEVICES
(CLOCKS, VOLTAGES, ESPECIALLY MEMORY CLOCK).
TO TERMINATE THIS PROGRAM PLEASE USE STANDARD 'CTRL-C' KEY COMBINATION.
Preparing StressTester for
#0 AMD Accelerated Parallel Processing:gfx1031
SetUp: workSize=1310720, memory=320 MB, workFactor=256, blocksNum=2,
computeUnits=20, groupSize=256, passIters=3200, testType=0,
inputAndOutput=no
Calibrating Kernel for
#0 AMD Accelerated Parallel Processing:gfx1031...
Calibration progress: 0% 12% 25% 37% 50% 62% 75% 87% 100%
Kernel calibrated for
#0 AMD Accelerated Parallel Processing:gfx1031
BestKitersNum: 17, Bandwidth: 351.083 GB/s, Performance: 4476.31 GFLOPS
Program build log:
AMD Accelerated Parallel Processing:gfx1031
:--------------------
:--------------------
KernelTime: 0.000955741s, itersPerWait: 314
#0 Results for comparison has been generated.
#0 AMD Accelerated Parallel Processing:gfx1031 passed PASS #10
Approx. bandwidth: 277.021 GB/s, Approx. perf: 3532.02 GFLOPS, elapsed: 0:00:38.760
^CSegmentation fault
Linux kernel (5.15.55 mainline) is reporting the following error:
[Mon Aug 8 15:51:20 2022] signal: gpustress-cli[4892] overflowed sigaltstack
Reducing the number of GPUs with the -L
argument doesn't make a difference, the issue is always present.
The execution environment is strictly identical between the two servers, however the hardware other than the GPU model is different.
Any idea of why only one server is impacted and not the other ? Feel free to ask if you need some additional logs.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.