Make babelstream runnable and test for different Accs #2299

mehmetyusufoglu · 2024-06-24T13:59:09Z

babelstream benchmarking was creating errors and had unnecessary parts (because it was developed as an extention of the test at Bristol Uni. )

Unnecesssary command line options are removed, 2 command line custom arguments are available

Run with Custom arguments:
./babelstream --array-size=1280000 --number-runs=10
Run with Catch2 arguments:
./babelstream --success
./babelstream -r a.xml
Run with Custom and catch2 arguments together:
./babelstream --success --array-size=1280000 --number-runs=10
Run to list custom and catch2 arguments:
./babelstream --help (or -h or -?)

Babelstream has 5 simple kernels for vector operations. Add, Mul, Copy, Triad and Dot product. And there is an Init kernel.
RunTimes are measured by using std::chrono
RESULTS for ./babelstream --array-size=33554432 --number-runs=100 are below. They are equal ( or slightly better) compared to previous implementation which runs 100 times and uses minimum time as well.


AcceleratorType:AccCpuSerial<1,unsigned int>
NumberOfRuns:100
Precision:double
DataSize(items):33554432
DeviceName:13th Gen Intel(R) Core(TM) i7-1360P
WorkDivInit :{gridBlockExtent: (33554432), blockThreadExtent: (1), threadElemExtent: (1)}
WorkDivCopy :{gridBlockExtent: (33554432), blockThreadExtent: (1), threadElemExtent: (1)}
WorkDivMult :{gridBlockExtent: (33554432), blockThreadExtent: (1), threadElemExtent: (1)}
WorkDivAdd  :{gridBlockExtent: (33554432), blockThreadExtent: (1), threadElemExtent: (1)}
WorkDivTriad:{gridBlockExtent: (33554432), blockThreadExtent: (1), threadElemExtent: (1)}
Kernels         Bandwidths(GB/s) MinTime(s) MaxTime(s) AvgTime(s) DataUsage(MB) 
 InitKernel      12.626          0.042521 0.044909 0.045561  536.87 
 CopyKernel      21.185          0.025342 0.027276 0.025485  536.87 
 MultKernel      21.357          0.025138 0.026368 0.025286  536.87 
 AddKernel       25.272          0.031865 0.041926 0.032137  805.31 
 TriadKernel     25.174          0.03199  0.033468 0.032193  805.31 



AcceleratorType:AccGpuCudaRt<1,unsigned int>
NumberOfRuns:100
Precision:double
DataSize(items):33554432
DeviceName:NVIDIA RTX A500 Laptop GPU
WorkDivInit :{gridBlockExtent: (32768), blockThreadExtent: (1024), threadElemExtent: (1)}
WorkDivCopy :{gridBlockExtent: (32768), blockThreadExtent: (1024), threadElemExtent: (1)}
WorkDivMult :{gridBlockExtent: (32768), blockThreadExtent: (1024), threadElemExtent: (1)}
WorkDivAdd  :{gridBlockExtent: (32768), blockThreadExtent: (1024), threadElemExtent: (1)}
WorkDivTriad:{gridBlockExtent: (32768), blockThreadExtent: (1024), threadElemExtent: (1)}
WorkDivDot  :{gridBlockExtent: (256), blockThreadExtent: (1024), threadElemExtent: (1)}
Kernels         Bandwidths(GB/s) MinTime(s) MaxTime(s) AvgTime(s) DataUsage(MB) 
 InitKernel      62.451          0.0085967 0.0086461 0.0086267  536.87 
 CopyKernel      89.58           0.0059932 0.0060739 0.0060465  536.87 
 MultKernel      89.359          0.006008 0.0060836 0.0060491  536.87 
 AddKernel       90.551          0.0088934 0.0089586 0.008937  805.31 
 TriadKernel     90.61           0.0088876 0.0089629 0.0089369  805.31 
 DotKernel       93.256          0.0057569 0.0058079 0.0057916  536.87

Full output file:

babelStreamOut.txt

Old implementation values for 4 kernels 100 runs in the following image:

SimeonEhrig · 2024-07-01T12:51:59Z

benchmarks/babelstream/src/benchmarkTest.cpp

+    {
+        dataTypeStr = "double-precision";
+    }
+    std::cout << "Data type: " << typeid(DataType).name() << " " << dataTypeStr << std::endl;


Do you know, if there is some reporting function to add the information as key value pair to the output? At the moment, it simply captures the stdout and write a string to the xml, which make it hard to parse.

<?xml version="1.0" encoding="UTF-8"?> <Catch2TestRun name="babelstream" rng-seed="1921823939" xml-format-version="2" catch2-version="3.3.2"> <TestCase name="Babelstream Bechmark by Alpaka Kernel - Single P. - TestAccs1D - 0" tags="[benchmark-test]" filename="/home/simeon/projects/alpaka/benchmarks/babelstream/src/benchmarkTest.cpp" line="418"> <BenchmarkResults name="CopyKernel" samples="100" resamples="100000" iterations="1" clockResolution="32.5898" estimatedDuration="1.62794e+07">  <mean value="162549" lowerBound="162235" upperBound="163277" ci="0.95"/> <standardDeviation value="2305.52" lowerBound="1253.62" upperBound="4617.02" ci="0.95"/> <outliers variance="0.0750291" lowMild="0" lowSevere="0" highMild="1" highSevere="8"/> </BenchmarkResults> <BenchmarkResults name="AddKernel" samples="100" resamples="100000" iterations="1" clockResolution="32.5898" estimatedDuration="2.07714e+07">  <mean value="208480" lowerBound="207810" upperBound="209855" ci="0.95"/> <standardDeviation value="4688.39" lowerBound="2825.43" upperBound="8760.11" ci="0.95"/> <outliers variance="0.161071" lowMild="0" lowSevere="0" highMild="1" highSevere="6"/> </BenchmarkResults> <BenchmarkResults name="MulKernel" samples="100" resamples="100000" iterations="1" clockResolution="32.5898" estimatedDuration="1.19507e+07">  <mean value="122044" lowerBound="121594" upperBound="122670" ci="0.95"/> <standardDeviation value="2676.42" lowerBound="2061.12" upperBound="3957.46" ci="0.95"/> <outliers variance="0.151683" lowMild="0" lowSevere="0" highMild="1" highSevere="1"/> </BenchmarkResults> <BenchmarkResults name="TriadKernel" samples="100" resamples="100000" iterations="1" clockResolution="32.5898" estimatedDuration="2.11882e+07">  <mean value="212046" lowerBound="211493" upperBound="212948" ci="0.95"/> <standardDeviation value="3544.51" lowerBound="2546.6" upperBound="5707.04" ci="0.95"/> <outliers variance="0.094408" lowMild="0" lowSevere="0" highMild="1" highSevere="9"/> </BenchmarkResults> <OverallResult success="true" skips="0"> <StdOut> ======================== Time:2024-07-01 14:21:45 Using alpaka accelerator: AccCpuSerial<1,unsigned int> Data type: f single-precision Data size: 1048576 elements Work division: {gridBlockExtent: (1048576), blockThreadExtent: (1), threadElemExtent: (1)} </StdOut> </OverallResult> </TestCase>

Yes I will use "INFO" macro of catch2. Good idea thanks. It turned out to be -r json also works for catch2 3.5.2. I am doing it for json, more readable.

json reporter seem to not include benchmark results only records REQUIRE results etc. XML reporter prints benchmark results as expected.

Only WARN has worked to log output other than using cout. I will use WARN. Because others only workd when there is REQUIRE/CHECK and it fails.

TEST_CASE("TESTCASE tesing trial", "[testcase-test]") { std::string dataType = "floating-point"; std::string dataSize = std::to_string(100000); int arraySize=1234; WARN(dataType); // This will be included as custom XML in the report INFO("\n normal info Data type is" << dataType); UNSCOPED_INFO("unscoped Data type is" << dataType); INFO("info example"); UNSCOPED_INFO("unscoped example"); INFO("normal info array size is" << arraySize); UNSCOPED_INFO("unscoped array size is" << arraySize); INFO("normal info data size is" << dataSize); UNSCOPED_INFO("unscoped data size is" << dataSize); CAPTURE(dataSize); CAPTURE(dataType); // !!!! CHECK(arraySize!=1234); // If this fails all INFO, capture and UNSCOPED_INFO prints. Otherwise only WARN prints SUCCEED("SUCCEED"); }

This will need additional operation to extract data from fields. May be I should use cout output directly and convert it to json by a script instead of dealing with this...

<Warning filename="/home/yusufo81/projects/alpaka-dir/alpaka/benchmarks/babelstream/src/benchmarkTest.cpp" line="146"> TimeStamp:2024-07-09 15:06:59 </Warning> <Warning filename="/home/yusufo81/projects/alpaka-dir/alpaka/benchmarks/babelstream/src/benchmarkTest.cpp" line="147"> DataSize:1048576 </Warning> <Warning filename="/home/yusufo81/projects/alpaka-dir/alpaka/benchmarks/babelstream/src/benchmarkTest.cpp" line="148"> DataSize:single-precision </Warning>

SimeonEhrig · 2024-07-01T12:52:32Z

benchmarks/babelstream/src/benchmarkTest.cpp

+        }
+    }
+
+    std::string getCurrentTimestamp();


Why the forward declaration?

There was a warning at one compiler "warning: no previous declaration for ‘std::string getCurrentTimestamp()’ [-Wmissing-declarations]" therefore i added a declaration. Alternatively, I can make the function static as well. Somehow one of the compilers forced for declaration.

I made the function static. Like functions in some other tests: e.g. alpaka/test/unit/mem/buf/src/BufTest.cpp

benchmarks/babelstream/src/benchmarkTest.cpp

psychocoderHPC · 2024-09-03T13:57:26Z

benchmarks/babelstream/src/babelStreamCommon.hpp

@@ -0,0 +1,433 @@
+#pragma once
+#include <boost/tokenizer.hpp>


I am not sure if this is part of boost compile time header.
If not this will introduce additional dependencies for alpaka, at least benchmarking.
Please remove the include and use best stl to split strings, I assume this is the reason why it is included.

The CI is maybe passing but the reason is only that all boost packages are installed. I tried it on our development system HAL

spack load [email protected] %[email protected] \ ^[email protected] \ ^[email protected] \ ^[email protected] +atomics +cuda cuda_arch=80 \ ^[email protected] \ ^[email protected] spack load [email protected] %[email protected]

error:

fatal error: boost/tokenizer.hpp: No such file or directory

yes i used stl removed boost tokenizer after seeing the CI fail. Now removed the header too.

I tested it, the include can be removed because it is not used

removed yes, i found CI error in the morning and removed boost tokenizer usage; but forgot tto remove header. now removed.

psychocoderHPC · 2024-09-03T14:06:29Z

@mehmetyusufoglu The CMake file for the benchmark is missing to include catch in case you compile without -DBUILD_TESTING=ON and only with -Dalpaka_BUILD_BENCHMARKS=ON compiling fails due to missing catch2 headers.

psychocoderHPC · 2024-09-03T14:08:40Z

~~The output on the terminal should looks like the old output. ~~
~~This helps external users if they compare this bable stream against other upstream implementations.~~
~~The user should not search in the code if we use min/max or average for the results.~~

Sry my fold, the output is nearly equal, only MB is now GB/s

psychocoderHPC · 2024-09-03T14:12:49Z

The output on the terminal should looks like the old output. This helps external users if they compare this bable stream against other upstream implementations. The user should not search in the code if we use min/max or average for the results.

Yes it is the same output. Additionally, i can add that we used minimum time as an extra comment. ok.

I updated my cmment above.

i can add that we used minimum time as an extra comment. ok.

Is this different compared too the original code?

If not not need to add the comment
If yes, it should show us ethe same as the old code

mehmetyusufoglu · 2024-09-03T14:13:41Z

The output on the terminal should looks like the old output. This helps external users if they compare this bable stream against other upstream implementations. The user should not search in the code if we use min/max or average for the results.

Yes it is the same output. Additionally, i can add that we used minimum time as an extra comment. ok.

I updated my cmment above.

i can add that we used minimum time as an extra comment. ok.

Is this different compared too the original code? If not no need to add the comment

No, same as original. It was not stated in the output. Therefore i did not mention. i will not state in the output as well so not change the code.

mehmetyusufoglu · 2024-09-03T14:27:17Z

@mehmetyusufoglu The CMake file for the benchmark is missing to include catch in case you compile without -DBUILD_TESTING=ON and only with -Dalpaka_BUILD_BENCHMARKS=ON compiling fails due to missing catch2 headers.

I am investigating but that was not creating a problem in CI. Here everything passed. https://gitlab.com/hzdr/crp/alpaka/-/commit/472bfed5d7045b6a94b04e9913b252ce1fd5e3dc
Locally no problem observed.

mehmetyusufoglu · 2024-09-05T20:07:28Z

The output on the terminal should looks like the old output. This helps external users if they compare this bable stream against other upstream implementations. The user should not search in the code if we use min/max or average for the results.

Yes it is the same output. Additionally, i can add that we used minimum time as an extra comment. ok.

I updated my cmment above.

i can add that we used minimum time as an extra comment. ok.

Is this different compared too the original code?
* If not not need to add the comment

* If yes, it should show us ethe same as the old code

It is the same with the previous code's output. Besides there are 4 additional fields as info: 1. Acc type, 2. data item size, 3. workdiv and 4. correct data-throughput for each kernel.

mehmetyusufoglu · 2024-09-05T20:12:54Z

@mehmetyusufoglu The CMake file for the benchmark is missing to include catch in case you compile without -DBUILD_TESTING=ON and only with -Dalpaka_BUILD_BENCHMARKS=ON compiling fails due to missing catch2 headers.
* I am investigating but that was not creating a problem in CI. Here everything passed. https://gitlab.com/hzdr/crp/alpaka/-/commit/472bfed5d7045b6a94b04e9913b252ce1fd5e3dc

* Locally no problem observed.

All tests pass at gitlab and github. No problem.

psychocoderHPC · 2024-09-10T09:52:22Z

@mehmetyusufoglu please amend and force push this PR again, the CI is fixed

psychocoderHPC · 2024-09-11T07:16:40Z

-Dalpaka_BUILD_BENCHMARKS=ON

The problem still exists. As I wrote if you only enable benchmarking catch will not found if you not provide an installed catch version externally.

cmake ~/workspace/alpaka -Dalpaka_ACC_GPU_CUDA_ENABLE=ON -Dalpaka_ACC_GPU_CUDA_ONLY_MODE=ON -Dalpaka_BUILD_BENCHMARKS=ON

alpaka/benchmarks/babelstream/src/babelStreamMainTest.cpp:3:10: fatal error: catch2/catch_session.hpp: No such file or directory
    3 | #include "catch2/catch_session.hpp"
      |          ^~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make[2]: *** [benchmarks/babelstream/CMakeFiles/babelstream.dir/build.make:77: benchmarks/babelstream/CMakeFiles/babelstream.dir/src/babelStreamMainTest.cpp.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:158: benchmarks/babelstream/CMakeFiles/babelstream.dir/all] Error 2

psychocoderHPC

#2299 (comment) must be fixed before we can merge this PR.
This error will not show up in the CI because we enable tests and benchmarking in the CI always together.

mehmetyusufoglu · 2024-09-11T13:11:49Z

#2299 (comment) must be fixed before we can merge this PR. This error will not show up in the CI because we enable tests and benchmarking in the CI always together.

@SimeonEhrig my latest commit would not solve the issue i think. What is your opinion?

SimeonEhrig · 2024-09-12T11:12:31Z

#2299 (comment) must be fixed before we can merge this PR. This error will not show up in the CI because we enable tests and benchmarking in the CI always together.

@SimeonEhrig my latest commit would not solve the issue i think. What is your opinion?

Something get broken. If I run cmake -Dalpaka_BUILD_BENCHMARKS=ON or cmake -DBUILD_TESTING=ON I get the following error message:

CMake Error at thirdParty/CMakeLists.txt:42 (find_package):
  Could not find a package configuration file provided by "Catch2" (requested
  version 3.5.2) with any of the following names:

    Catch2Config.cmake
    catch2-config.cmake

  Add the installation prefix of "Catch2" to CMAKE_PREFIX_PATH or set
  "Catch2_DIR" to a directory containing one of the above files.  If "Catch2"
  provides a separate development package or SDK, be sure it has been
  installed.

This CMake path should be used, if the CMake variable alpaka_USE_INTERNAL_CATCH2 is OFF. But setting -Dalpaka_USE_INTERNAL_CATCH2=ON also does not solve the problem.

psychocoderHPC · 2024-09-12T13:26:07Z

CMakeLists.txt

 cmake_dependent_option(alpaka_USE_INTERNAL_CATCH2 "Use internally shipped Catch2" ON BUILD_TESTING OFF)
+cmake_dependent_option(alpaka_USE_INTERNAL_CATCH2 "Use internally shipped Catch2" ON alpaka_BUILD_BENCHMARKS OFF)


I think you must merge this line and the line above see https://cmake.org/cmake/help/latest/module/CMakeDependentOption.html

If you add two lines the last line will have the highest priority.

psychocoderHPC · 2024-09-12T13:26:18Z

CMakeLists.txt

@@ -48,7 +48,9 @@ option(alpaka_INSTALL_TEST_HEADER "Install headers of the namespace alpaka::test

 include(CMakeDependentOption)
 cmake_dependent_option(alpaka_CHECK_HEADERS "Check all alpaka headers as part of the tests whether they can be compiled standalone." OFF BUILD_TESTING OFF)
+cmake_dependent_option(alpaka_CHECK_HEADERS "Check all alpaka headers as part of the benchmarks whether they can be compiled standalone." OFF alpaka_BUILD_BENCHMARKS OFF)


I think you must merge this line and the line above see https://cmake.org/cmake/help/latest/module/CMakeDependentOption.html

If you add two lines the last line will have the highest priority.

I have added an additional local variable as (alpaka_BUILD_BENCHMARKS OR BUILD_TESTING).

But; Benchmarks were locally not compilable and i got the error,
/home/yusufo81/projects/alpaka-dir/alpaka/benchmarks/babelstream/src/babelStreamMainTest.cpp:5:10: fatal error: alpaka/alpaka.hpp: No such file or directory 5 | #include <alpaka/alpaka.hpp> | ^~~~~~~~~~~~~~~~~~~

to solve this i had to change the condition to call add_subdirectory("test/") at line 167. But with this change, setting benchmark cmake variable became the same thing of setting BUILD_TESTING. It is a kind of work-around.
\# Only build the tests if alpaka is the top-level project and BUILD_TESTING or alpaka_BUILD_BENCHMARKS is ON \# Note: alpaka_BUILD_BENCHMARKS is added with an OR besides BUILD_TESTING below. \# Because the benchmark which uses catch2 is not compilable locally without adding test directory as below \# or without setting BUILD_TESTING cmake variable to ON. See PR2299.

if(CMAKE_PROJECT_NAME STREQUAL PROJECT_NAME AND (BUILD_TESTING OR alpaka_BUILD_BENCHMARKS)) add_subdirectory("test/") endif()

psychocoderHPC · 2024-09-16T09:51:39Z

CMakeLists.txt

+# Note: alpaka_BUILD_BENCHMARKS is added with an OR besides BUILD_TESTING below.
+# Because the benchmark which uses catch2 is not compilable locally without adding test directory as below
+# or without setting BUILD_TESTING cmake variable to ON. See PR2299.
+if(CMAKE_PROJECT_NAME STREQUAL PROJECT_NAME AND (BUILD_TESTING OR alpaka_BUILD_BENCHMARKS))


we do not want to build testing if benchmark only is enabled.

psychocoderHPC · 2024-09-16T09:52:23Z

I will provide a fix after lunch!

Co-authored-by: René Widera <[email protected]> Co-authored-by: Simeon Ehrig <[email protected]>

mehmetyusufoglu marked this pull request as draft June 24, 2024 13:59

mehmetyusufoglu force-pushed the updateBenchmarkBabelStr branch 11 times, most recently from 1965c6b to b38be37 Compare June 27, 2024 14:26

mehmetyusufoglu changed the title ~~[Wip] Make babelstream runnable and use Catch2 advanced benchmarking~~ Make babelstream runnable and use Catch2 advanced benchmarking Jun 27, 2024

mehmetyusufoglu marked this pull request as ready for review June 27, 2024 15:00

mehmetyusufoglu marked this pull request as draft June 27, 2024 15:00

mehmetyusufoglu marked this pull request as ready for review June 27, 2024 15:48

mehmetyusufoglu force-pushed the updateBenchmarkBabelStr branch 2 times, most recently from 6ffd219 to 0a5f6bd Compare June 28, 2024 12:27

SimeonEhrig reviewed Jul 1, 2024

View reviewed changes

mehmetyusufoglu force-pushed the updateBenchmarkBabelStr branch 11 times, most recently from 897c296 to 97286cc Compare July 10, 2024 11:05

mehmetyusufoglu force-pushed the updateBenchmarkBabelStr branch from 64372e3 to 6f755f1 Compare September 3, 2024 12:10

psychocoderHPC requested changes Sep 3, 2024

View reviewed changes

mehmetyusufoglu force-pushed the updateBenchmarkBabelStr branch from 6f755f1 to 846856a Compare September 3, 2024 14:07

mehmetyusufoglu force-pushed the updateBenchmarkBabelStr branch from 846856a to d088742 Compare September 3, 2024 16:15

mehmetyusufoglu closed this Sep 5, 2024

mehmetyusufoglu reopened this Sep 5, 2024

psychocoderHPC requested changes Sep 11, 2024

View reviewed changes

mehmetyusufoglu force-pushed the updateBenchmarkBabelStr branch 4 times, most recently from 8fde9e5 to 93fccb0 Compare September 12, 2024 11:07

psychocoderHPC requested changes Sep 12, 2024

View reviewed changes

psychocoderHPC reviewed Sep 16, 2024

View reviewed changes

mehmetyusufoglu force-pushed the updateBenchmarkBabelStr branch from 57a3e98 to 3057e10 Compare September 16, 2024 10:58

Babelstream benchmark activation and refactoring

1840791

Co-authored-by: René Widera <[email protected]> Co-authored-by: Simeon Ehrig <[email protected]>

mehmetyusufoglu force-pushed the updateBenchmarkBabelStr branch from 3057e10 to 1840791 Compare September 16, 2024 17:27

psychocoderHPC approved these changes Sep 17, 2024

View reviewed changes

psychocoderHPC merged commit a017c60 into alpaka-group:develop Sep 17, 2024
22 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Make babelstream runnable and test for different Accs #2299

Make babelstream runnable and test for different Accs #2299

mehmetyusufoglu commented Jun 24, 2024 •

edited

Loading

SimeonEhrig Jul 1, 2024

mehmetyusufoglu Jul 8, 2024 •

edited

Loading

mehmetyusufoglu Jul 8, 2024

mehmetyusufoglu Jul 9, 2024 •

edited

Loading

mehmetyusufoglu Jul 9, 2024

SimeonEhrig Jul 1, 2024

mehmetyusufoglu Jul 3, 2024

mehmetyusufoglu Jul 3, 2024

psychocoderHPC Sep 3, 2024

psychocoderHPC Sep 3, 2024

mehmetyusufoglu Sep 3, 2024 •

edited

Loading

psychocoderHPC Sep 3, 2024

mehmetyusufoglu Sep 3, 2024

psychocoderHPC commented Sep 3, 2024

psychocoderHPC commented Sep 3, 2024 •

edited

Loading

psychocoderHPC commented Sep 3, 2024 •

edited

Loading

mehmetyusufoglu commented Sep 3, 2024 •

edited

Loading

mehmetyusufoglu commented Sep 3, 2024

mehmetyusufoglu commented Sep 5, 2024 •

edited

Loading

mehmetyusufoglu commented Sep 5, 2024

psychocoderHPC commented Sep 10, 2024

psychocoderHPC commented Sep 11, 2024 •

edited

Loading

psychocoderHPC left a comment

mehmetyusufoglu commented Sep 11, 2024

SimeonEhrig commented Sep 12, 2024

psychocoderHPC Sep 12, 2024

psychocoderHPC Sep 12, 2024

mehmetyusufoglu Sep 12, 2024 •

edited

Loading

psychocoderHPC Sep 16, 2024

psychocoderHPC commented Sep 16, 2024

		cmake_dependent_option(alpaka_USE_INTERNAL_CATCH2 "Use internally shipped Catch2" ON BUILD_TESTING OFF)
		cmake_dependent_option(alpaka_USE_INTERNAL_CATCH2 "Use internally shipped Catch2" ON alpaka_BUILD_BENCHMARKS OFF)

Make babelstream runnable and test for different Accs #2299

Make babelstream runnable and test for different Accs #2299

Conversation

mehmetyusufoglu commented Jun 24, 2024 • edited Loading

Choose a reason for hiding this comment

mehmetyusufoglu Jul 8, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mehmetyusufoglu Jul 9, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mehmetyusufoglu Sep 3, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

psychocoderHPC commented Sep 3, 2024

psychocoderHPC commented Sep 3, 2024 • edited Loading

psychocoderHPC commented Sep 3, 2024 • edited Loading

mehmetyusufoglu commented Sep 3, 2024 • edited Loading

mehmetyusufoglu commented Sep 3, 2024

mehmetyusufoglu commented Sep 5, 2024 • edited Loading

mehmetyusufoglu commented Sep 5, 2024

psychocoderHPC commented Sep 10, 2024

psychocoderHPC commented Sep 11, 2024 • edited Loading

psychocoderHPC left a comment

Choose a reason for hiding this comment

mehmetyusufoglu commented Sep 11, 2024

SimeonEhrig commented Sep 12, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mehmetyusufoglu Sep 12, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

psychocoderHPC commented Sep 16, 2024

mehmetyusufoglu commented Jun 24, 2024 •

edited

Loading

mehmetyusufoglu Jul 8, 2024 •

edited

Loading

mehmetyusufoglu Jul 9, 2024 •

edited

Loading

mehmetyusufoglu Sep 3, 2024 •

edited

Loading

psychocoderHPC commented Sep 3, 2024 •

edited

Loading

psychocoderHPC commented Sep 3, 2024 •

edited

Loading

mehmetyusufoglu commented Sep 3, 2024 •

edited

Loading

mehmetyusufoglu commented Sep 5, 2024 •

edited

Loading

psychocoderHPC commented Sep 11, 2024 •

edited

Loading

mehmetyusufoglu Sep 12, 2024 •

edited

Loading