Thank you. In my OpenCL code in FindBestResultKernel, I am replacing that index field value in temporary variables by index of element in the array. However, final result is copy of element from source array, thus the index field will be kept.

I had to reset GPU performance stats listed here GPU speed because the script did not sum points correctly when the app runtime was very short and the app used less than 100s of CPU time. Mostly AMD cards running OpenCL apps were affected and I finally found the bug when I noticed RX Vega at the bottom with speed index of 0.

---

M4 Project homepage
M4 Project wiki

Will there be Intel gpu?

Will there be Intel gpu?

+1

Not sure how easy or difficult it is to port the AMD openCL app to Intel?

The current OpenCL version is highly optimized for AMD cards and won't work on Intel GPUs for sure.
Older version did work but it was painfully slow and I've encountered lots of random driver/app crashes, it crashed so often that I could not even verify if it returned correct results, so the support was dropped.

---

M4 Project homepage
M4 Project wiki

Ok, I understand, thank you for your reply

4 min 14 sec for gtx 960m with run_m4.bat

---

Is anyone else having issues with 1.10 (cuda_fermi) failing immediately on NVIDIA cards? All tasks fail within 2-3 seconds with 195 error/exit code.

Is anyone else having issues with 1.10 (cuda_fermi) failing immediately on NVIDIA cards? All tasks fail within 2-3 seconds with 195 error/exit code.

What is the error message?

On Windows 7 my card runs fine.

I had a short problem and all tasks failed with:

<core_client_version>7.6.33</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code 195 (0xc3)
</message>
<stderr_txt>
wrapper [26118 765 +dev#]: starting (GPU #0)
17:49:16 (1608): wrapper: running enigma_cuda.exe (-d 0 -R -o results.txt 04trigr.cur 04unigr.cur 04ciphertext)
wrapper [26118 765 +dev#]: starting (GPU #0)
17:53:26 (2200): wrapper: running enigma_cuda.exe (-d 0 -R -o results.txt 04trigr.cur 04unigr.cur 04ciphertext)
CUDA error: C:/Proj/Other/Enigma/enigma-cuda/enigma-cuda-lib/cuda_code.cu:784:  initialization errorInitialization failed: CUDA error: C:/Proj/Other/Enigma/enigma-cuda/enigma-cuda-lib/cuda_code.cu:784:  initialization error
app exit status: 0x2
17:53:32 (2200): called boinc_finish

</stderr_txt>
]]>

But after a restart everything seems fine now.

<core_client_version>7.8.2</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code 195 (0xc3)
</message>
<stderr_txt>
wrapper [26118 765 +dev#]: starting (GPU #0)
14:48:45 (2208): wrapper: running enigma_cuda.exe (-d 0 -R -o results.txt 04trigr.cur 04unigr.cur 04ciphertext)
app exit status: 0xc0000135
14:48:46 (2208): called boinc_finish

</stderr_txt>
]]>

Is the error output I am getting. I tried going back to windows 7, but to the same result. Perhaps it is a driver issue? Which one are you using?

If it is any help my AMD cards run fine.

NVIDIA GeForce GTX 1080 (4095MB) driver: 385.41 OpenCL: 1.2

Microsoft Windows 7
Home Premium x64 Edition, Service Pack 1, (06.01.7601.00)

Interesting. We have almost the same setup except for BOINC version and driver version. I am currently using 385.69. I will roll back to your version and see if that helps.

Do you have the visual c++ redistributable installed ?
It is the most common thing that causes all the tasks to fail with error 195.

---

M4 Project homepage
M4 Project wiki

That was the issue. Thanks TJM!

I forgot the redistributable comes with the AMD drivers, but not the NVIDIA.

Hi :)

Both cuda and opencl versions work fine on my machines, the only errors reported are due to drivers losing contact during the same night on both PCs with GTXs (dear Murphy :)).

---

Download and install Visual Studio 2015 c++ redistributable package.

Need to download the 32 bit version. 64 bit doesn't do the trick.

W10, BOINC 7.8.3, Visual C 2015 32 BIT installed (and 64 bit)

nVidia NVS4200M - 512mb - driver 385.08

Once I figured out we need the 32 bit Runtime and not the 64 bit it runs fine on my ASUS ROG notebook and didn't crash after 2 seconds on my Dell here. One died after 6 seconds and the other after 13 seconds. I was so hopeful when it ran more than 2 seconds... SIGH

Compute capability shows as 2.1 / OpenCL: 1.1

(unknown error) - exit code 195 (0xc3)</message>
<stderr_txt>
wrapper [26118 765 +dev#]: starting (GPU #0)
19:13:27 (7212): wrapper: running enigma_cuda.exe (-d 0 -R -o results.txt 04trigr.cur 04unigr.cur 04ciphertext)
CUDA error: C:/Proj/Other/Enigma/enigma-cuda/enigma-cuda-lib/cuda_code.cu:663:  unspecified launch failureRuntime error: CUDA error: C:/Proj/Other/Enigma/enigma-cuda/enigma-cuda-lib/cuda_code.cu:663:  unspecified launch failure
app exit status: 0x3

I've gotten 2 tasks that have gone over the typical times and are still running although at a 100%. I aborted one and looks like I'll have to abort another one.

Somehow I´m not getting GPU workunits.
"No tasks available for GPU", as stated via the manager log.
Nearly 2K WU´s are unsent.

"Run test applications" in Your preferences. It helped in my case.