CUDA error 'an illegal memory access was encountered'

[nzaar9]

Error in latest release 5c

Setting log level to 2
[21:20:44][0x00001630] Using SSE2: YES
[21:20:44][0x00001630] Using AVX: YES
[21:20:44][0x00001630] Using AVX2: YES
[21:20:44][0x00001630] Benchmarking CUDA worker (CUDA-DJEZO) GeForce 840M (#0) M=2
[21:20:45][0x00001630] Benchmark starting... this may take several minutes, please wait...
[21:20:45][0x0000150c] CUDA error 'an illegal memory access was encountered' in func 'eq_cuda_context<RB, SM, SSM, THREADS, PACKER>::solve' line 2095

[ImrieA]

I get the same error on both Windows and Ubuntu.
3x NVIDIA GTX 750ti rig

[unsalted]

Same on 1x GTX 1070 -- after about 2hrs

[kleine1979]

I have the same problem in my GTX 750ti. Instantly.

[betobmx]

same problem on gtx 750 ti, any solution

[fishouttah20]

same on gtx 750 ti

[VascoVisser]

same here, also gtx 750 ti

[VascoVisser]

CUDA_TROMP works for me. In CMakeLists.txt set the cuda tromp option to true and recompile.

[astronaute77]

Can someone please provide the recompiled file?
I'm also getting the same message :/

[kulla37]

VacscoVisser, do you mean to disable all others and leave only CUDA_TROMP?

[nelson6340]

Yeahhh, this version: nheqminer_v0.4b.zip works for me, on GTX 750 ti

[zhumingyu]

@VascoVisser hey,guy!how to modify the code?thx

[zhumingyu]

@S74nk0 hey,pls how to modify the code?thx

[AR0x7E7]

push same error here on a

• #0 Quadro K1200 | SM version: 5.0 | SM count: 4
• Win 7 64 bit
• CUDA 8.0.61 + 8.0.61.2

[locmer]

Running into this too GTX 750 ti 2 Gb with Mac OS X release 5c

[zammildin]

Am I? Im sorry....how to get rid of CUDA error. Pls advise...Im just trying to mine...(newbies) Sorry again...

…

On 29 October 2017 at 04:10, locmer ***@***.***> wrote: Running into this too GTX 750 ti 2 Gb with Mac OS X release 5c — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#223 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AfmeEXWQqivJFdcO9-nG0wDdxrGUzXD1ks5sw4o0gaJpZM4LjKVr> .

[loven-doo]

I have fixed this bug!!!
enable all lines "__syncthreads();" in cuda_djezo/equi_miner.cu
and recompile it
Works 10 times faster than cuda_tromp