cudaError 78

Question

cudaError 78

Opened this issue 7 years ago · 6 comments

I don't work Cudaminer

cudaminer -o stratum+tcp://jp.lapool.me:3015 -u okada.a -p 0908
*** CudaMiner for nVidia GPUs by Christian Buchner ***
This is version 2014-02-28 (beta)
based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler
Cuda additions Copyright 2013,2014 Christian Buchner
LTC donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm
BTC donation address: 16hJF5mceSojnTD3ZTUDqdRhDyPJzoRakM
YAC donation address: Y87sptDEcpLkLeAuex6qZioDbvy1qXZEj4
[2017-09-18 19:59:08] NVML GPU monitoring is not available.
[2017-09-18 19:59:08] 1 miner threads started, using 'scrypt' algorithm.
[2017-09-18 19:59:08] Starting Stratum on stratum+tcp://jp.lapool.me:3015
[2017-09-18 19:59:09] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaFree(0)' (salsa_kernel.cu line 242)

[2017-09-18 19:59:09] GPU #0: GeForce GTX 1060 6GB with compute capability 6.1
[2017-09-18 19:59:09] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaDeviceSetCacheConfig(kernel->cache_config())' (salsa_kernel.cu line 431)

[2017-09-18 19:59:09] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaDeviceSetSharedMemConfig(kernel->shared_mem_config())' (salsa_kernel.cu line 432)

[2017-09-18 19:59:09] GPU #0: interactive: 0, tex-cache: 0 , single-alloc: 0
[2017-09-18 19:59:09] GPU #0: 32 hashes / 4.0 MB per warp.
[2017-09-18 19:59:10] GPU #0: Performing auto-tuning (Patience...)
[2017-09-18 19:59:10] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaMalloc((void **) &d_idata, mem_size)' (salsa_kernel.cu line 609)

[2017-09-18 19:59:10] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaMalloc((void **) &d_odata, mem_size)' (salsa_kernel.cu line 610)

[2017-09-18 19:59:11] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaMemcpy(d_idata, h_idata, mem_size, cudaMemcpyHostToDevice)' (salsa_kernel.cu line 615)

[2017-09-18 19:59:11] GPU #0: maximum total warps (BxW): 0
[2017-09-18 19:59:11] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaFree(d_odata)' (salsa_kernel.cu line 745)

[2017-09-18 19:59:11] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaFree(d_idata)' (salsa_kernel.cu line 746)

[2017-09-18 19:59:11] GPU #0: 0.00 hash/s with configuration T0x0
[2017-09-18 19:59:11] GPU #0: using launch configuration T0x0
[2017-09-18 19:59:12] workio thread dead, waiting for workers...
[2017-09-18 19:59:12] worker threads all shut down, exiting.

Any thoughts?

Answer 1 · 2017-09-18T11:20:52.000Z

Someone reported a similar error here https://bitcointalk.org/index.php?topic=2161577.0 I would suggest to add the correct -gencode arch=compute_XX,code=sm_XX for your specific GPU model to the Makefile (Makefile.am?) In your case probably -gencode arch=compute_60,code=sm_61 or -gencode arch=compute_61,code=sm_61 - not sure which one is better. Also be sure to use CUDA 8.0 with Pascal support if it's available for your platform. Christian 2017-09-18 13:14 GMT+02:00 Deelas <notifications@github.com>:

…

I don't work Cudaminer cudaminer -o stratum+tcp://jp.lapool.me:3015 -u okada.a -p 0908 *** CudaMiner for nVidia GPUs by Christian Buchner *** This is version 2014-02-28 (beta) based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler Cuda additions Copyright 2013,2014 Christian Buchner LTC donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm BTC donation address: 16hJF5mceSojnTD3ZTUDqdRhDyPJzoRakM YAC donation address: Y87sptDEcpLkLeAuex6qZioDbvy1qXZEj4 [2017-09-18 19:59:08] NVML GPU monitoring is not available. [2017-09-18 19:59:08] 1 miner threads started, using 'scrypt' algorithm. [2017-09-18 19:59:08] Starting Stratum on stratum+tcp://jp.lapool.me:3015 [2017-09-18 19:59:09] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaFree(0)' (salsa_kernel.cu line 242) [2017-09-18 19:59:09] GPU #0: GeForce GTX 1060 6GB with compute capability 6.1 [2017-09-18 19:59:09] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaDeviceSetCacheConfig(kernel->cache_config())' ( salsa_kernel.cu line 431) [2017-09-18 19:59:09] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaDeviceSetSharedMemConfig(kernel->shared_mem_config())' ( salsa_kernel.cu line 432) [2017-09-18 19:59:09] GPU #0: interactive: 0, tex-cache: 0 , single-alloc: 0 [2017-09-18 19:59:09] GPU #0: 32 hashes / 4.0 MB per warp. [2017-09-18 19:59:10] GPU #0: Performing auto-tuning (Patience...) [2017-09-18 19:59:10] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaMalloc((void **) &d_idata, mem_size)' (salsa_kernel.cu line 609) [2017-09-18 19:59:10] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaMalloc((void **) &d_odata, mem_size)' (salsa_kernel.cu line 610) [2017-09-18 19:59:11] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaMemcpy(d_idata, h_idata, mem_size, cudaMemcpyHostToDevice)' ( salsa_kernel.cu line 615) [2017-09-18 19:59:11] GPU #0: maximum total warps (BxW): 0 [2017-09-18 19:59:11] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaFree(d_odata)' (salsa_kernel.cu line 745) [2017-09-18 19:59:11] GPU #0: cudaError 78 (a PTX JIT compilation failed) calling 'cudaFree(d_idata)' (salsa_kernel.cu line 746) [2017-09-18 19:59:11] GPU #0: 0.00 hash/s with configuration T0x0 [2017-09-18 19:59:11] GPU #0: using launch configuration T0x0 [2017-09-18 19:59:12] workio thread dead, waiting for workers... [2017-09-18 19:59:12] worker threads all shut down, exiting. Any thoughts? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#159>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAZb0xs3DhQdtwrxD1uYcFlnTEmtBUdeks5sjlB7gaJpZM4PawNu> .

Answer 2 · 2017-09-18T11:53:36.000Z

Thank you for your reply.
Where can I add it to Makefile(Makefile.am)?

Answer 3 · 2017-09-18T13:28:12.000Z

if you're building from source code (on Linux) edit the Makefile and rebuild. If you're just running a Windows binary, don't run ancient binaries from 2014. They have not been tested with GPUs past the Kepler generation. Take one of the 2017 releases done by others (and optimized specifically for Pascal GPUs). Many of these releases are listed on this blog. Note that I am no longer the author/publisher of these forks. http://cryptomining-blog.com/tag/cudaminer/ 2017-09-18 13:53 GMT+02:00 Deelas <notifications@github.com>:

…

Thank you for your reply. Where can I add it to Makefile(Makefile.am)? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#159 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAZb0xMxpIAiSr7k9Cb66lJUwdlKKKwkks5sjlnAgaJpZM4PawNu> .

Answer 4 · 2017-09-19T12:48:37.000Z

I am using linux, so I will try to write Makefile

Answer 5 · 2017-09-19T14:33:26.000Z

I've just looked at my old CudaMiner github repo in the file Makefile.am you'll likely want to replace the -arch options for all cuda build targets with the -gencode statement I've suggested earlier. You'll have to run autoconf automake following that change. Christian 2017-09-19 14:48 GMT+02:00 Deelas <notifications@github.com>:

…

I am using linux, so I will try to write Makefile — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#159 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAZb0_u8G-WSsC1dfifFhrLc0Z3hMGhjks5sj7gmgaJpZM4PawNu> .

Answer 6 · 2017-09-19T14:39:18.000Z

Thanks many times
Problem solved with your advice and http://denor.daa.jp/steamos%E3%81%A7cudaminer