xmrig-nvidia icon indicating copy to clipboard operation
xmrig-nvidia copied to clipboard

Program compile log: nvrtc: error: invalid value for --gpu-architecture (-arch)

Open jerrynpc opened this issue 4 years ago • 3 comments

what can i do·?for error

  • ABOUT XMRig-NVIDIA/2.14.4 MSVC/2015
  • LIBS libuv/1.29.1 CUDA/9.0 OpenSSL/1.1.1b microhttpd/0.9.63
  • CPU Intel(R) Core(TM) i5-9300H CPU @ 2.40GHz x64 AES
  • GPU #0 PCI:0000:01:00 GeForce RTX 2060 @ 1200/7001 MHz 26x90 6x25 arch:75 SMX:30 MEM:5006/6144 MiB
  • ALGO cryptonight, donate=5%
  • POOL #1 donate.v2.xmrig.com:3333 variant auto
  • COMMANDS hashrate, health, pause, resume [2019-07-16 16:16:46] use pool donate.v2.xmrig.com:3333 159.89.38.204 [2019-07-16 16:16:46] new job from donate.v2.xmrig.com:3333 diff 1000225 algo cn/r height 1879609 [2019-07-16 16:16:47] nvrtcCompileProgram failed: NVRTC_ERROR_INVALID_OPTION [2019-07-16 16:16:47] Program compile log: nvrtc: error: invalid value for --gpu-architecture (-arch)

[CUDA] Error gpu 0: <cryptonight_gpu_hash>:878 "device kernel image is invalid" 请按任意键继续. . .

jerrynpc avatar Jul 16 '19 08:07 jerrynpc

Same issus here

$ ./xmrig-nvidia.exe

  • ABOUT XMRig-NVIDIA/2.14.4 MSVC/2015
  • LIBS libuv/1.29.1 CUDA/9.0 OpenSSL/1.1.1b microhttpd/0.9.63
  • CPU Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz x64 AES
  • GPU #0 PCI:0000:01:00 GeForce RTX 2070 @ 1815/7001 MHz 30x108 6x25 arch:75 SMX:36 MEM:6738/8192 MiB
  • ALGO cryptonight, donate=1%
  • POOL #1 pool.hashvault.pro:3333 variant=auto
  • COMMANDS 'h' hashrate, 'e' health, 'p' pause, 'r' resume [2019-08-02 21:54:10] use pool pool.hashvault.pro:3333 139.180.138.233 [2019-08-02 21:54:10] new job from pool.hashvault.pro:3333 diff 20000 algo cn/r height 1891993 [2019-08-02 21:54:11] nvrtcCompileProgram failed: NVRTC_ERROR_INVALID_OPTION [2019-08-02 21:54:11] Program compile log: nvrtc: error: invalid value for --gpu-architecture (-arch)

[CUDA] Error gpu 0: <cryptonight_gpu_hash>:878 "device kernel image is invalid"

kanztu avatar Aug 02 '19 13:08 kanztu

For RTX you should use version build with CUDA 10+. Thank you.

xmrig avatar Aug 02 '19 14:08 xmrig

Despite using the CUDA 10+ release, I am still receiving this error on Win10 with an RTX 3090.

trevorprater avatar Jan 22 '21 05:01 trevorprater