Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

CUDA failed at autolykos.cu: 304 #53

Open
nils-trubkin opened this issue Aug 27, 2022 · 2 comments
Open

CUDA failed at autolykos.cu: 304 #53

nils-trubkin opened this issue Aug 27, 2022 · 2 comments

Comments

@nils-trubkin
Copy link

Can't start miner

RTX3080 + 2070S
Driver 516.94

ERROR:  CUDA failed at autolykos.cu: 304
CUDA error code 700
Miner is now terminated
@tjayz
Copy link

tjayz commented Aug 31, 2022

2022-08-30 21:48:11,177 INFO [main thread] Using 1 GPU devices
2022-08-30 21:48:11,177 INFO [main thread] Using configuration file ./config.json
2022-08-30 21:48:11,178 INFO [main thread] Block getting URL:
http://127.0.0.1:3000/mining/candidate
2022-08-30 21:48:11,178 INFO [main thread] Solution posting URL:
http://127.0.0.1:3000/mining/solution
2022-08-30 21:48:11,184 INFO [main thread] Got new block in main thread, block data: {"MSG":"A7A93691D6769841577B705747DD328D8963E8C3C9E7E9319376857D5DC1146E","B":499820569754874971280969147457920358060901175036868974332871072,"EXTRANONCE1":"00","EXTRANONCE2SIZE":7,"HEIGHT":827769} 2022-08-30 21:48:11,646 INFO [GPU 0 miner] GPU 0 started
ERROR: CUDA failed at src/mining.cu: 310
CUDA error code 700
Miner is now terminated

I get same error as poster this one as well. Mining to own node 4.0.41 or woolypooly, built ergostratumproxy from source, run from ./auto.out with nvidiatoolkit 11.6.2 driver 511.09 wsl2 ubuntu or debian

@tjayz
Copy link

tjayz commented Aug 31, 2022

#51

Timbers007 added a commit to Timbers007/Autolykos2_NV_Miner that referenced this issue Oct 15, 2022
@Timbers007
Fix prehash crashing on newer NVIDIA drivers
36cdbc0 
Using any NVIDIA driver that is part of a 5XX release branch (such as any drivers released after and including 510.39.01 for Linux or 511.23 for Windows) will cause the miner to crash with cudaErrorIllegalAddress (700) at the prehash stage. Downgrading to an older driver such as 470.141.03 has been a stop-gap solution to getting the miner running again without this patch.

The issue is coming from this line: https://github.com/mhssamadani/Autolykos2_NV_Miner/blob/4de5cea0952a86d5d287d9788146e0840cb9b61d/secp256k1/src/prehash.cu#L251

Simply signaling to the compiler to disable unrolling for this one loop seems to be enough to make the driver happy with no measurable performance difference as far as I have measured.

Fixes mhssamadani#51 mhssamadani#53
@Timbers007 Timbers007 mentioned this issue Oct 15, 2022
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@nils-trubkin @tjayz