-
Notifications
You must be signed in to change notification settings - Fork 29
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.
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
Nano-Vanity compiles, but won't run with the gpu flag (Error passing generic uint32_t to const uint32_t) #52
Comments
Part 2 of error:
|
So, I had this working fine about a year ago. Something either in the Rust version or windows update, has ruined this. I can't figure out what it is either. I can no longer get what used to work for me.... to work anymore. |
Ok, so I figured out the issue for myself using a 3070 GPU on a Lenovo Legion 7 2021 AMD model and I'm convinced I can get it to work for you too. I simply rolled-back my GPU drivers to the dates that I recall it was working for me... So I downloaded and installed this set of driver from Aug. 2021. I didn't even have to restart my laptop, it just worked after that. Now, for yours... you'll want to try version 471.96 below. Once installed, you'll want to I imagine you're running the most recent branch of nano-vanity... I have a slightly modified version, so if you still get issues after downloading / installing / runnning with the August 31st 2021 drivers, let me know and I'll pass the cleaned version of my modified branch to you for further testing. Edit: If you're using a different Windows version like 8... you'll want to use this. |
Wow, thank you so much for the detailed research and workaround @adrianegraphene! That indeed resolved this issue for me! I am now able to run with the -g flag and get >2,000,000 keys/second 🙌 |
You're welcome. Glad to hear you got it working! |
Something in the 40xx series release broke this. The last driver 517.48 before that date still works. |
Any chance for a fix for this? 3000 series cards and older have the option to downgrade their driver version, but 4000 series and anything after will be unable to use the GPU flag going forward. Thanks for supporting this software for so long! |
I was kind of hoping nvidia would fix it, since it definitely seems to be a bug in their drivers. IIRC you can patch the OpenCL kernel a bit to make it work. I'll take a look. |
The kernel compilation is currently just hanging on my system, so I'm afraid I can't really build that alternative kernel. I think if you change Perhaps you could try opening a bug report with nvidia. |
This solved it for me: #56 No idea if it's the right way to do it per what @PlasmaPower suggested - or how it impacts AMD, etc. But it works fine for me on latest nvidia driver. |
This is fixed with #56. My guess is AMD is (still?) not working though |
Nano-vanity compiles, but won't run with the GPU flag, even after installing GPU drivers from scratch + Visual Studio + CUDA Toolkit.
Also tried this, but no luck so far.
Nvidia 1080 Ti
Driver version: 527.56
Cuda:
Error:
The text was updated successfully, but these errors were encountered: