CUDA error: too many resources requested for launch

[jkrauska]

Tried running on my hardware (2080), getting a cuda error.

~/src/gpusnarks/build/test$ ./fft_test
Field size: 32, Field count: 262144
A address: 0x7f2ae4d54010 Last element: 0
CUDA Devices: 1, Field size: 32, Field count: 262144
threads 65536, blocks 65, threads 1024
CUDA error: too many resources requested for launch

smi info

$ nvidia-smi
Tue May 14 21:55:38 2019
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 418.67       Driver Version: 418.67       CUDA Version: 10.1     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce RTX 208...  Off  | 00000000:01:00.0 Off |                  N/A |
| 17%   45C    P0     1W / 260W |      0MiB / 10989MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

[MariusVanDerWijden]

Thanks for the report!
The current master does not work properly, as it tries to use more registers than there are on a GPU. This is because I'm currently working on the switch from 32bit integers to my own 256bit library.
You can test an outdated version of the code here: https://github.com/MariusVanDerWijden/gpusnarks/releases/tag/v0.0.1

I'm currently working on the 256bit finite field library, however I can not allocate too much time, as the Ethereum Foundation has still not payed the prize money from ETHParis and I'm a starving student ;)