Closed rsmmnt closed 5 years ago
There is an error that occurs from time to time:
2019-05-03 08:02:12,234 INFO [GPU 0 miner] GPU 0 read new block data ERROR: CUDA failed at src/prehash.cu: 1019 Miner is now terminated
It's at cudaMemcpy call and seems like it tries to read from wrong chunk of memory
CUDA_CALL(cudaMemcpy( (void *)&len, (void *)(invalid + 2 * N_LEN), INDEX_SIZE_8, cudaMemcpyDeviceToHost ));
so probably kernels before that call can yield corrupted value of invalid
invalid
fixed by https://github.com/ergoplatform/Autolykos-GPU-miner/pull/38
There is an error that occurs from time to time:
2019-05-03 08:02:12,234 INFO [GPU 0 miner] GPU 0 read new block data ERROR: CUDA failed at src/prehash.cu: 1019 Miner is now terminated
It's at cudaMemcpy call and seems like it tries to read from wrong chunk of memory
so probably kernels before that call can yield corrupted value of
invalid