You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
on RTX4090 running llama3-8B, allocate_kv_data_batchsize=10, I encountered the error of the following:
CUDA Error: misaligned address (716) /code/gemv/include/attention/prefill.cuh: line 1918 at function cudaLaunchKernel((void*)kernel, nblks, nthrs, args, smem_size, stream)
terminate called after throwing an instance of 'std::invalid_argument'
what(): BatchPrefillWithPagedKVCache Failed with error: misaligned address
It happens in the 2nd Cycle
The text was updated successfully, but these errors were encountered:
on RTX4090 running llama3-8B, allocate_kv_data_batchsize=10, I encountered the error of the following:
CUDA Error: misaligned address (716) /code/gemv/include/attention/prefill.cuh: line 1918 at function cudaLaunchKernel((void*)kernel, nblks, nthrs, args, smem_size, stream)
terminate called after throwing an instance of 'std::invalid_argument'
what(): BatchPrefillWithPagedKVCache Failed with error: misaligned address
It happens in the 2nd Cycle
The text was updated successfully, but these errors were encountered: