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
PR #1654 has been introduced using large GRF mode automatically.
Could we make the cout in these lines be triggered by a debug-only flag so that normal users could safely ignore this? Those should be treated as warnings in my opinion.
std::cout << "(I): Kernel has now " << n_spills << " spills" << std::endl;
}
xpu train AlbertForQuestionAnswering
// We wish those won't exposed to the normal user
(I): Detected 9472 spills, recompiling the kernel using large GRF mode
(I): Kernel has now 512 spills
(I): Detected 20032 spills, recompiling the kernel using large GRF mode
(I): Kernel has now 10816 spills
(I): Detected 33600 spills, recompiling the kernel using large GRF mode
(I): Kernel has now 25408 spills
The text was updated successfully, but these errors were encountered:
I would like to explain my concern as to why we don't set the grf_mode=auto in the triton config from the inductor side.
The reason is that, from the PyTorch inductor side, we are currently trying to keep the same config with CUDA/HIP, so that we would avoid possible unalignments. Large GRF mode for compiling kernels is an optimization for XPU only, thus we would like to hide the complexity from the users familiar with CUDA.
BTW, I am not quite familiar with the differences between different grf_modes. So if there are any concerns, please point out and let's have a discussion.
PR #1654 has been introduced using large GRF mode automatically.
Could we make the
cout
in these lines be triggered by a debug-only flag so that normal users could safely ignore this? Those should be treated as warnings in my opinion.intel-xpu-backend-for-triton/third_party/intel/backend/driver.c
Lines 188 to 201 in 614efe2
The text was updated successfully, but these errors were encountered: