Open
Conversation
- pulp_conv_naive_fp32.c/h: add dw_kernel_weight_grad_padded and dw_kernel_input_grad_padded for DW ConvGrad with non-zero padding or stride != 1. Use precomputed valid-range loops (ho_min/ho_max, wo_min/wo_max) instead of inner-loop conditionals to avoid a GCC -O3 -ffast-math miscompile on RISC-V. - pulp_conv_dw_fp32.c: dispatch to padded kernels in both pulp_conv_dw_fp32_bw_param_grads_cl and pulp_conv_dw_fp32_bw_input_grads_cl when padding or stride != 1; forward stride/padding fields that were previously missing. - pulp_im2col_fp32.c: remove overly-strict validity check that returned early (leaving the buffer uninitialized) whenever (Hin - Hk + pad) was not divisible by stride. - pulp_conv2d_fp32.c: pass actual Lpad/Rpad/Upad/Dpad to im2col in pulp_conv2d_fp32_bw_param_grads_cl (previously hard-coded to 0).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
pulp_conv_naive_fp32.c/h: add dw_kernel_weight_grad_padded and dw_kernel_input_grad_padded for DW ConvGrad with non-zero padding or stride != 1. Use precomputed valid-range loops (ho_min/ho_max, wo_min/wo_max) instead of inner-loop conditionals to avoid a GCC -O3 -ffast-math miscompile on RISC-V.
pulp_conv_dw_fp32.c: dispatch to padded kernels in both pulp_conv_dw_fp32_bw_param_grads_cl and pulp_conv_dw_fp32_bw_input_grads_cl when padding or stride != 1; forward stride/padding fields that were previously missing.
pulp_im2col_fp32.c: remove overly-strict validity check that returned early (leaving the buffer uninitialized) whenever (Hin - Hk + pad) was not divisible by stride.
pulp_conv2d_fp32.c: pass actual Lpad/Rpad/Upad/Dpad to im2col in pulp_conv2d_fp32_bw_param_grads_cl (previously hard-coded to 0).