Skip to content

cherry pick reorder #309

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
azhai219 wants to merge 2 commits into
base: v3.10_for_ie_master
Choose a base branch
from
Open

cherry pick reorder #309

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f692013
Revert "Revert "cpu: x64: matmul: add missing VDISPATCH_REORDER_IC ch…
azhai219 Apr 17, 2026
da00166
cpu: x64: matmul: fix VDISPATCH_REORDER_IC use with memory_desc_reshape
asztukow Apr 17, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 11 additions & 7 deletions src/cpu/x64/matmul/brgemm_matmul_reorders.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,10 +79,14 @@ status_t calculate_plain_transpose_blocks(dim_t &batch, dim_t &M, dim_t &K,
}

memory_desc_t src_md_reduced, dst_md_reduced;
CHECK(memory_desc_reshape(
src_md_reduced, src_md, non_unit_dim, non_unit_dims));
CHECK(memory_desc_reshape(
dst_md_reduced, dst_md, non_unit_dim, non_unit_dims));
VDISPATCH_REORDER_IC(memory_desc_reshape(src_md_reduced, src_md,
non_unit_dim, non_unit_dims)
== status::success,
VERBOSE_UNSUPPORTED_TENSOR_LAYOUT, "src");
VDISPATCH_REORDER_IC(memory_desc_reshape(dst_md_reduced, dst_md,
non_unit_dim, non_unit_dims)
== status::success,
VERBOSE_UNSUPPORTED_TENSOR_LAYOUT, "dst");

const memory_desc_wrapper id(src_md_reduced), od(dst_md_reduced);

Expand Down Expand Up @@ -386,7 +390,7 @@ status_t brgemm_matmul_copy_reorder_t::execute_body(
= dst_d.size() - dst_d.additional_buffer_size();
const size_t s8s8_comp_size_bytes = kernel_conf.s8s8_compensation_required
? dst_d.additional_buffer_size(
memory_extra_flags::compensation_conv_s8s8)
memory_extra_flags::compensation_conv_s8s8)
: 0;
const size_t zp_comp_offset_bytes
= comp_offset_bytes + s8s8_comp_size_bytes;
Expand Down Expand Up @@ -469,7 +473,7 @@ status_t brgemm_matmul_copy_reorder_t::execute_body(
const auto src_offset = !kernel_conf.blocked_B
? get_blk_off(src_d, sdt_sz, batch, k, n)
: get_blk_off(src_d, sdt_sz, batch, k_blk_idx,
n_blk_idx);
n_blk_idx);
ker_exec_ctx.src
= (void *)&src[src_offset / src_typesz_scale];
ker_exec_ctx.tr_src = (void *)&dst[get_blk_off(
Expand All @@ -483,7 +487,7 @@ status_t brgemm_matmul_copy_reorder_t::execute_body(
const auto src_offset = !kernel_conf.blocked_B
? get_blk_off(src_d, sdt_sz, batch, k, n)
: get_blk_off(src_d, sdt_sz, batch, k_blk_idx,
n_blk_idx);
n_blk_idx);
ker_exec_ctx.src
= (void *)&src[src_offset / src_typesz_scale];
const auto dst_offset = get_blk_off(
Expand Down
Loading