Skip to content

Avoid int overflow in block_map heuristics for large matmuls #375

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
EylonKrause wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Avoid int overflow in block_map heuristics for large matmuls #375

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
21 changes: 15 additions & 6 deletions ruy/block_map.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -136,9 +136,14 @@ BlockMapTraversalOrder GetTraversalOrder(
int rhs_scalar_size, const CpuCacheParams& cpu_cache_params) {
static constexpr bool kAnyFractal =
RUY_OPT(FRACTAL_Z) | RUY_OPT(FRACTAL_U) | RUY_OPT(FRACTAL_HILBERT);
const int working_set_size =
(lhs_scalar_size * rows_after_rectangularness_division +
rhs_scalar_size * cols_after_rectangularness_division) *
// Promote to int64 before multiplying: for a large matmul this product
// (bytes touched) exceeds INT_MAX, and a plain int computation would
// overflow (UB) and wrap negative, wrongly selecting a linear traversal for
// a working set that does not fit in cache. Mirrors GetTentativeThreadCount
// in trmul.cc, which promotes the same rows*cols*depth product to int64.
const std::int64_t working_set_size =
(std::int64_t{lhs_scalar_size} * rows_after_rectangularness_division +
std::int64_t{rhs_scalar_size} * cols_after_rectangularness_division) *
depth;
if (kAnyFractal && (working_set_size > cpu_cache_params.local_cache_size)) {
if (RUY_OPT(FRACTAL_HILBERT) &&
Expand Down Expand Up @@ -258,9 +263,13 @@ int GetCacheLocalityScore(int block_size_log2, int rows, int cols, int depth,
}
const int block_rows = std::min(1 << block_size_log2, rows);
const int block_cols = std::min(1 << block_size_log2, cols);
const int total_read_bytes =
(lhs_scalar_size * block_rows + rhs_scalar_size * block_cols) * depth;
const int total_read_bytes_log2 = ceil_log2(total_read_bytes);
// Promote to int64 before multiplying to avoid signed overflow (UB) for a
// large depth; the result only feeds ceil_log2 below (see GetTraversalOrder).
const std::int64_t total_read_bytes =
(std::int64_t{lhs_scalar_size} * block_rows +
std::int64_t{rhs_scalar_size} * block_cols) *
depth;
const int total_read_bytes_log2 = static_cast<int>(ceil_log2(total_read_bytes));
const int nonlocality_log2 =
total_read_bytes_log2 - floor_log2(cpu_cache_params.local_cache_size);
// The values here have been tuned on ARM Cortex-A55.
Expand Down