Improve std::vector allocations and insertions performance#37880

Draft

mateusznowakTT wants to merge 54 commits intomainfrom

mateusznowakTT/optimize_std_vector_allocations

Commits on Feb 12, 2026

Optimize std::vector allocations with reserve()

mateusznowakTT
and
claude
committed
Add reserve() to matmul, all_gather, and groupnorm operations

mateusznowakTT
and
claude
committed
Replace list initialization with reserve+emplace for better performance

mateusznowakTT
and
claude
committed
Add reserve() to transpose and CCL operations

mateusznowakTT
and
claude
committed
Fix: Correct reserve() call for std::set in groupnorm_mcast

mateusznowakTT
and
claude
committed
Optimize vector allocations: permute, reshard, llama matmul, groupnorm

mateusznowakTT
and
claude
committed
Optimize vector allocations: rms_allgather and groupnorm_sharded

mateusznowakTT
and
claude
committed
Optimize vector allocations: sdpa, sort, reduce, adam, sliding_window

mateusznowakTT
and
claude
committed
Optimize vector allocations across 20 files: ccl, moreh, concat, mesh, tensor

mateusznowakTT
and
claude
committed

Commits on Feb 13, 2026

Commits on Feb 15, 2026

Commits on Feb 16, 2026