Skip to content

ggml-cpu: keep the generic SIMD vec kernel accumulators in registers - #40

Open
codspeed-hq[bot] wants to merge 1 commit into
masterfrom
codspeed-optim-keep-the-simd-vec-kernel-accumulators-in-registers-1785335630636
Open

ggml-cpu: keep the generic SIMD vec kernel accumulators in registers#40
codspeed-hq[bot] wants to merge 1 commit into
masterfrom
codspeed-optim-keep-the-simd-vec-kernel-accumulators-in-registers-1785335630636

ggml-cpu: keep the generic SIMD vec kernel accumulators in registers

0ebdd46
Select commit
Loading
Failed to load commit list.
CodSpeed HQ / CodSpeed Performance Analysis succeeded Jul 29, 2026

Performance Gate Passed

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 4 improved benchmarks
✅ 24 untouched benchmarks

Performance Changes

Mode Benchmark BASE HEAD Efficiency
Simulation mul_mat[f32] 346.1 ms 232.6 ms +48.81%
Simulation ffn_swiglu 1,067.7 ms 727.9 ms +46.69%
WallTime lm_head[f16] 60.6 ms 42.6 ms +42.19%
WallTime prompt_layer[f32] 247.1 ms 195.1 ms +26.62%

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.


Comparing codspeed-optim-keep-the-simd-vec-kernel-accumulators-in-registers-1785335630636 (0ebdd46) with master (46819c9)

Open in CodSpeed