Judge an AIE matmul against its MAC unit's peak

[asyms]

Stacked on #160.

The AIE estimator took a matmul's ideal cycles from a fixed rate of 32 bf16 ops per cycle per tile. A measured call of the library's matmul runs at about 150 MACs per cycle, so the reported efficiency came out above 1.

A matmul kernel's ideal now comes from the kernel library's MAC tile, one of which the matmul unit computes per cycle. For the AIE2P library that is 8x8x8: the bfp16 mmul<8,8,8> the bf16 matmul issues, 512 MACs per cycle, the same rate as int8 on that tile. On AIE2 the native bf16 tile is 4x8x4, 128 MACs per cycle. Vector kernels, and nodes without a library, keep the per-datatype rate. A measured call is still what prices the node.

Test: tests/unit/test_aie_cost_estimator.py. Fast suite: 911 passed.

[github-actions]

Stream AIE Metrics Regression Guard

⚠ 7 cell(s) flagged (total_latency > 0.1% tol): hardware_swiglu[fusemax], hardware_two_conv[eyeriss_like_dual_core], hardware_two_conv[eyeriss_like_quad_core], hardware_two_conv[fusemax], hardware_two_conv[meta_prototype], hardware_two_conv[simba_small], hardware_two_conv[tpu_like_quad_core]

16 of 16 cells captured

Provenance: baseline 32b8eaa94931143f86c3a135c774febd1d8b96ae | date 2026-10-04 | Python 3.12.3 | backend ortools_gscip
Note: mip_gap: null — OR-Tools GSCIP
Columns: array fill = per-layer PE-array spatial fill (dataflow quality); MAC eff (e2e) = useful MACs / (chip peak MACs/cycle × total latency), the true fraction of the chip's compute used (incl. idle cores, temporal stalls & transfers).

hardware_swiglu — 8 hardware (⚠ 1 flagged)

seq_len=256, embedding_dim=2048, hidden_dim=8192, bf16; layer-fused tiles seq=16/embedding=128/hidden=32

Hardware	total_latency (base → cur)	Δ%	array fill	MAC eff (e2e)	note
eyeriss_like_dual_core	149684492 → 149684492	+0.00%	52%	26%
eyeriss_like_quad_core	101581557 → 101581557	+0.00%	52%	19%
eyeriss_like_single_core	303038604 → 303038604	+0.00%	52%	25%
fusemax	233766951 → 510918663	⚠ ↑+118.56%	6.2%	0.75%
meta_prototype	101711933 → 101711933	+0.00%	100%	6.2%
simba	125344507.60 → 125344507.60	+0.00%	65%	1.0%
simba_small	106627594 → 106627594	+0.00%	65%	3.0%
tpu_like_quad_core	101580877 → 101580877	+0.00%	97%	3.1%

hardware_two_conv — 8 hardware (⚠ 6 flagged)

batch=1, in_ch=8, H=32, W=32, out_ch1=16, out_ch2=32, kernel=3x3, bf16 (generic auto-tiling, not layer-fused)

Hardware	total_latency (base → cur)	Δ%	array fill	MAC eff (e2e)	note
eyeriss_like_dual_core	74631 → 74761	⚠ ↑+0.17%	51%	23%
eyeriss_like_quad_core	39917 → 40209	⚠ ↑+0.73%	51%	22%
eyeriss_like_single_core	115003 → 115003	+0.00%	68%	31%
fusemax	187612 → 22546	⚠ ↓-87.98%	1.6%	0.48%	low array utilization
meta_prototype	15341 → 15439	⚠ ↑+0.64%	51%	19%
simba	4610.67 → 4610.67	+0.00%	50%	4.9%
simba_small	8706 → 8998	⚠ ↑+3.35%	50%	16%
tpu_like_quad_core	12300 → 12592	⚠ ↑+2.37%	42%	11%

To regenerate baseline: python scripts/analysis/render_metrics_comment.py --update-baseline