paper(gqa): derive (m,l,O) AR per-PE cost from kernel topology
Replaces the unsourced 32 KB/layer placeholder in the analytical chart with T*(N-1)/N where T = h_q * S_q * (d_head*2 + 8) ~ 2.1 KB (per-PE (m, l, O) payload) and N is the participant count of the reduce-only chain per stage. Cases 2/3/6 analytical bars now match the simulator-measured numbers within ~10% (was 6-30x over). Also: - Bumps the measurement S_kv from 8 K to 64 K to verify Cases 1, 2, 3, 6 are genuinely S_kv-independent and Cases 4, 5 scale linearly. - In-bar descriptor on the paired chart is now wrapped to fit the bar width, black, no background box, and centered between the analytical and measured bars so the label applies to both. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
Binary file not shown.
|
Before Width: | Height: | Size: 134 KiB After Width: | Height: | Size: 132 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 182 KiB After Width: | Height: | Size: 183 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 185 KiB After Width: | Height: | Size: 182 KiB |
Reference in New Issue
Block a user