ywkang/kernbench2
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

analytical-viz: roofline tab — drop formula from step-latency header

Browse Source 
Revert the multi-line formula annotation on the 'Latency per decode
step' header; make it a single-line markdown title matching the
'Cost per token' header on the right. Both headline plots now render
at the same height so the plot area lines up.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
Mukesh Garg
2026-07-29 14:14:24 -07:00
parent a1c3c28f62
commit bda76cbd66
1 changed files with 1 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/analytical_visualization/app.py
+1 -5
View File
@@ -2255,11 +2255,7 @@ with tab_roofline:
# Plot A — Latency per decode step (undivided by B) # Plot A — Latency per decode step (undivided by B)
with _hcol1: with _hcol1:
st.markdown( st.markdown("**Latency per decode step** (one forward pass)")
"**Latency per decode step** (one forward pass)\n\n"
"`t_step = N·b/W + 2·N·B/C + B·S_kv·kv_bpt/W`\n\n"
"` = weight_fetch + compute·B + KV_read·B`"
)
_figA, _axA = plt.subplots(figsize=(6, 4.2)) _figA, _axA = plt.subplots(figsize=(6, 4.2))
_axA.plot(_xs_head, [p.weight_s * 1e3 for p in _step_pts], _axA.plot(_xs_head, [p.weight_s * 1e3 for p in _step_pts],
"^-", color="#ffbe0b", "^-", color="#ffbe0b",