You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Move ci/scripts/metrics_unify/flamegraph.py to openvm-prof (in a flamegraph.rs file)
I think we could directly use inferno without a subprocess call to cargo-flamegraph.
The text was updated successfully, but these errors were encountered:
What about samply, (flamegrpah.rs recomend it in their README.md). I tested on the sdk you just need define a [profile.profiling] and works out-of-box.
What about samply, (flamegrpah.rs recomend it in their README.md). I tested on the sdk you just need define a [profile.profiling] and works out-of-box.
Here I am referring to this script which is for circuit flamegraphs where we emit custom metrics (similar to linux perf but specific to VM / STARKs). The python script parses them into a format compatible with flamegraph. Indeed you could also output a profile.json that could be given to samply -- I haven't looked into this but it would be cool.
jonathanpwang
changed the title
Add flamegraphs to openvm-prof
Add circuit flamegraphs to openvm-profFeb 28, 2025
Move
ci/scripts/metrics_unify/flamegraph.py
toopenvm-prof
(in aflamegraph.rs
file)I think we could directly use
inferno
without a subprocess call tocargo-flamegraph
.The text was updated successfully, but these errors were encountered: