[Feature] Add idle (gap) and wall time (span) rows to layer detail re… - #1
Open
akao-amd wants to merge 5 commits into
Open
[Feature] Add idle (gap) and wall time (span) rows to layer detail re…#1akao-amd wants to merge 5 commits into
akao-amd wants to merge 5 commits into
Conversation
- PythonSourceIndex gains get_callstack() / _find_callstack(): collects all python_function frames enclosing a kernel launch timestamp, sorted outermost-first (widest span first), stripping boilerplate wrappers. - KernelDetail gains a callstack slot populated during correlation. - KernelCorrelator.correlate() stores _callstack on each kernel event. - ReportGenerator.write_callstack_markdown() writes a Markdown file covering the same representative module instances as the Excel detail tabs, with per-kernel metadata tables and full call stack blocks. - New --callstack-md PATH argument; works standalone without -o. - Auto-selection uses self_kernel_time (not total) so container modules don't outrank the leaf layer types that own the actual work. Co-Authored-By: Claude <noreply@anthropic.com>
Generates a self-contained HTML file (no server needed) with: - Tab per selected module type showing instance count and kernel count - Kernel timeline chart: bar per kernel (execution order on X), Y = avg duration (µs) across all layer instances, error bars show min/max variance - Category bar chart: horizontal bars summing avg duration per category; updates live as kernels are reassigned via drag-and-drop - Category buckets: one chip per unique kernel type (not one per occurrence); dragging a chip reassigns all occurrences of that kernel in the layer; chip shows repeat count (e.g. "×2") when the kernel appears multiple times - Hover tooltips on all bars and chips with full kernel name, module path, category, avg/min/max, input dims, and launch source - Collapsible table view with all kernels in execution order - Light/dark mode via CSS custom properties and prefers-color-scheme - Categorical palette from the validated dataviz reference (8-slot fixed order, CVD-safe adjacent ordering) Works standalone without -o. Use --detail-module to target a specific layer type (e.g. --detail-module GptOssDecoderLayer). Auto-selection uses self_kernel_time ranking so container wrappers don't outrank leaf modules. Co-Authored-By: Claude <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.