madengine v2 with unified framework for local and distribution#57
Open
madengine v2 with unified framework for local and distribution#57
Conversation
… structure that emphasizes its core strengths in MAD package integration and distributed model execution
…, and K8s; Expanded command line interface;
…environment variables, which is particularly useful in CI/CD environments, containerized deployments, or when you want to avoid storing credentials in files
[WIP] Enhanced distributed execution with runners
…into coketaste/refactor
Fix multi-result on k8s
Update node list config for SLURM
…engine into coketaste/refactor-dis
Silent error check via subprocess
Update the additional context and manifest printout
Update the interface of report to-html
* feat(madengine): ROCm path override and RPD e2e fix ROCM_PATH / --rocm-path: - Add get_rocm_path() and wire through Context, ROCmToolManager, gpu_tool_factory, gpu_validator, container_runner, run orchestrator; add --rocm-path to run CLI - Unit tests for get_rocm_path, Context, ROCmToolManager, run --help; update fixtures and test_get_cached_managers for (vendor, rocm_path) cache RPD e2e: - pre_scripts/trace.sh: install nlohmann-json3-dev (Ubuntu) and json-devel (CentOS) so rocmProfileData rpd_tracer build finds nlohmann/json.hpp * Updated docs * Addd madengine logo icon * Resize the logo icon
…performance, metric, and status show in the correct column (#77) Use header index to replace fixed column order
…to coketaste/refactor-dis
…super) (#78) K8s orchestrator now uses the same reporting path as single-node Docker when multiple_results CSV is present, so both produce the same artifacts: perf.csv, perf_entry.json, perf_entry.csv, perf_super.json, perf_entry_super.json/csv, and perf_super.csv. - Add Docker-compatible reporting path in _collect_results: - _build_common_info_dict() for common_info (args, gpu_arch, etc.) - _ensure_perf_csv_exists() so update_perf_csv can read perf.csv - Call update_perf_csv, update_perf_super_json, update_perf_super_csv with scripts_base_dir so --config from models.json resolves correctly - Multi-node multiple_results: resolve one CSV path per run - _resolve_multiple_results_csv(): single pod → use that CSV; multi-pod → merge all pod CSVs with sum/average rules - _merge_multi_node_multiple_results_csv(): align rows by index; performance aggregated by metric type (throughput→sum, latency→avg, memory→max); extra columns by _aggregation_for_extra_column (sum/avg/max/first) - _aggregation_for_extra_column() for consistent multi-node semantics - Keep legacy row-by-row _write_to_perf_csv when reporting module unavailable; record failure when no CSV found - job.yaml.j2: no functional change required; existing copy block and find fallback for multiple_results already support this refactor
Resolve merge conflicts by keeping refactor-dis (v2) and discarding main (v1) changes: - Remove src/madengine/mad.py and src/madengine/tools/run_models.py (deleted in v2, accept deletion over main's modifications) - Resolve rocenv_tool.py conflict: keep current-branch version for unknown GPU device handling - Resolve tests/fixtures/dummy/models.json: keep v2 fixture set (dummy_superset and full model list) over main's therock-only entry
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
2 participants
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.
Motivation
Technical Details
Test Plan
Test Result
Submission Checklist