Under 10 hrs runtime: Resource adequacy #225
Conversation
for more information, see https://pre-commit.ci
…/solver-benchmark into Resource_Adequacy
…conflicting files
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
for more information, see https://pre-commit.ci
runner/run_benchmarks.py
Outdated
| @@ -207,23 +207,36 @@ def benchmark_solver(input_file, solver_name, timeout, solver_version): | |||
| if "XDG_RUNTIME_DIR" in os.environ: | |||
| command.append("--user") | |||
|
|
|||
| # I modify this part | |||
There was a problem hiding this comment.
This comments seems like a leftover that should be removed.
There was a problem hiding this comment.
Thanks @KristijanFaust-OET just deleted this :)
| command.extend( | ||
| [ | ||
| "--scope", | ||
| f"--property=MemoryMax={memory_limit_bytes}", # Set resident memory limit |
There was a problem hiding this comment.
These comments IMO can come in handy for the future. It was quite an effort to figure out how to prevent system wide OOMs, and keywords like resident play a crucial role in making it happen and understanding it. I think we should preserve both these comment and the one below, so that there isn't any confusion why we have these here in the future.
runner/run_benchmarks.py
Outdated
| "python", | ||
| f"{Path(__file__).parent / 'run_solver.py'}", | ||
| solver_name, | ||
| input_file, | ||
| solver_version, | ||
| ] | ||
| ) | ||
| # End of modification |
There was a problem hiding this comment.
This also looks like an leftover that should be removed.
There was a problem hiding this comment.
Thanks @KristijanFaust-OET just deleted this :)
for more information, see https://pre-commit.ci
Hi @siddharth-krishna and @danielelerede-oet,
Please take a look at this PR. The benchmark
DCOPF-Carolinas_uc_6Mnow runs to completion in 8.55 hours using HiGHS. In the previous run results from Sid, I noticed that when the runtime exceeded 600 seconds, the termination condition was marked asTO(timeout). I made a small change inrun_benchmarks.pyso that specifying--timeout 0disables the timeout constraint. You can now run:This may already be handled differently elsewhere in the repository when running the full benchmark suite, but just wanted to flag it in case you notice changes.
I'm currently on holiday, but feel free to leave any comments and I’ll respond as soon as I can.
Best,
Luis