[GH-ISSUE #2] Small enhancement proposal, save the workflow json along with benchmark json #1

New issue

Open

opened 2026-10-26 22:11:19 +00:00 by bsakaguchi · 1 comment

bsakaguchi commented

2026-10-26 22:11:19 +00:00

Owner

Originally created by @jovan2009 on GitHub (Nov 2, 2025).
Original GitHub issue: https://github.com/Comfy-Org/comfyui-benchmark/issues/2

Hi, I recently found out about this project, I find it very useful and I started using it in my ComfyUI installation. I noticed that (if not configured otherwise) by default it saves benchmark results for every workflow executed in a subfolder in ComfyUI output folder. Very soon that folder has a lot of benchmark files in it. I would propose, if it is not much trouble, that along with the benchmark json in the same folder to be saved the workflow itself that was run, with a naming convention that would make it easy to pair them both later if the user wants to rerun the same workflow again and compare results (or even save it with the original name, sorting the files in the folder by date would also accomplish this pairing). It's almost impossible, otherwise, to remember or identify what exactly was the workflow that was executed, that corresponds to a particular benchmark result, in order to make a comparison later (after a week or a month), apples to apples.

Originally created by @jovan2009 on GitHub (Nov 2, 2025). Original GitHub issue: https://github.com/Comfy-Org/comfyui-benchmark/issues/2 Hi, I recently found out about this project, I find it very useful and I started using it in my ComfyUI installation. I noticed that (if not configured otherwise) by default it saves benchmark results for every workflow executed in a subfolder in ComfyUI output folder. Very soon that folder has a lot of benchmark files in it. I would propose, if it is not much trouble, that along with the benchmark json in the same folder to be saved the workflow itself that was run, with a naming convention that would make it easy to pair them both later if the user wants to rerun the same workflow again and compare results (or even save it with the original name, sorting the files in the folder by date would also accomplish this pairing). It's almost impossible, otherwise, to remember or identify what exactly was the workflow that was executed, that corresponds to a particular benchmark result, in order to make a comparison later (after a week or a month), apples to apples. <img width="706" height="573" alt="Image" src="https://github.com/user-attachments/assets/f6cb63a7-cd20-48b9-97b3-9967423efe33" />

bsakaguchi commented

2026-10-26 22:11:20 +00:00

Author

Owner

@Kosinkadink commented on GitHub (Nov 26, 2025):

Thanks for the feedback, I will try to add this in soon and make it configurable via the config.yaml!

@Kosinkadink commented on GitHub (Nov 26, 2025): Thanks for the feedback, I will try to add this in soon and make it configurable via the config.yaml!

bsakaguchi referenced this issue

2026-10-26 22:11:23 +00:00

[PR #1] [MERGED] add support for benchmark worflow node #4