Documentation for outputted results

[nurSaadat]

Hello!

I was running perf-cost experiment and I still don't understand if I am getting my measurements right. I am using the generated result.csv to build box-and-whiskers graphs and the time metric looks weird. I'd like to ask, what is the unit of measure for time that I can see in my result.csv table? Are those milliseconds or seconds?

I get some ridiculously big time measurements, I was sure those were milliseconds, but dividing them by 1000 didn't help that much. I am attaching some files and graph pictures to see the results of the experiments I performed. UPD:

• The name is above the figure.
• The time value is already divided by 1000
• Those are warm invocations, exec_time (benchmark time)

amazon_compression

amazon_thumbnailer

amazon_uploader

azure_compression

azure_thumbnailer

azure_uploader

google_compression

google_thumbnailer

google_uploader

azure_uploader.csv azure_thumbnailer.csv azure_compression.csv amazon_compression.csv google_compression.csv google_thumbnailer.csv amazon_uploader.csv google_uploader.csv amazon_thumbnailer.csv

[nurSaadat]

@mcopik I would love to hear your opinion on this matter

[nurSaadat]

The problem is, I am trying to replicate your paper results for a reproducibility challenge, but the results are not the same and I have no idea how to get there.

Even the order of magnitude looks like not compatible

[mcopik]

@nurSaadat We updated the main branch with new and expanded documentation on the benchmarks. We now clearly state that all time measurements are in microseconds.

Hope it helps - please let me know via email or here if there's anything else that is missing in our documentation or incorrect, we'll fix it!

[nurSaadat]

@mcopik Thank you very much! That clears my concerns.