starred/flyimg

Fork 0

mirror of https://github.com/flyimg/flyimg.git synced 2026-04-25 09:45:50 +03:00

[GH-ISSUE #427] Benchmarks clarification request #768

New issue

Closed

opened 2026-03-15 15:20:45 +03:00 by kerem · 5 comments

kerem commented

2026-03-15 15:20:45 +03:00

Owner

Originally created by @peraronsson on GitHub (Mar 9, 2024).
Original GitHub issue: https://github.com/flyimg/flyimg/issues/427

Hello, just a few questions regarding the benchmarks:

Latencies: does this mean "response time per request"? (which is what I want to know)
Bytes Out: this is always zero, so I guess it's some sort of bug?
Test Image: what is the original size of Rovinj-Croatia.jpg?

I'm trying to understand where Flyimg would be in this comparison.

Originally created by @peraronsson on GitHub (Mar 9, 2024). Original GitHub issue: https://github.com/flyimg/flyimg/issues/427 Hello, just a few questions regarding the [benchmarks]( https://github.com/imgproxy/imgproxy/blob/master/BENCHMARK.md): - **Latencies**: does this mean "response time per request"? (which is what I want to know) - **Bytes Out**: this is always zero, so I guess it's some sort of bug? - **Test Image**: what is the original size of `Rovinj-Croatia.jpg`? I'm trying to understand where Flyimg would be in [this comparison](https://github.com/imgproxy/imgproxy/blob/master/BENCHMARK.md#results).

kerem closed this issue

2026-03-15 15:20:50 +03:00

kerem commented

2026-03-15 15:20:56 +03:00

Author

Owner

@sadok-f commented on GitHub (Mar 10, 2024):

Hello @peraronsson
Thank you for opening this issue!
the actual benchmark in the Flyimg repo is outdated, I think it's a good idea to update it and maybe use the Apache ab tool.
regarding your questions:

yes the Latencies mean the response time per request
BytesOut is the number of sent bytes in the HTTP body (From vegeta tool documentation)
Rovinj-Croatia.jpg Original size is 57.3 KB

Actually I'll update the benchmark script to use ab tool to have a meaningful comparison with the other tools you mentioned.

@sadok-f commented on GitHub (Mar 10, 2024): Hello @peraronsson Thank you for opening this issue! the actual benchmark in the Flyimg repo is outdated, I think it's a good idea to update it and maybe use the Apache `ab` tool. regarding your questions: - yes the Latencies mean the response time per request - BytesOut is the number of sent bytes in the HTTP body (From vegeta tool [documentation](https://github.com/tsenart/vegeta)) - `Rovinj-Croatia.jpg` Original size is 57.3 KB Actually I'll update the benchmark script to use `ab` tool to have a meaningful comparison with the other tools you mentioned.

kerem commented

2026-03-15 15:21:01 +03:00

Author

Owner

@sadok-f commented on GitHub (Mar 10, 2024):

@peraronsson can you check the latest benchmark I just released?
https://github.com/flyimg/flyimg/blob/main/docs/full-documentation.md#benchmark

@sadok-f commented on GitHub (Mar 10, 2024): @peraronsson can you check the latest benchmark I just released? https://github.com/flyimg/flyimg/blob/main/docs/full-documentation.md#benchmark

kerem commented

2026-03-15 15:21:06 +03:00

Author

Owner

@peraronsson commented on GitHub (Mar 10, 2024):

@peraronsson can you check the latest benchmark I just released? https://github.com/flyimg/flyimg/blob/main/docs/full-documentation.md#benchmark

This looks really good, and thank you so much for doing this! 😃
If I understand the results correctly, Flyimg is a lot more performant than the "competitors" in the table. Do you agree?

The only thing I am missing now is a url to the image used for the test. If you want to make the tests comparable to the referenced table, I guess you should use the same image size as they did. Maybe you already have - the image name is the same - but it's not apparent without a link to the image (and perhaps size and weight as additional info).

The source image is a photo of Wat Arun (JPEG, 7360x4912, 29MB).

The tools were requested to resize it to fit 500x500.

Thanks again for the quick response. Amazing!

@peraronsson commented on GitHub (Mar 10, 2024): > @peraronsson can you check the latest benchmark I just released? https://github.com/flyimg/flyimg/blob/main/docs/full-documentation.md#benchmark This looks really good, and thank you so much for doing this! 😃 If I understand the results correctly, Flyimg is a lot more performant than the "competitors" in the table. Do you agree? The only thing I am missing now is a url to the image used for the test. If you want to make the tests comparable to the referenced table, I guess you should use the same image size as they did. Maybe you already have - the image name is the same - but it's not apparent without a link to the image (and perhaps size and weight as additional info). > The source image is a [photo of Wat Arun](https://upload.wikimedia.org/wikipedia/commons/b/b7/The_sculptures_of_two_mythical_giant_demons%2C_Thotsakan_and_Sahatsadecha%2C_guarding_the_eastern_gate_of_the_main_chapel_of_Wat_Arun%2C_Bangkok.jpg) (JPEG, 7360x4912, 29MB). > The tools were requested to resize it to fit 500x500. Thanks again for the quick response. Amazing!

kerem commented

2026-03-15 15:21:11 +03:00

Author

Owner

@sadok-f commented on GitHub (Mar 12, 2024):

actually yes I used the same image in the benchmark script to resize it to 500x500.
But the source of the image is local (the benchmark script downloads it and then deletes it).
I'm not sure about other solutions, it seems some of them are using a local version as well.

@sadok-f commented on GitHub (Mar 12, 2024): actually yes I used the same image in the benchmark script to resize it to 500x500. But the source of the image is local (the benchmark script downloads it and then deletes it). I'm not sure about other solutions, it seems some of them are using a local version as well.

kerem commented

2026-03-15 15:21:17 +03:00

Author

Owner

@sadok-f commented on GitHub (Mar 13, 2024):

hey @peraronsson ,
I'm going to close this issue if any help is still needed please feel free to comment and I'll open it.
Thank you again!

@sadok-f commented on GitHub (Mar 13, 2024): hey @peraronsson , I'm going to close this issue if any help is still needed please feel free to comment and I'll open it. Thank you again!