On Jul 04 2019, Daniel Jagszent <[email protected]> wrote:
> Hello Nikolaus,
>
> currently S3QL uses a fixed buffer size of 64KB. With that buffer size
> I can get upload speeds of 100 MBit/s. Not that bad. But when I
> increase that constant from 64KB to e.g. 4MB the same setup can
> saturate a 1GBit/s network connection.
>
> Do you have any reservations against making BUFSIZE configurable with
> an extra argument to most of the command line tools so that – if one
> chooses – can tailor the BUFSIZE to the system/use case.
>
> I just wanted to ask in advance before going thru the hassle of
> creating a pull request.
Thanks for checking! My immediate question is: is there a reason to not
just bump the hardcoded buffer size?
Adding an option typically means that 50% of the people go with the
default (even when it's sub-optimal), 40% pick a value that's even
worse, and 10% actually benefit. So I'd like to avoid options whenever
possible.
Did you check other buffer sizes or just 4 MB? There's probably an
optimum value that we could pick.
Best,
-Nikolaus
--
GPG Fingerprint: ED31 791B 2C5C 1613 AF38 8B8A D113 FCAC 3C4E 599F
»Time flies like an arrow, fruit flies like a Banana.«
--
You received this message because you are subscribed to the Google Groups
"s3ql" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/s3ql/87muhurzjl.fsf%40thinkpad.rath.org.
For more options, visit https://groups.google.com/d/optout.
