mirage-unix: increase TCP receive buffer size #510
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
On `AMD Ryzen 9 7950X` using Fedora39 and OCaml 4.14.1 testing with iperf on localhost, with 4K buffer: ``` [ 1] 0.00-10.00 sec 31.1 GBytes 26.7 Gbits/sec ``` With 64KiB buffer: ``` [ 1] 0.00-10.00 sec 49.9 GBytes 42.9 Gbits/sec ``` The max size for various TSO/GSO/GRO packet offloads in Linux is 64KiB, so going beyond that is unlikely to be useful [1]. [1]: although there are plans for larger than 64KiB offloads https://netdevconf.info/0x15/slides/35/BIG%20TCP.pdf Signed-off-by: Edwin Török <[email protected]>
|
(this won't necessarily help |
|
Thanks, this looks reasonable to me. |
hannesm
added a commit
to hannesm/opam-repository
that referenced
this pull request
Mar 26, 2024
CHANGES: * TCP: add `src : flow -> ipaddr * int`, implemented by `getsockname` on unix (mirage/mirage-tcpip#511 @hannesm) * TCP unix stack: increase TCP buffer size (was 4096, is now 65536) (mirage/mirage-tcpip#510 @edwintorok) * TCP: adapt to mirage-flow 4.0: add ``val shutdown : flow -> [ `read | `write | `read_write ] -> unit Lwt.t`` (mirage/mirage-tcpip#512 @hannesm, review by @djs55)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Using the
iperfcode fromtest_iperf.ml(which seems to be compatible withiperf2, but notiperf3) I noticed withstracethatreadsystem calls use a 4KiB buffer.Tweaking this resulted in a significant increase in speed, at least on localhost.
On
AMD Ryzen 9 7950Xusing Fedora39 and OCaml 4.14.1 testing with iperf on localhost, with 4K buffer:With 64KiB buffer:
The max size for various TSO/GSO/GRO packet offloads in Linux is 64KiB, so going beyond that is unlikely to be useful [1].
OCaml itself would also use a 64KiB buffer in the Unix module.
[1]: although there are plans for larger than 64KiB offloads https://netdevconf.info/0x15/slides/35/BIG%20TCP.pdf