public inbox for [email protected]
 help / color / mirror / Atom feed
* [PATCH 5.14 00/12] for-next optimisations
@ 2021-06-14 22:37 Pavel Begunkov
  2021-06-14 22:37 ` [PATCH 01/12] io_uring: keep SQ pointers in a single cacheline Pavel Begunkov
                   ` (13 more replies)
  0 siblings, 14 replies; 15+ messages in thread
From: Pavel Begunkov @ 2021-06-14 22:37 UTC (permalink / raw)
  To: Jens Axboe, io-uring

There are two main lines intervened. The first one is pt.2 of ctx field
shuffling for better caching. There is a couple of things left on that
front.

The second is optimising (assumably) rarely used offset-based timeouts
and draining. There is a downside (see 12/12), which will be fixed
later. In plans to queue a task_work clearing drain_used (under
uring_lock) from io_queue_deferred() once all drainee are gone.

nops(batch=32):
    15.9 MIOPS vs 17.3 MIOPS
nullblk (irqmode=2 completion_nsec=0 submit_queues=16), no merges, no stat
    1002 KIOPS vs 1050 KIOPS

Though the second test is very slow comparing to what I've seen before,
so might be not represantative.

Pavel Begunkov (12):
  io_uring: keep SQ pointers in a single cacheline
  io_uring: move ctx->flags from SQ cacheline
  io_uring: shuffle more fields into SQ ctx section
  io_uring: refactor io_get_sqe()
  io_uring: don't cache number of dropped SQEs
  io_uring: optimise completion timeout flushing
  io_uring: small io_submit_sqe() optimisation
  io_uring: clean up check_overflow flag
  io_uring: wait heads renaming
  io_uring: move uring_lock location
  io_uring: refactor io_req_defer()
  io_uring: optimise non-drain path

 fs/io_uring.c | 226 +++++++++++++++++++++++++-------------------------
 1 file changed, 111 insertions(+), 115 deletions(-)

-- 
2.31.1


^ permalink raw reply	[flat|nested] 15+ messages in thread

end of thread, other threads:[~2021-06-15 21:39 UTC | newest]

Thread overview: 15+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2021-06-14 22:37 [PATCH 5.14 00/12] for-next optimisations Pavel Begunkov
2021-06-14 22:37 ` [PATCH 01/12] io_uring: keep SQ pointers in a single cacheline Pavel Begunkov
2021-06-14 22:37 ` [PATCH 02/12] io_uring: move ctx->flags from SQ cacheline Pavel Begunkov
2021-06-14 22:37 ` [PATCH 03/12] io_uring: shuffle more fields into SQ ctx section Pavel Begunkov
2021-06-14 22:37 ` [PATCH 04/12] io_uring: refactor io_get_sqe() Pavel Begunkov
2021-06-14 22:37 ` [PATCH 05/12] io_uring: don't cache number of dropped SQEs Pavel Begunkov
2021-06-14 22:37 ` [PATCH 06/12] io_uring: optimise completion timeout flushing Pavel Begunkov
2021-06-14 22:37 ` [PATCH 07/12] io_uring: small io_submit_sqe() optimisation Pavel Begunkov
2021-06-14 22:37 ` [PATCH 08/12] io_uring: clean up check_overflow flag Pavel Begunkov
2021-06-14 22:37 ` [PATCH 09/12] io_uring: wait heads renaming Pavel Begunkov
2021-06-14 22:37 ` [PATCH 10/12] io_uring: move uring_lock location Pavel Begunkov
2021-06-14 22:37 ` [PATCH 11/12] io_uring: refactor io_req_defer() Pavel Begunkov
2021-06-14 22:37 ` [PATCH 12/12] io_uring: optimise non-drain path Pavel Begunkov
2021-06-15 12:27 ` [PATCH 5.14 00/12] for-next optimisations Pavel Begunkov
2021-06-15 21:38 ` Jens Axboe

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox