Thread (3 messages) 3 messages, 2 authors, 2013-10-30

Re: [PATCH v2] can: c_can: Speed up rx_poll function

From: Markus Pargmann <hidden>
Date: 2013-10-30 07:54:57
Also in: linux-can, lkml

Possibly related (same subject, not in this thread)

On Tue, Oct 29, 2013 at 09:24:05AM -0700, Joe Perches wrote:
On Tue, 2013-10-29 at 09:58 +0100, Markus Pargmann wrote:
quoted
On Tue, Oct 29, 2013 at 01:34:48AM -0700, Joe Perches wrote:
quoted
On Tue, 2013-10-29 at 09:27 +0100, Markus Pargmann wrote:
quoted
This patch speeds up the rx_poll function by reducing the number of
register reads.
[]
quoted
125kbit:
  Function                               Hit    Time            Avg             s^2
  --------                               ---    ----            ---             ---
  c_can_do_rx_poll                     63960    10168178 us     158.977 us      1493056 us
With patch:
  c_can_do_rx_poll                     63939    4268457 us     66.758 us       818790.9 us

1Mbit:
  Function                               Hit    Time            Avg             s^2
  --------                               ---    ----            ---             ---
  c_can_do_rx_poll                     69489    30049498 us     432.435 us      9271851 us
With patch:
  c_can_do_rx_poll                    103034    24220362 us     235.071 us      6016656 us
[]
quoted
Yes I just measured the timings again:
[]
quoted
./perf_can_test.sh 125000 30
[]
quoted
  c_can_do_rx_poll                     63941    3764057 us     58.867 us       776162.2 us 
Good, it's slightly faster still.
quoted
./perf_can_test.sh 1000000 30
[]
quoted
  c_can_do_rx_poll                    207109    24322185 us     117.436 us      171469047 us 
[]
quoted
It is interesting that the number of hits for c_can_do_rx_poll is twice as much
as it was with find_next_bit.
How is this possible?  Any idea?
Perhaps the can core received more messages in the previous patch
version while being in the poll function. This way it can handle more
messages without a new interrupt/poll call.

The new patch is faster so it does not receive as many new messages as
the old one, leading to more interrupts and less handled messages per
poll call.

Regards,

Markus Pargmann

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help