Re: smrproxy v2

Liste des GroupesRevenir à cl c++ 
Sujet : Re: smrproxy v2
De : chris.m.thomasson.1 (at) *nospam* gmail.com (Chris M. Thomasson)
Groupes : comp.lang.c++
Date : 25. Nov 2024, 23:12:37
Autres entêtes
Organisation : A noiseless patient Spider
Message-ID : <vi2sol$31jrd$1@dont-email.me>
References : 1 2 3 4 5 6 7 8
User-Agent : Mozilla Thunderbird
On 11/25/2024 3:31 AM, jseigh wrote:
On 11/24/24 19:47, Chris M. Thomasson wrote:
On 11/24/2024 4:09 PM, jseigh wrote:
 ...
 
>
Actually, I remember way back where a scenario that had a lot of writes would start to mess with a deferred reclamation wrt a polling thread type of “scheme”. Too many deferred nodes would start to "pile up". Basically, the single polling thread was having trouble keeping up with all of them. The interlocked versions seemed to perform sort of "better" in a sense during periods that had a lot of frequent “writes”. Of course clever use of node caches helps heavy write periods. Anyway, some of the tests just used a mutex for writes, others used lock-free and would generate high loads of them that would push and pop nodes and defer them to the poll thread to test how much load it (poll thread) could take.
  Using deferred reclamation to implement a lock-free queue will slow
things down.  Also cache line thrashing will limit how fast queue
operations will go no matter what.
 Better to parallelize.  You can take a really bad lock-free queue
implementation, parallelize it, and then tell everyone how fast
your queue is.
LOL! Yeah. In layers. Per thread queues, per cpu queues and some hashed global ones. Then we get to play with single consumer/producer, multi consumer single producer, and all that jazz... ;^)
Actually it was a RCU algorithm wrt deferred reclamation that worked great with read mostly. Then exposing it to a lot of writes sort of messed around with it, in a bad way. ;^)
The writes were able to swamp things under heavy artificial load.

 Single queue at 1 op/sec.
Parallel queue on 1000000 processors, 1000000 op/sec.

Date Sujet#  Auteur
17 Oct 24 * smrproxy v266jseigh
17 Oct 24 +* Re: smrproxy v241Chris M. Thomasson
17 Oct 24 i`* Re: smrproxy v240jseigh
18 Oct 24 i `* Re: smrproxy v239Chris M. Thomasson
18 Oct 24 i  +* Re: smrproxy v22Chris M. Thomasson
18 Oct 24 i  i`- Re: smrproxy v21Chris M. Thomasson
18 Oct 24 i  `* Re: smrproxy v236jseigh
25 Oct 24 i   `* Re: smrproxy v235Chris M. Thomasson
25 Oct 24 i    `* Re: smrproxy v234jseigh
27 Oct 24 i     `* Re: smrproxy v233Chris M. Thomasson
27 Oct 24 i      `* Re: smrproxy v232jseigh
27 Oct 24 i       +* Re: smrproxy v228Chris M. Thomasson
28 Oct 24 i       i`* Re: smrproxy v227jseigh
28 Oct 24 i       i `* Re: smrproxy v226Chris M. Thomasson
28 Oct 24 i       i  `* Re: smrproxy v225jseigh
28 Oct 24 i       i   `* Re: smrproxy v224Chris M. Thomasson
28 Oct 24 i       i    +- Re: smrproxy v21Chris M. Thomasson
29 Oct 24 i       i    `* Re: smrproxy v222jseigh
29 Oct 24 i       i     +* Re: smrproxy v23Chris M. Thomasson
29 Oct 24 i       i     i`* Re: smrproxy v22jseigh
29 Oct 24 i       i     i `- Re: smrproxy v21Chris M. Thomasson
29 Oct 24 i       i     +* Re: smrproxy v25Chris M. Thomasson
29 Oct 24 i       i     i`* Re: smrproxy v24jseigh
29 Oct 24 i       i     i `* Re: smrproxy v23Chris M. Thomasson
30 Oct 24 i       i     i  `* Re: smrproxy v22jseigh
1 Nov 24 i       i     i   `- Re: smrproxy v21Chris M. Thomasson
29 Oct 24 i       i     `* Re: smrproxy v213Chris M. Thomasson
29 Oct 24 i       i      `* Re: smrproxy v212Chris M. Thomasson
30 Oct 24 i       i       `* Re: smrproxy v211jseigh
4 Nov 24 i       i        `* Re: smrproxy v210Chris M. Thomasson
4 Nov 24 i       i         `* Re: smrproxy v29jseigh
4 Nov 24 i       i          +* Re: smrproxy v23Muttley
4 Nov 24 i       i          i`* Re: smrproxy v22Chris M. Thomasson
9 Nov 24 i       i          i `- Re: smrproxy v21Chris M. Thomasson
12 Dec 24 i       i          `* Re: smrproxy v25Chris M. Thomasson
12 Dec 24 i       i           `* Re: smrproxy v24jseigh
12 Dec 24 i       i            `* Re: smrproxy v23Chris M. Thomasson
13 Dec 24 i       i             `* Re: smrproxy v22jseigh
26 Dec 24 i       i              `- Re: smrproxy v21Chris M. Thomasson
28 Oct 24 i       `* Re: smrproxy v23Chris M. Thomasson
28 Oct 24 i        +- Re: smrproxy v21Chris M. Thomasson
28 Oct 24 i        `- Re: smrproxy v21Chris M. Thomasson
29 Oct 24 +* Re: smrproxy v24Chris M. Thomasson
30 Oct 24 i+- Re: smrproxy v21Chris M. Thomasson
30 Oct 24 i`* Re: smrproxy v22Chris M. Thomasson
30 Oct 24 i `- Re: smrproxy v21Chris M. Thomasson
30 Oct 24 +* Re: smrproxy v22Chris M. Thomasson
30 Oct 24 i`- Re: smrproxy v21jseigh
2 Nov 24 +- Re: smrproxy v21Chris M. Thomasson
21 Nov 24 `* Re: smrproxy v217jseigh
23 Nov 24  `* Re: smrproxy v216jseigh
23 Nov 24   +* Re: smrproxy v23Chris M. Thomasson
23 Nov 24   i`* Re: smrproxy v22jseigh
24 Nov 24   i `- Re: smrproxy v21Chris M. Thomasson
24 Nov 24   `* Re: smrproxy v212jseigh
25 Nov 24    +* Re: smrproxy v27Chris M. Thomasson
25 Nov 24    i`* Re: smrproxy v26jseigh
25 Nov 24    i +* Re: smrproxy v22Chris M. Thomasson
25 Nov 24    i i`- Re: smrproxy v21Chris M. Thomasson
25 Nov 24    i `* Re: smrproxy v23Chris M. Thomasson
25 Nov 24    i  `* Re: smrproxy v22jseigh
25 Nov 24    i   `- Re: smrproxy v21Chris M. Thomasson
27 Nov 24    `* Re: smrproxy v24jseigh
9 Dec 24     `* Re: smrproxy v23jseigh
11 Dec 24      `* Re: smrproxy v22jseigh
12 Dec 24       `- Re: smrproxy v21Chris M. Thomasson

Haut de la page

Les messages affichés proviennent d'usenet.

NewsPortal