Re: Many threads in Linux
- From: Scott Gifford <sgifford@xxxxxxxxxxxxxxxx>
- Date: Thu, 13 Dec 2007 17:59:24 -0500
Eric Sosman <esosman@xxxxxxxxxxxxxxxxxxxx> writes:
Scott Gifford wrote:
[...]
Here are the results I saw:
THREADS VIRT MEM RES MEM REQ/S
------- -------- ------- -----
1000 18m 9m 35218
10000 159m 79m 35216
20000 317m 159m 33659
30000 474m 238m 33733
40000 632m 318m 35183
50000 789m 398m 28864
60000 947m 477m 28018
It does start to slow down a bit, but overall the performance is
not too bad.
"A bit?" It's a twenty percent plunge! Are your friends
hedge fund managers, maybe?
It's about what I expected; my experience with servers has generally
been that throughput decreases under a high enough load. It's
actually slightly better than the "plunge" I saw with a simple
epoll-based echo server I wrote for comparison, although the epoll
server was still about 10% faster overall. See message
<lyve73oxyt.fsf@xxxxxxx> for the numbers and a link to the code:
http://groups.google.com/group/comp.programming.threads/msg/c1af24bd26a4860a
-----Scott.
.
- References:
- Many threads in Linux
- From: Scott Gifford
- Re: Many threads in Linux
- From: David Schwartz
- Re: Many threads in Linux
- From: Scott Gifford
- Re: Many threads in Linux
- From: Marcel Müller
- Re: Many threads in Linux
- From: Scott Gifford
- Re: Many threads in Linux
- From: Eric Sosman
- Re: Many threads in Linux
- From: Scott Gifford
- Re: Many threads in Linux
- From: Scott Gifford
- Re: Many threads in Linux
- From: Eric Sosman
- Many threads in Linux
- Prev by Date: OT: Book
- Next by Date: Re: Design Questions on Termination
- Previous by thread: Re: Many threads in Linux
- Next by thread: Re: Many threads in Linux
- Index(es):
Relevant Pages
|