Messages in this thread | | | Date | Thu, 25 Jun 1998 19:51:17 +0200 (CEST) | From | Rik van Riel <> | Subject | RE: Thread implementations... |
| |
On Thu, 25 Jun 1998, Amsden, Zachary wrote:
> Possibly. But according to Squid's own doc, the primary problems > in Squid performance are > > 1) Not enough memory
Can indeed be alleviated by sendfile()...
> 2) Too slow disks
With a filesystem that can't handle the extreme dir sizes that Squid uses. ReiserFS or another advanced filesystem will speed up this bottleneck by more than just a considerable amount...
> which as far as I can see aren't helped by sendfile. They also > say "CPU limitations are rarely encountered except in very large > caches".
This situation is also better handled by tree-based filesystems and better buffer/cache administration.
> But too slow disks and not enough memory are directly > related problems. If the kernel has frequently used > disk files in the buffer cache, it still needs to copy > the data to the user space server, then copy it back > out onto a socket. With sendfile(), it can directly > send data from the buffer cache without any user space > involvement. In addition, we no longer need user > space buffers in the server, cutting down on useless > "memory duplication", allowing for more buffer cache.
This is very very true and might be able to speed up large servers quite a lot.
Rik. +-------------------------------------------------------------------+ | Linux memory management tour guide. H.H.vanRiel@phys.uu.nl | | Scouting Vries cubscout leader. http://www.phys.uu.nl/~riel/ | +-------------------------------------------------------------------+
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |