lkml.org 
[lkml]   [1999]   [Jun]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: 2.3.7: where are the gains?

On Tue, 22 Jun 1999, Richard Gooch wrote:

> Hi, all. A quick test compiling the 2.3.7 kernel tree shows 2.3.7
> perform slightly worse than 2.3.5 for make dep. PII 233 with IDE disc.

compiling the kernel is very CPU-intensive, about 95% of the time is spent
in user-space.

nevertheless the speedup does seem to show up:

> 2.3.7:
> time make zImage
> 291.870u 17.570s 5:14.29 98.4% 0+0k 0+0io 192480pf+0w
^^^^^^^

> 2.3.5:
> time make zImage
> 291.590u 18.810s 5:14.38 98.7% 0+0k 0+0io 194125pf+0w
^^^^^^^

an about 10% speedup for the kernel-only component. Considering that
compiling the kernel is doing mostly reads (well it writes object files
but the source is bigger, unless you compile with -g), and is accessing
files - this is not bad at all i think.

even for pure single-threaded cached writes the speedup is rather visible:

2.2.10 (this is essentially 2.3.6 wrt. file IO)

[root@moon fileben]# ./fileben FILE 60000 1 2 1
WRITE: 150.27 MB/sec

2.3.8-pre2:

[root@moon fileben]# ./fileben FILE 60000 1 2 1
WRITE: 249.58 MB/sec

a 66% speedup. Of course it depends on the application - if the
application is purely doing reads then there will be no speedup, reads
were pretty much close to the physical limit in previous kernels already.

-- mingo


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:52    [W:0.156 / U:0.248 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site