We seem to be seeing that both reiserfs and ext2 got slower in pre5, suggesting
that something in VFS may need inspection.
Since I usually am eager to learn of any performance losses in code I write, I
have forwarded this to linux-kernel on the assumption that others are the same.
Anyone know the reason for this offhand?
Hans
attached mail follows:
Hans Reiser wrote:
>
> Did we lose performance in our most recent version? Can you compare the most
> recent version with the benchmarks of the versions shortly after you added your
> speedup? Or is it that a test condition changed? Or is it that I imagine that
> we used to be faster at creates and writes?
>
> Please analyze this for your lazy boss.
>
> Hans
Hi !
Dbench results :
NEW OLD
linux-2.3.99-pre5 linux-2.3.99-pre3
# dbench 5
reiserfs-3.6.4 : 33.6730 MB/sec 32.6549 MB/sec
ext2 : 30.2088 MB/sec 28.1817 MB/sec
# dbench 10
reiserfs-3.6.4 : 24.3463 MB/sec 25.4903 MB/sec
ext2 : 19.0197 MB/sec 19.6588 MB/sec
# dbench 20
reiserfs-3.6.4 : 15.8885 MB/sec 17.5582 MB/sec
ext2 : 12.8872 MB/sec 14.2822 MB/sec
# dbench 50
reiserfs-3.6.4 : 5.74655 MB/sec 5.80893 MB/sec
ext2 : 4.98903 MB/sec 5.02426 MB/sec
Dual Celeron-500, 128 MB RAM, 8 GB scsi HDD
New results for pre5 are a bit worse for reiserfs and for ext2 too.
Reiserfs faster in both cases.
Best,
Yura.
-
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/
This archive was generated by hypermail 2b29 : Sun Apr 23 2000 - 21:00:12 EST