On Fri, Aug 20, 1999 at 07:09:06PM -0400, James Manning wrote:
> hw_raid0_4+4
> -------Sequential Output-------- ---Sequential Input-- --Random--
> -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
> MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
> 2047 20878 92.6 51166 40.8 21132 46.8 25705 94.6 49572 70.6 597.1 6.3
>
> hw_raid0_8_on_1
> -------Sequential Output-------- ---Sequential Input-- --Random--
> -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
> MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
> 2047 21774 98.5 29975 24.0 13814 30.4 25131 92.2 27747 39.1 607.1 5.6
So here is what I got:
hw_raid0_8_on_1
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
2047 6589 98.9 25846 22.9 13467 27.3 6944 99.1 44127 37.6 430.0 7.1
hw_raid0_3+3+2
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
2047 6577 98.7 26098 23.1 14235 28.4 6867 98.0 49663 42.1 431.0 7.4
hw_raid5_8_on_1
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
2047 6593 99.0 23438 20.7 11830 23.7 6856 97.8 44106 38.7 412.5 5.8
hw_raid5_3+3+2
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
2047 6573 98.6 22023 19.4 12216 24.5 6939 99.0 50148 44.5 395.0 7.3
caveat: those runs were done weeks ago, just once.
That was before I started taking best-of-n runs.
Your 1 bus results are much worse than mine for reading,
but your 2 bus results are much better than mine for writing.
I don't know what makes for the big differences.
You have a quad Xeon in SMP mode (?).
I have a dual P2/400 in uniprocessor mode.
You have ?? MB ram.
I have 512MB ram.
You have Cheetahs.
I have IBM Ultrastar 18ZX.
We both have DAC1164P w/ 32MB ram.
Yours is firmware version ??
Mine has firmware 5.07-0-79, Stripe Size: 64KB, Segment Size: 8KB, WB cache
You used the DAC960 driver 2.2.2.
I used the DAC960 driver 2.2.1.
You used Linux kernel 2.x.x.
I used Linux kernel 2.2.10 + no patches (well, ok, tulip driver v.91).
I think my raid5 results maintain
Any ideas?
Jan Edler
NEC Research Institute
