On 17 Dec 1999, Marc SCHAEFER wrote:
> Yes. However you will note that a true hardware RAID controller will
> normally lower your CPU usage compared to a software RAID. However,
> main CPU is nowadays far cheaper than specialized CPUs like you can
> find on hardware RAID PCI adapters ...
>
Especially because during the idle time of disks you can still use the
main CPU whereas the hardware raid CPU is completely useless I guess.
> Are you sure you run U2LVD ? Look at the dmesg output for the aic7xxx
> driver. Is there an archive for linux-raid somewhere ? I posted a few
> benchmarks made on upto 8 Seagate Cheetah with RAID5 and RAID0, changing
> things such as splitting in many host adapters, etc. It looks like
> you can't get very good performance on U2LVD with more than three
> Cheetah per channel.
How can I tell from /var/log/messages whether this is actually using
U2LVD? It tells me
(scsi0) <Adaptec AHA-294X Ultra2 SCSI host adapter> found at PCI 20/0
(scsi0) Wide Channel, SCSI ID=7, 32/255 SCBs
(scsi0) Downloading sequencer code... 374 instructions downloaded
scsi0 : Adaptec AHA274x/284x/294x (EISA/VLB/PCI-Fast SCSI) 5.1.20/3.2.4
<Adaptec AHA-294X Ultra2 SCSI host adapter>
> What disk is that ? I was thinking the 18 GB IBMs were 10'000 RPM, around
> 18-24 MByte/s. This looks more like a 7200 RPM drive of older generation
> (IBM DDRS can do 12-13 MByte/s).
You are right those are 7200RPM disks.
>
> Or it could be fs overhead: are you using mke2fs -b 4096 ?
We used a blocksize of 4096B and a chunksize of 32kB. Sometimes you can
see that people suggest a chunksize of 128kB. Is that actually helping?
>
> Or it could be that your partition is at the end of the disk ?
All the partitions were last on the disk (that means the last 8GB of the
18GB drives) but the disk was empty during the test so the files will
probably be somewhat in the center.
>
> PS: how much memory do you have ? With 256 MB memory I like to benchmark
> with 2 GB files.
>
Memory size is 512MB. First I tried with 256MB, 512MB, 1GB, and 2Gb files,
and didn't see a significant change from 1GB to 2GB so I assume memory
cache/buffers are saturated enough to get correct results.
Klaus
>>>>>>>>>>>>> new areacode 516 -> 631 <<<<<<<<<<<<<
+#$*&^$>@$+)$@(*&@$>@+@)$)(&*@$>@$+@$)_*@$>@$+@$*)@>@$+_)@<@)(*$^@&@>?@!+@)$)(
Klaus Schroer [EMAIL PROTECTED]
Biology Department
Brookhaven Nat'l Lab phone: ++1 631 344 7306
Upton NY 11973 fax: ++1 631 344 3407
