Re: How to make Linux scale up WRT bandwidth and size?

Philip Blundell (pb@nexus.co.uk)
Fri, 30 Apr 1999 10:53:38 +0100

Messages sorted by: [ date ][ thread ][ subject ][ author ]
Next message: Trond Myklebust: "Re: Horrible NFS performace"
Previous message: Alexander Viro: "Re: [PATCH] missing lock_kernel() and assorted races"
Maybe in reply to: David Kastrup: "How to make Linux scale up WRT bandwidth and size?"

In message <19990430112758.B4502@boole.suse.de>, "Dr. Werner Fink" writes:
>On Thu, Apr 29, 1999 at 05:25:21PM -0400, Nat Lanza wrote:
>> claims a transfer rate of 18-28 Mbytes/sec.
>
>Hmmm ... how long this stays at this value for a 2 Gig file
>read or write at *once*?

If you are reading or writing in a straight line they can probably sustain
that throughput indefinitely. The only thing that causes it to drop is seek
time.

I benchmarked some earlier Cheetah drives about 6-9 months ago. Reading and
writing 256k blocks randomly distributed over the disk they managed a
sustained throughput of a little under 10MB/s. These are quite fast drives.

-
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/

Next message: Trond Myklebust: "Re: Horrible NFS performace"
Previous message: Alexander Viro: "Re: [PATCH] missing lock_kernel() and assorted races"
Maybe in reply to: David Kastrup: "How to make Linux scale up WRT bandwidth and size?"