Subject: Re: [linux-audio-user] Ardour, Jack, and 2.6 kernels + XRuns with the Audiophile 24/96 M-Audio
From: Steve Harris (S.W.Harris_AT_ecs.soton.ac.uk)
Date: Tue Jun 01 2004 - 12:11:38 EEST
On Tue, Jun 01, 2004 at 01:25:28 -0700, Malcolm Baldridge wrote:
> I've heard of PCI bus greedy graphics cards (in the pre-AGP era), but
> bus-greedy SCSI controllers take the cake.
*shrug* its a symptom of the same thing - graphics people think video is
the most important, storage people think its disk thorughput, but we
know its audio latency ;)
In general, people who add SCSI controllers to thier systems probably are
mostly worried about disk i/o (thats when I use SCSI, certainly). In
audio workstations disk i/o is not generally a problem though, in my
experience.
>From the SCSI controller manufacturers p.o.v it makes perfect sense -
holding onto the bus for a bit longer probably gives them a few percent
advantage over the competetitors, and most server motherboards I've seen
in the last few years seem to have one PCI bus per slot, more or less.
Another potential problem with SCSI is that the disks tend to be tuned for
server applications, where flat throughput is the critical thing - eg.
using larger ordering buffers to get the maximum throughput for each
traversal of the platters. IDE and SATA drives tend to be tuned more for
workstation performance where latency is more significant (the user doesnt
want to wait 30secs for the word document to load because chkdisk is
running, even though the overall throughput is higher). In practice it may
not matter as the audio i/o load is pretty low, but its something to bear
in mind.
- Steve
This archive was generated by hypermail 2b28 : Tue Jun 01 2004 - 12:33:25 EEST