Messages in this thread | | | Date | Mon, 2 Jun 1997 12:00:56 -0700 | From | "Leonard N. Zubkoff" <> | Subject | Re: SCSI disks |
| |
Looks like this discussion may have moved to linux-scsi a bit too soon. Here are three relevant messages:
From lnz@dandelion.com Mon Jun 2 10:06:50 1997 Date: Mon, 2 Jun 1997 09:54:04 -0700 From: "Leonard N. Zubkoff" <lnz@dandelion.com> To: linux-scsi@vger.rutgers.edu In-reply-to: <Pine.LNX.3.96.970602101648.700B-100000@max1-36> (message from Edward Welbon on Mon, 2 Jun 1997 11:00:29 -0500 (CDT)) Subject: Re: Why SCSI disks rather than IDE disks? Re: SCSI disks Precedence: bulk
Date: Mon, 2 Jun 1997 11:00:29 -0500 (CDT) From: Edward Welbon <welbon@bga.com>
[snip]
The disk mechanism itself is not at issue as those are in principle identical.
[snip]
This is certainly true, but omits the important fact that *none* of the highest performance disk drives are available with an IDE interface. All the high end disk drives also have sophisticated head motion optimization algorithms which can substantially improve performance when the drive is given multiple commands at a time to work on. The IDE interface has no way of even giving the disk more than one command at a time. IDE looks good on paper for sequential I/O performance, but not for random I/O on multitasking systems. Here's an example of the measured performance of two SCSI disks for 512 byte block random I/O. The Quantum Fireball Tempest is available in both IDE and SCSI; the SCSI interface on this disk is poor, but doesn't affect the random I/O particularly. The Quantum Atlas is SCSI-only and is even the previous generation (I don't have an Atlas II). The Queue Depth is the number of concurrent commands given to the drive.
Queue Depth = 1 Queue Depth = 15 Queue Depth = 31 ================ ================ ================ Fireball Tempest 47 ops/sec 52 ops/sec 50 ops/sec Atlas 71 ops/sec 129 ops/sec 151 ops/sec
As with all performance comparisons, what's crucial is the performance on the workload that's important to the people making the hardware selection. If your disk usage is not heavily multitasking and you don't use more than one disk per channel concurrently, then IDE should equal 2nd-tier SCSI in raw performance, if not in robustness (can IDE drives automatically reassign a bad block if an error occurs? SCSI can). For any serious multitasking system with more disks than IDE channels, SCSI is a big win even over bus-mastering EIDE.
Leonard
From welbon@bga.com Mon Jun 2 10:34:00 1997 X-Authentication-Warning: max1-26: welbon owned process doing -bs Date: Mon, 2 Jun 1997 12:32:59 -0500 (CDT) From: Edward Welbon <welbon@bga.com> X-Sender: welbon@max1-36 To: "Leonard N. Zubkoff" <lnz@dandelion.com> cc: linux-scsi@vger.rutgers.edu Subject: Re: Why SCSI disks rather than IDE disks? Re: SCSI disks In-Reply-To: <199706021654.JAA17464@dandelion.com> Subliminal-Message: Run linux now
On Mon, 2 Jun 1997, Leonard N. Zubkoff wrote:
> I said:
> > The disk mechanism itself is not at issue as those are in principle > > identical.
> This is certainly true, but omits the important fact that *none* of the > highest performance disk drives are available with an IDE interface. > All the high end disk drives also have sophisticated head motion > optimization algorithms which can substantially improve performance when > the drive is given multiple commands at a time to work on.
I stand corrected. I had completely fogotten about this very important fact. I wonder if one could simulate this aspect of SCSI function with a sufficiently intelligent IDE controller (though at that point, there would be significantly less cost advantage to IDE).
Ed Welbon; welbon@bga.com;
From lnz@dandelion.com Mon Jun 2 11:34:39 1997 Date: Mon, 2 Jun 1997 11:19:16 -0700 From: "Leonard N. Zubkoff" <lnz@dandelion.com> To: welbon@bga.com CC: linux-scsi@vger.rutgers.edu In-reply-to: <Pine.LNX.3.96.970602122825.700I-100000@max1-36> (message from Edward Welbon on Mon, 2 Jun 1997 12:32:59 -0500 (CDT)) Subject: Re: Why SCSI disks rather than IDE disks? Re: SCSI disks Precedence: bulk
Date: Mon, 2 Jun 1997 12:32:59 -0500 (CDT) From: Edward Welbon <welbon@bga.com>
On Mon, 2 Jun 1997, Leonard N. Zubkoff wrote:
> This is certainly true, but omits the important fact that *none* of the > highest performance disk drives are available with an IDE interface. > All the high end disk drives also have sophisticated head motion > optimization algorithms which can substantially improve performance when > the drive is given multiple commands at a time to work on.
I stand corrected. I had completely fogotten about this very important fact. I wonder if one could simulate this aspect of SCSI function with a sufficiently intelligent IDE controller (though at that point, there would be significantly less cost advantage to IDE).
I'd say that's very unlikely, as I doubt the IDE interface specification feeds back enough information about the rotational position of the platter. For that matter, all modern disks divide the disk platter radially into zones with different numbers of sectors per track, and I doubt that information is available either. Without it, one cannot even determine whether sector N and sector M are on the same track. It's impossible to do optimal seek optimization without a pretty detailed model of the disk layout and capabilities.
Leonard
| |