home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.amiga.hardware
- Path: netnews.upenn.edu!dsinc!scala!news
- From: dave.haynie@scala.com (Dave Haynie)
- Subject: Re: EIDE vs. IDE , was: Reports from CeBit
- Sender: news@scala.scala.com (Usenet administrator)
- Message-ID: <1996Mar26.225231.9678@scala.scala.com>
- Date: Tue, 26 Mar 1996 22:52:31 GMT
- Reply-To: dave.haynie@scala.com (Dave Haynie)
- References: <4j0jnq$et9@leol.net-link.net> <4j3es5$u0@kaon.kuai.se>
- Nntp-Posting-Host: gator
- Organization: Scala Computer Television, US Research Center
-
- In <4j3es5$u0@kaon.kuai.se>, flex@kuai.se (Anders Karlsson) writes:
- >In a message of 23 Mar 96 Mike Williams wrote to :
- >
- > MW> Actually I think you're both right. EIDE boards drive 4 drives, but
- > MW> they list it as two controllers.
-
- Right. The basic IDE bus is still the same with EIDE. You have what
- began life as a subset of the ISA bus, put on a connector. The bus
- supports one "master" and one "slave" drive, via a 16-bit data path
- and a small number of address and handshake signals.
-
- The original idea of IDE was a simple one -- rather than waste an ISA
- bus slot with the typical ST-506 hard disk controller, you could
- integrate the controller on the drive itself and just run the ISA bus
- (or at least the pieces needed for the controller) over to the drive.
- A few simple standards were defined for where the disk controller
- would be mapped in ISA bus space and what it looked like, bit-wise,
- and IDE was born.
-
- > MW> On a PC you can disable the secondary controller if there's
- > MW> nothing connected to it.
-
- And in fact, the second controller wasn't new with EIDE either. Way
- back in the dark ages, I had an IDE card controller for an AT Bridge
- Card that supported two (very vanilla) IDE buses. Once folks started
- integrating IDE into the motherboard and SuperI/O chips, there was
- rarely a need for two IDE buses. After all, who could possibly need
- more than two hard disk drives?
-
- The push for multiple IDE buses is based on the use of IDE for CD-ROM,
- which happened to come along around the same time as EIDE.
-
- >Hmm.. That's why SCSI is still the better option.
-
- SCSI is, in general, better. It's also more expensive. And if you're
- counting pennies, you can get a real fast IDE system for less than a
- cheezy SCSI system. The best SCSI adaptors are still faster than the
- best EIDE adaptors. Either will outrun the fastest single drive
- available today, though SCSI is designed to allow interleaved acccess
- when you have multiple drives. So it has an advantage beyond simple
- transfer speeds when you're on a multiple drive, multitasking system.
-
- Of course, low-end systems aren't generally expected to have multiple
- drives. While everyone would like to think their A1200 or next year's
- low-end Power Amiga is a baby SGI or something, it really is a low-end
- computer. You can't expect it to come with every advantage of your
- favorite graphics workstation.
-
- > MW> However, each controller usually supports mode 3 and sometimes
- > MW> mode 4 transfers. I don't have any idea what that means, except
- > MW> that it's quite a bit faster than standard IDE, which I think is
- > MW> mode 2.
-
- The original IDE ran at ISA bus speeds, naturally. When IDE was
- nothing more than an existing ISA bus run through buffers, this made
- all kinds of sense. But as the world moved to systems not centered
- around the ISA bus, IDE was enhanced, yielding EIDE. Actually, there
- are a number of specific enhancements, some of which you can get
- without really being EIDE. One is the addressing problem workaround,
- which allowed addressing of drives beyond the ~500MB limit (yes, it's
- confusing, mainly because both IDE and MS-DOS set drive space limits,
- and what you got on PCs as a maximum drive size was effectively the
- intersection of these two limits). Another are new transfer speed
- "modes", which run the bus faster. After all, once its no longer part
- of the ISA bus proper, an IDE bus shouldn't be stuck at 4MB/s maximum
- speeds. So you get up to around 13.5MB/s transfers across EIDE in mode
- 4. That's of course based on burst out of a drive cache, you're not
- going to find an IDE drive that gives you a sustained transfer at that
- rate.
-
- EIDE is quite usable under a good OS, even a multitasking OS. It can
- also be butt-ugly under a bad driver or OS, as can SCSI. If you know
- you're going to be doing some heavy drive crunching with several hard
- disks going at once, you're going to want SCSI. Ideally, you're going
- to want the new SCSI Fast-20, which once again doubles the synchronous
- rate of SCSI-2, up to 20MB/s on 8-bit SCSI, 40MB/s on 16-bit
- SCSI. Active termination, short runs, and shielded cable
- recommended. I don't know of any Fast-20 implementation on the Amiga
- (though SCSI-2 FS at 16-bit goes just as fast as Fast-20 at 8-bit),
- but they'll probably show up, certainly on high-end Power Amigas at
- least as options (no, I don't know of any specific plans, just that
- given any high-end Power Amiga, Fast-20 SCSI will be so easy to do
- it'll happen).
-
- Dave Haynie | ex-Commodore Engineering | for DiskSalv 3 &
- Sr. Systems Engineer | Hardwired Media Company | "The Deathbed Vigil"
- Scala Inc., US R&D | Ki No Kawa Aikido | info@iam.com
-
- "Feeling ... Pretty ... Psyched" -R.E.M.
-
-