solomon@chaos.utexas.edu (Thomas Solomon) (02/06/91)
I have seen some information about EISA disk controllers. Apparently, this disk controllers can provide transfer rates that are substantially higher than anything that PC's have had before. Austin Computer Systems has an EISA controller for their 486 machine which, they claim, can transfer at rates up to 33 MB/sec (the key words there being "up to"). Does anyone know the scoop about these controllers? Specifically, how fast can they write to disk in a _continuous_ mode. We are interested in extremely fast disk transfers for some image processing that we will be doing. Ideally, we would like to be able to handle 8 MB/sec transfers. This would allow us to record video images in real time from a digitizer without use of a VCR. However, if we can get 4 MB/sec, we may be content with that. Thanks. Tom Solomon solomon@chaos.utexas.edu
phil@brahms.amd.com (Phil Ngai) (02/07/91)
In article <43779@ut-emx.uucp> solomon@chaos.utexas.edu (Thomas Solomon) writes: |Does anyone know the scoop about these controllers? Specifically, how |fast can they write to disk in a _continuous_ mode. We are interested The question is how fast is their disk interface? ST-506, for example is only 5 megabits per second. ESDI is 10, 15 or maybe 20 mbs. IDE and SCSI get around this to some extent by putting a controller on the drive, but you still have a serial path from the read/write heads and the same issues apply except that you get to use Zone Bit Recording. Some are also cranking up the spindle speed to 4400 RPM. Even so, the highest bit rate I've seen is 25 megaBITs/second (and this only on the outer most cylinders). So your CONTINUOUS 4 megabyte/second seems very unlikely. -- Would you trust a government that didn't trust you?
poffen@sj.ate.slb.com (Russ Poffenberger) (02/08/91)
In article <43779@ut-emx.uucp> solomon@chaos.utexas.edu (Thomas Solomon) writes: >I have seen some information about EISA disk controllers. Apparently, >this disk controllers can provide transfer rates that are substantially >higher than anything that PC's have had before. Austin Computer Systems >has an EISA controller for their 486 machine which, they claim, can >transfer at rates up to 33 MB/sec (the key words there being "up to"). > >Does anyone know the scoop about these controllers? Specifically, how >fast can they write to disk in a _continuous_ mode. We are interested >in extremely fast disk transfers for some image processing that we >will be doing. Ideally, we would like to be able to handle 8 MB/sec >transfers. This would allow us to record video images in real time >from a digitizer without use of a VCR. However, if we can get >4 MB/sec, we may be content with that. > This is a very elusive subject. It doesn't matter one bit how fast the controller is if the disk can't handle it. The tranfser rate to/from the disk (for large sequential access so no cache is involved) is purely a function of the disk geometry. It all relates to two things, the number of sectors per track, and the rotational speed of the disks. Most disks run at 3600 RPM, some new ones will soon have 5400. Sectors per track range from < 20 to > 80. Obviously, the drives with the larger number of sectors per track will deliver better performance. Even then, the fastest interface available only reaches 3MB/sec (IPI-2) for single head per surface drives. 6MB/sec for dual head drives. SCSI-2 promises greater performance, but there is little hardware available for this right now. Even so, these are VERY expensive, a disk and controller would cost you (If they were available for PC, whcih they probably aren't) several thousand dollars. Still, the disk is always the limiting factor, not much you can do about this. Don't get caught up in market hype, you will be disappointed. Russ Poffenberger DOMAIN: poffen@sj.ate.slb.com Schlumberger Technologies UUCP: {uunet,decwrl,amdahl}!sjsca4!poffen 1601 Technology Drive CIS: 72401,276 San Jose, Ca. 95110 (408)437-5254
pasquale@sgl (Pasquale Leone) (02/08/91)
There was some talk a while back about a DTC 3280 (16bit) scsi card for the ibm pcat. The article mentioned that this board could be had for $129. This sounds like a very good price for a good scsi board considering I'm currently sufferring with an ST-02. Can anybody confirm this price or maybe steer me to other inexpensive but good scsi cards. pasquale@sgl.ists.ca
src@scuzzy.in-berlin.de (Heiko Blume) (02/14/91)
pcg@cs.aber.ac.uk (Piercarlo Grandi) writes: >Your best chance IMNHO is to get an Adaptec EISA controller and >some nice HP or Imprimis 600MB/1.2GB disk with SCSI-2 and synchronous >mode support. perhaps the wren runner would be more suited, due to it's 10.6ms average seek time. i think they now have ~300 and ~600 MB models. -- Heiko Blume <-+-> src@scuzzy.in-berlin.de <-+-> (+49 30) 691 88 93 public source archive [HST V.42bis]: scuzzy Any ACU,f 38400 6919520 gin:--gin: nuucp sword: nuucp uucp scuzzy!/src/README /your/home
src@scuzzy.in-berlin.de (Heiko Blume) (02/14/91)
poffen@sj.ate.slb.com (Russ Poffenberger) writes: >Even then, the fastest interface available only reaches >3MB/sec (IPI-2) for single head per surface drives. 6MB/sec for dual head >drives. CDC/Imprimis/Seagate's catalog has a parallel transfer disk drive with four x four heads. they claim 9.9MB/s sustained transfer rate with 4K blocks, 12MB/s peak. it has an ISI interface, which might be a problem to get a ISA/EISA bus controller for :-) -- Heiko Blume <-+-> src@scuzzy.in-berlin.de <-+-> (+49 30) 691 88 93 public source archive [HST V.42bis]: scuzzy Any ACU,f 38400 6919520 gin:--gin: nuucp sword: nuucp uucp scuzzy!/src/README /your/home
pcg@cs.aber.ac.uk (Piercarlo Grandi) (02/19/91)
On 13 Feb 91 17:36:36 GMT, src@scuzzy.in-berlin.de (Heiko Blume) said:
src> pcg@cs.aber.ac.uk (Piercarlo Grandi) writes:
pcg> Your best chance IMNHO is to get an Adaptec EISA controller and
pcg> some nice HP or Imprimis 600MB/1.2GB disk with SCSI-2 and synchronous
pcg> mode support.
src> perhaps the wren runner would be more suited, due to it's 10.6ms
src> average seek time. i think they now have ~300 and ~600 MB models.
Nit picking time: Imprimis, Seagate and Wren are really the same thing
now. I was indeed thinking of the Imprimis WREN Roadrunner in my
paragraph above.
More precisely: Seagate is a drive manufacturer, Imprimis is (more or
less) one of their divisions that they bought from CDC, and WREN (I, II,
III, IV, V, VI, and so on until they run out of roman numerals :->) is
the main line of 5.25" drives manufactured and marketed by Seagate's
Imprimis division.
Other suitable drives may be the Fujitsu 2263a and a whole load of
others. But HP and Seagate/Imprimis/WREN are especially highly regarded.
For information on SCSI drives look at the mac and next newsgroups,
where they are constantly debated.
--
Piercarlo Grandi | ARPA: pcg%uk.ac.aber.cs@nsfnet-relay.ac.uk
Dept of CS, UCW Aberystwyth | UUCP: ...!mcsun!ukc!aber-cs!pcg
Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk