| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 2891.1 | Ultra DMA/33 | TLE::INGRAM | oops | Fri Jan 24 1997 17:03 | 9 |
|
It's officially known as Ultra DMA/33. For more info, check out
http://www.via.com.tw/dma33.htm for starters.
Essentially, it's a protocol which doubles the existing 16.6MB/sec
maximum ATA data rate and adds CRC for data integrity.
Larry
|
| 2891.2 | Where's The Beef? | WRKSYS::TATOSIAN | The Compleat Tangler | Fri Jan 24 1997 18:33 | 17 |
| I'd hazard to say that it's a (somewhat feeble) attempt to avoid
complete obsolescence in the face of UltraSCSI...
Still limited to a pair of drives, no pended operations/command
queueing/seek optimization/etc, it's hard to see that there's any
significant advantage to the user once the smoke dissipates. Even in
the cases provided at Quantum's web site (which you'll note are
contrived to make Ulta DMA33 look its best) one could make an easy
argument that the "buzz" amounts to little net gain.
I'd rather see decent size buffers in the drive than cranking up the
transfer rate. A 64KB buffer in a 4GB drive seems stupid on the face of
it...
I'd bin the concept under "MarketWare" and move to UWSCSI...
/dave (But then, I could be wrong ;^)
|
| 2891.3 | exactly backwards | CSC32::J_MANNING | | Sat Jan 25 1997 14:53 | 14 |
|
In fact, if you look at the web site mentioned in .1, it says:
The protocol also allows has disk drive manufacturers to reduce data
transfer buffer sizes because more data can be passed to the host
computer in less time.
This is exactly opposite of what they should do. Even 7200 RPM drives
cannot deliver data at more than 6-7 MB/sec sustained so the only
advantage to this new protocol would be for burst tranfers from the
on-disk cache.
John
|
| 2891.4 | | PCBUOA::KRATZ | | Mon Jan 27 1997 10:14 | 2 |
| Like it or not, if Intel builds it in to their chipsets, it will
come...
|
| 2891.5 | | WRKSYS::TATOSIAN | The Compleat Tangler | Mon Jan 27 1997 11:17 | 1 |
| No argument there - but that doesn't mean one must actually use it ;^)
|
| 2891.6 | | DANGER::ARRIGHI | Life is an else-if construct | Mon Jan 27 1997 15:41 | 4 |
| If you all can convince the drive manufacturers to price their scsi
drives like their ata drives, then I'll agree with you.
Tony (who only needs 2 hard drives on his home machine.
|
| 2891.7 | FIC response | BRITE::FYFE | Use it up, wear it out, make it do, or do without. | Wed Jan 29 1997 08:04 | 36 |
|
Subj: PA-2005/CyrixPR200+ PCI bus out of spec (37.5mhz); Why?
Dear Sir,
Thank you for raising this question with FIC Tech Support. I have
now put some answers re this issue on our FAQ at :
http://rd.fic.com.tw/faq/faqmb.htm in the Cyrix processor
section.
The reason why the PCI bus runs out of spec is because when the
system bus is running at 75 MHz, required for the 6x86 PR200+,
chipset only supports a 37.5/75MHz clock ratio.
If you use an older type peripheral that doesn't support the
37.5MHz bus frequency you may not get optimum performance from
it. Therefore we recommend using the newer PCI peripherals which
support the 37.5MHz PCI bus. I have posted the list of
recommended PCI peripherals in our FAQ section.
Future FCI solutions based on upcoming PCIsets will be able to
support both PCI bus frequencies with the 6x86 PR200+. Therefore
if you are wanting to run old PCI peripherals you may consider
waiting two months or so for these future solutions. Otherwise if
you want a top peforming 6x86 PR200+ board now, then the PA-2005
is an excellent choice.
I hope I have answered all you questions. If you have any more
please get back in contact with us at Tech Support.
Sincerely,
Fleur Shand
Marketing
|
| 2891.8 | In case you were wondering ... | WRKSYS::INGRAHAM | Andy | Wed Jan 29 1997 17:53 | 4 |
| The previous reply (-.1) refers to a discussion in note 2876.
Did you type in their response yourself? Did they call themselves
both FIC and FCI, or is that just a typo?
|
| 2891.7 | read this last night .... | BRITE::FYFE | Use it up, wear it out, make it do, or do without. | Fri Feb 07 1997 09:45 | 2 |
|
the Intel TX chipset does/will_have ATA-33 extensions ....
|