| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 99.1 | 2.1 vs 2.0? what are differences | BBPBV1::WALLACE | john wallace @ bbp. +44 860 675093 | Tue Apr 08 1997 16:48 | 11 |
| See 134 for the V2.1 stuff (answer = "not yet").
Actually, when I did some research on this for a customer a while ago,
I thought I was given the answer "2.1 is the same functionally as 2.0
but with minor editorial improvements in wording etc to resolve
ambiguities and such".
I can't find where I put that info. Am I misremembering ?
regards
john
|
| 99.2 | | CIMCAD::PIERSON | | Tue Apr 08 1997 17:23 | 16 |
| welllllllll.
Coupla different things are getting blurred together, here.
134.4(?) discusses _backplanes_.
.0 inquires about SBC's, specifically, about the EBM3x-px, aka
Rushmore. As near as i can tell, Ebm3x-px IS PCI 2.1 compliant...
There ARE technical Differences between PCI 2.0 and PCI 2.1, tho they
are some where between oscure and abstruse. (eg: I dunno). However
Digital Semiconductor saw fit to sell separate devices for each, so
there IS a difference....
regards
dwp
|
| 99.3 | guess i wasn't paying enough attention... | BBPBV1::WALLACE | john wallace @ bbp. +44 860 675093 | Tue Apr 08 1997 18:05 | 2 |
| Ooops. Sorry (wish I could remember my sources for "2.0 = 2.1"; maybe
Laurie Pegrum's chips seminar in Munich?).
|
| 99.4 | | CIMCAD::PIERSON | | Tue Apr 08 1997 20:48 | 10 |
| Checked with the nearest wizard. He sez:
>Dave,
>Most of the differences are editorial. The only substantive changes I'm aware
>of are the 66 MHz timing values and a rule that a Target must respond with
>data in 16 clock ticks or terminate the transaction in RETRY.
regards
dwp
|
| 99.5 | PCI 2.1 protocol did change... | POBOXA::DUNCAN | | Thu Apr 10 1997 20:16 | 13 |
|
There are new ordering rules and a "delayed transaction" type that was
not specified in 2.0. The delayed transaction makes PCI a pended bus
and makes PCI to PCI bus bridges and peer-to-peer transactions
possible without deadlock.
For most applications the delayed transaction is transparent because
it looks like a retry.
There are changes to class-codes used by BIOS to associate devices to
drivers.
/SHD
|