home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.arch
- Path: sparky!uunet!wupost!sdd.hp.com!nigel.msen.com!yale.edu!jvnc.net!nuscc!ntuix!eoahmad
- From: eoahmad@ntuix.ntu.ac.sg (Othman Ahmad)
- Subject: Re: BUSES
- Message-ID: <1992Jul30.130720.7567@ntuix.ntu.ac.sg>
- Organization: Nanyang Technological University - Singapore
- References: <1992Jul28.180537.14749@StarConn.com>
- Date: Thu, 30 Jul 1992 13:07:20 GMT
- Lines: 42
-
- In article <1992Jul28.180537.14749@StarConn.com> jpp@StarConn.com (John Pettitt) writes:
-
- :
- : On a more normal level an ISA bus will peak out at around 3.3 MB a
- : second half duplex (600 ns per transaction 2 bytes per). In the real
- This is the standard memory cycle at 8Mhz with a certain wait state. Surely
- 600nS transaction time is too conservative. Today, static RAM can operate at
- 20nS access time.
- For a sequential access mode, there is no address setup time. The
- transaction time is determined by the access time and propagation delay
- from card to microprocessor. This technique is called the synchronous
- mode for the EISA. At 33Mbyte /second, that makes the transaction time to be
- 121 nS for a 32-bit wide transfer. For a 16-bit wide ISA, the skewing effect
- should be smaller so let us assume that the transaction time be 110nS. That
- makes the peak transfer rate to be 18Mbytes/second.
- The 18Mbytes/s figure assumes that there is no-wait state inserted by
- the PC chipset, using the 0WS, and no address setup time is necessary because
- the transfer is sequential and the address is generated by the ISA card itself.
- The card vendor must design the transfer protocol himself, unlike
- EISA, which is standardised. The advantage is huge because of the large
- number of ISA machines.
- The card which needs these high speed would be the communication cards,
- video cards and harddisk controller cards.
-
- My question is: Has anybody tried to experiment with the 0WS signal of the ISA?
- If they have, may I know the problems which are faced?
-
- I'm not interested in theoretical discussion because I had asked the same
- thing but am not satisfied with the answers or suggestions.
-
- Since John Pettitt mentioned that he is a retired I/O system designer, maybe
- someone else may volunteer. If John Pettitt had done some experiments with
- the 0WS, it would be nice to hear his views.
-
- Thank you for any comment.
-
- --
- Othman bin Ahmad, School of EEE,
- Nanyang Technological University, Singapore 2263.
- Internet Email: eoahmad@ntuix.ntu.ac.sg
- Bitnet Email: eoahmad@ntuvax.bitnet
-
-