home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!wupost!csus.edu!ucdavis!endive.eecs.ucdavis.edu!wilken
- From: wilken@endive.eecs.ucdavis.edu (Kent Wilken)
- Newsgroups: comp.arch
- Subject: Re: DEC Alpha AXP System Performance
- Message-ID: <19225@ucdavis.ucdavis.edu>
- Date: 14 Nov 92 20:33:30 GMT
- References: <BxIM38.L9F.2@cs.cmu.edu> <15445@auspex-gw.auspex.com> <BxL3IH.KtH.2@cs.cmu.edu>
- Sender: usenet@ucdavis.ucdavis.edu
- Organization: U.C. Davis - Department of Electrical Engineering and Computer Science
- Lines: 42
-
- In article <BxL3IH.KtH.2@cs.cmu.edu> lindsay+@cs.cmu.edu (Donald Lindsay) writes:
- >In article <15445@auspex-gw.auspex.com> guy@Auspex.COM (Guy Harris) writes:
- >>Hmm. As I remember, somebody from DEC at a session on topics such as
- >>OSF/1-on-MIPS on Tuesday night at the San Antonio USENIX claimed that
- >>HP-PA was "out of gas"; ...
- > ...
- >
- >What he may have meant, is that HP won't be able to use offchip
- >primary cache much longer. That was difficult at 66 MHz, and is
- >really impressive at 100 MHz. Even HP can't do it at 200 MHz, surely. <---
- > ...
- >
- >Don D.C.Lindsay Carnegie Mellon Computer Science
-
- A person from HP previously addressed your incredulity regarding
- the feasibility of 200MHz off-chip caches:
-
- #From: linley@hpcuhe.cup.hp.com (Linley Gwennap)
- #Newsgroups: comp.arch
- #Subject: Re: Re: Offchip Primary Cache
- #Message-ID: <32580093@hpcuhe.cup.hp.com#
- #Date: 5 Mar 92 21:46:37 GMT
- #References: <1992Mar03.233219.176162@cs.cmu.edu#
- #Organization: Hewlett Packard, Cupertino
- #
- #(Donald Lindsay) discusses HP cache access rates:
- #> I'm impressed that HP can do it at 100 MHz. Can you do it at 200 MHz?
- #
- #There is no reason to believe that 200 MHz will not be possible. 100 MHz
- #primary caches on the PA7100 are accomplished with standard PCB technology,
- #ceramic PGA packaging, and off-the-shelf 9ns TTL I/O SRAMs. SRAMs get
- #faster every year, multichip modules are getting cheaper, ECL I/O circuits
- #are well understood, future VLSI technologies will improve timing
- #overhead, etc, etc. Two years ago these same discussions were taking place
- #about whether external primary caches could ever run at 50 MHz.
- #
- #Some of the work going on to define new high-speed I/O protocol standards
- #for DRAM and SRAM memory devices also look interesting for application
- #in future cache memory designs.
- #
-
-
-