Article View: demon.tech.pc
Article #69004Re: Caught by cache?
From: Roger Hamlett
Date: Wed, 17 Apr 1996 00:00
Date: Wed, 17 Apr 1996 00:00
50 lines
3342 bytes
3342 bytes
In article: <oKMISAAuP+cxEwZ2@mwb.u-net.com> Marcus Bainbridge <marcus@mwb.u-net.com> writes: > John Naismith <John@naismith.demon.co.uk> wrote: > > >I've just checked & these chips have 128 pins on each. Seems a little excessive > >for 64 bit cache ? > > > >64 bit data + 32 bit address + ? > > > >Anyone know what the hell they are? > > Well, if you have a part number... :-> > > With QFP and similar devices, they come in standard pin arrays, so there > could be a lot of power pins (ISTR something about quad cache having a > particular problem with routing power around the silicon..?) > > But if they are cache controllers, then then have to access 64-bit data > and 20-bit cache address (say), plus chip selects, interfaces with main > memory (could be on a different bus and multiplexed by the cache > controller, so call it another 64-bit data and 32-bit DRAM address). > All adds up... > > If I could find the manual for my motherboard (never trust a woman to > tidy up - you'll never find anything ever again...), I could find out a > bit more about the Triton chipset. Maybe it's on the Intel CDs at > work..? I have two motherboards here, both using the same chips. One uses just two chips, Intel Partno:S82438fFX. This board has a COAST socket, and these _are_ the cache interface chips from the PCIset. The other has four, all the same size package, and in this case, the extra two, are the memory. It appears that machines implementing pipeline burst mode cache, have these larger cache controller chips. The Triton chipset, uses them on such machines. Adding to the above comment, about the number of pins needed, you also have to add the memory control bits, from both sides, and the tag data. It uses a lot of pins. Machines without P/B cache, seem to use a single larger chip for the cache control, but the layout described, is common, in a lot of motherboards. The puzzling/worrying part, is the addition of the strange labelling as 'write-back', and the strange utilisation of the pins. Given the recent spate of stories about fakes, the only real solution, is to use a diagnostic, that tells you what you have actually got, and stop worrying about how the particular manufacturer has implemented it. Best Wishes -- --------------------------------------------------------------------- | EMail roger@ttelmah.demon.co.uk | | A beard! A beard! cried Fly Nicholas. 'By God, that's a good one!'| | (Chaucer) |
Technical Details
Message-ID:
<279049516wnr@ttelmah.demon.co.uk>Path:
arc1-ca.pugleaf.net!unknown!not-for-mailReferences:
<415183800wnr@draycot.demon.co.uk> <708079308wnr@xabcs.demon.co.uk> <549011010wnr@naismith.demon.co.uk> <61gucDAfVUcxEw5c@mwb.u-net.com> <619927758wnr@naismith.demon.co.uk> <oKMISAAuP+cxEwZ2@mwb.u-net.com>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 1036bc,7858595a79b6a65d X-Google-Attributes: gid1036bc,public From: Roger Hamlett <roger@ttelmah.demon.co.uk> Subject: Re: Caught by cache? Date: 1996/04/17 Message-ID: <279049516wnr@ttelmah.demon.co.uk> X-Deja-AN: 148065558 x-nntp-posting-host: ttelmah.demon.co.uk references: <415183800wnr@draycot.demon.co.uk> <708079308wnr@xabcs.demon.co.uk> <549011010wnr@naismith.demon.co.uk> <61gucDAfVUcxEw5c@mwb.u-net.com> <619927758wnr@naismith.demon.co.uk> <oKMISAAuP+cxEwZ2@mwb.u-net.com> x-broken-date: Wednesday, Apr 17, 1996 21.29.49 x-mail2news-path: disperse.demon.co.uk!post.demon.co.uk!ttelmah.demon.co.uk organization: None reply-to: roger@ttelmah.demon.co.uk newsgroups: demon.tech.pc