Message-Id: <3.0.32.19990520004516.00aa4930@pop.xs4all.nl> X-Sender: diep AT pop DOT xs4all DOT nl X-Mailer: Windows Eudora Pro Version 3.0 (32) Date: Thu, 20 May 1999 00:45:18 +0100 To: pgcc AT delorie DOT com, pgcc AT delorie DOT com From: Vincent Diepeveen Subject: Re: K7 potentials Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Reply-To: pgcc AT delorie DOT com At 05:48 PM 5/19/99 -0400, Adam Schrotenboer wrote: >I was wondering if there is any speculation, q's, etc about what the K7 >may be capable of. well i don't expect too much. K6-3 should be way faster too than k6-2, yet a k6-3 at 475Mhz was for my 32 bits programs and some other 32 bits integerprograms slower than a PII-450, although really close to it. Rounding off it's a PII-450 speed. Xeon is under linux about 7% faster than a PII of equal Mhz for me. Of course when using the parallel version, the shared memory rocks on the Xeon. So will it at a K7. So whether K7 core is way quicker... ...first see then believe. >For example: > Will the K7 be less "register starved" than the rest of the x86 >archictecture, or was this simply not possible while maintaining x86 >compatibility???? > If it will have more registers, how much difference will this make >performance or optimization-wise?? > Will this be a EGCS thing, or will it be more likely PGCC?? > Anybody know when K7's will be available at a decent price (yes, >this is probably near to impossible to determine, but it could make an >interesting point for discussion.) > Anything more that is AMD &| K7 related. > >I'm assuming that a different scheduler will be used, either that or the >Haifa sched w/ full processor specs. > >Anyway, I'm bringing this up due to the (hopefully) soon release of the >AMD K7 in June or July (I hope so, really I do) > >