Author: Tom Kerrigan
Date: 18:25:09 06/30/03
Go up one level in this thread
On June 30, 2003 at 21:16:35, Sune Fischer wrote: >On June 30, 2003 at 21:13:21, Tom Kerrigan wrote: > >>On June 30, 2003 at 21:11:06, Sune Fischer wrote: >> >>>On June 30, 2003 at 20:51:58, Tom Kerrigan wrote: >>> >>>>On June 29, 2003 at 16:57:02, Sune Fischer wrote: >>>> >>>>>On June 29, 2003 at 16:08:02, Jeremiah Penery wrote: >>>>> >>>>> >>>>>>>Measuring specint for a new chip using a compiler that in effect cripples it, is >>>>>>>not only incompetent, it's misleading, IMO. >>>>>> >>>>>>If it's incompetent to use a compiler that makes faster binaries than another >>>>>>compiler, then I think you can call everyone incompetent. >>>>> >>>>>Well yes, it is. >>>>> >>>>>If Carl Lewis has a broken leg, we should postpone the race until he is back in >>>>>shape, anything else will just be a race with a broken leg. >>>>> >>>>>It is however amazing how fast the chip is running, even with it being on >>>>>crutches. >>>>> >>>>>-S. >>>> >>>>It is not on crutches. The x86-64 ISA is more or less the same as the x86-32 >>>>ISA. >>>> >>>>-Tom >>> >>>Seeing is believing. >>> >>>If I'm not mistaken the specint uses Crafty as an integer benchmark, obviously >>>this is 64-bit *integer*, hence huge difference between 332 abd 64 bit, hence >>>the whole thing is more or less rediculous. >>> >>>-S. >> >>What would you suggest? That AMD retract its SPEC submissions and ask vendors >>not to make SPEC submissions until performance of 64-bit binaries on the Opteron >>is to your liking? Sure. That'll boost sales. > >Ah I see, you think running 32 bit binaries is testing a 64 bit chip, hurray for >progress - why did AMD even bother? > >-S. Chicken and egg. 64 bit is obviously good (that's why they bothered) but the chip has to be available for a while for people to take advantage of 64 bit. Until then, its 32 bit performance is what's interesting to users (and what's accurately measurable). -Tom
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.