Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Alighieri who wrote (114496)	6/5/2000 11:02:00 PM
From: Daniel Schuh	Read Replies (1) \| Respond to of 1574772

Al, who knows the veracity of that bit, but it wouldn't surprise me if Willy looked a bit weak on existing stuff, especially gaming code with lots of FP. Remember, the entire x87 FP instruction set is deprecated. Intel will have some careful orchestration of benchmarks to do in the launch run-up there. Should be entertaining. Cheers, Dan.

To: Alighieri who wrote (114496)	6/6/2000 12:58:00 AM
From: Charles R	Respond to of 1574772

Al, <Interesting. Fellows here claim to have some performance numbers of a Willamette (Pentium IV). Translation, which I am posting below, was taken from here: "During this first day in Computex, I was able to throw(cast) a blow of test on a machine based around the Intel's future processor Willamette (future " Pentium IV " ) put rythm to 800 Mhz (where are the 1.5 Ghz?). The motherboard was for its part based on a chipset i850, and used naturally some memory Rambus. I was able to make a fast benchmark on this machine results? With a GEFORCE 2 GTS, 230 in CPU 3D mark and 4157 in the 3D mark, under 3D Mark 2000 while with a PIII 600 a similar configuration obtained 388 in CPU 3D mark and 5136 in the 3D mark!". These numbers seem a bit on the low side. Mind you, microarchitecturally, Willamette's fpu isn't anything to scream about, but I suspect you'll see higher scores in future (eg, less beta-y) revisions and with better drivers.> Scumbria has been beating the architectural performance issue to death on this thread for at lest a year now. Complexity and newer generations typically do not produce higher IPC. Typically a balanced deeply pipelined machine will have lower IPC than a balanced less-deeply pipelined machine. Wilamette numbers are unlikley to be that good except in the cases where the SIMD-128 stuff is used. And for the power aspect, there have been plenty of rumors about heat dissipation problems on Wilamette. Rumors put Merced and Wilamette samples at good 30-50% below the MHz numbers touted by Intel. Not a great sign for products that are supposed to be launching in H2. Chuck