SI
SI
discoversearch

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: Elmer who wrote (115415)6/11/2000 3:00:00 AM
From: pgerassi  Respond to of 1577883
 
Elmer:

Go to JC's web page: jc-news.com

Go to QMC and Moldyn benchmarks page.

Your vaunted Cmine at 1Ghz on i820 PC800 is beaten by a 770Mhz Athlon for Moldyn. It is 34% slower than an 1Ghz K75 Athlon. Look at all the Athlons faster than the top Cmine.
In QMC, a 1Ghz PIII was 39% slower than a 1Ghz K75 Athlon.

These are real world benches, not a made up SPEC bench that allows a non-real world compiler (software is coded for the whole market not just one CPU type) to be used. When a third party compiler is used, Cmine does not measure up to Athlon.

With all the games that are played with benches, one should try to minimiize the differences in environment. Same NB, same software, same compiler, same memory, and so on. Notice on QMC, RBMS slows down as memory size goes up. When Willie comes out, it will have to compete against Mustang, on a DDR 266 chipset. It may even have to go against Sledgehammer. If it does, it will be flattened.

Classic K7 is 7 gen, Tbird is 7.1 gen, Mustang is 7.25, and Sledgehammer will be 7.5.

Pete



To: Elmer who wrote (115415)6/11/2000 11:11:00 AM
From: Scumbria  Read Replies (2) | Respond to of 1577883
 
Elmer,

Your constrained use of the word "yield" continues to defy my understanding. Designers use the word to yield to represent the speed distribution and functionality of the manufacturing.

i.e. "Athlon is yielding 30% at 1GHz, 40% at 900MHz, 20% at 800MHz, 10% non-functional".

This provides far more information than "PIII is yielding 90%", because most of the PIII parts are too slow to sell (except as jewelry.)

Scumbria



To: Elmer who wrote (115415)6/11/2000 11:51:00 AM
From: TimF  Read Replies (1) | Respond to of 1577883
 
Easy to say when you don't have any links to reference. The links that do exist show exactly the opposite. CuMine destroys
Athlon in both SPECint and SPECfp. I'll do what you are unwilling or too embarrassed to do. I'll post links that backup my
claims:

specbench.org.
specbench.org.
spec.org.
spec.org.


Note the links show the Athlon's configuration to include
"Secondary Cache:
512KB(I+D), 333MHz"

This means they are K75s not T-birds. Since AMD is not making any more K75s this is probably not an important comparison.

Tim