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: Tony Viola who wrote (47681)	1/28/1999 3:01:00 PM
From: Tenchusatsu	Read Replies (1) \| Respond to of 1572319

Tony, slight correction. Celeron 400 outperforms K6-3 (w/ 1 MB L3 cache) on every High End benchmark that Anand ran. Business Winstone 99 is still K6-3's domain. Tenchusatsu

To: Tony Viola who wrote (47681)	1/28/1999 3:02:00 PM
From: Scumbria	Read Replies (1) \| Respond to of 1572319

Tony, K7, some folks on this thread are beginning to call it overdesigned. Xeon = Celeron + a large coherent L2 cache. Intel figured out a long time ago that you don't need a killer core to compete in the workstation/server market. All you need is: 1. Adequate MHz 2. A large L2 cache 3. MP compliant caches and bus K7 is an attempt to do 1-3 + the killer core. If it works out, Dirk Meyer will be a hero. If it doesn't work out, everyone will say "they should have done the easy part, and not worried about the killer core." The K7 seems to have a lot of risk, derived from it's complexity and die size. It will be interesting to see what happens the next few months. Scumbria

To: Tony Viola who wrote (47681)	1/28/1999 3:03:00 PM
From: Elmer	Respond to of 1572319

Re: "Joey, this is pretty awesome, isn't it, that Celeron 400 (?Celery by each and every Intel hater there is) outperforms K6-3 on every benchmark Anand ran. " And the Celeron is AVAILABLE. EP

To: Tony Viola who wrote (47681)	1/28/1999 6:57:00 PM
From: Petz	Read Replies (3) \| Respond to of 1572319

The Celeron 400 emphatically did NOT outperform K6-3 on every benchmark Anand ran. Read the whole review, the Celeron 400 only did well on a few handselected "highend" applications and non-3DNow-enhanced games. This is just Anand's attempt to counterbalance the flack he got from his last article, when he was mockingly called "AMD's teenage love." To counteract this appearance of bias, he set out to find something that runs faster on a Celeron. He found it, in the multiprocessor versions of three esoteric programs -- Microstation SE, Photoshop 4.0 and Visual C++ 5.0. He didn't worry that both Visual C++ 5.0 and Photoshop 4.0 are OBSOLETE VERSIONS, or that no one in their right mind would shell out an extra $1000 to run multiprocessor versions on a single CPU. On business Winstone 99, the K6-3-450 is 20% faster than the Celery 400 (14% faster on Windows NT). Also worth noting is that the Celery 400 is within 1% of the Pentium II-400 and 5% of the Pentium II-450 on every Windows NT benchmark that was run, illustrating that the PII-400 and 450 are useless products from a business user's point of view.. All the graphics benchmarks (2/3 in the highend test) and gaming benchmarks are also useless because he admits to using beta drivers for his Riva TNT graphics card which have a poor Socket 7 implementation. How else can you explain the fact that without 3DNow acceleration on the "crusher" demo, Sharkyextreme.com got 35.5 frames/sec with a K6-3-400, while Anand only got only 21. His conclusions are also baseless. He blames lower frame rates on some games on the K6-3 FPU, but doesn't notice that the K6-3 450 is only 7% faster than the K6-3 350. Obviously, its not the FPU holding back performance! In fact, when the video card performance is removed from the picture by running in "software mode", the K6-3-450 is 37% faster than the Celeron 400. Since a Celeron 400 is only 12% faster than a Celeron 300A in this benchmark, it indicates that a Celeron 700 might be as fast as a K6-3-450! <ggg> Obviously, there's something wrong with Anand's benchmarks. The deck was also "stacked against AMD" by only using OpenGL games, when there are five times as many DirectX games on the market. DirectX games are automatically accelerated by 3DNow. Sometimes you have to dig to find the truth Petz