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: Paul Engel who wrote (42309)	11/27/1998 9:58:00 PM
From: Jim McMannis	Read Replies (2) \| Respond to of 1572631

This new Winstone 99 is very interesting. Seems like every year with a new Winstone it seems to favor Intel chips. Another report using Winstone 98 showed the Pentium II clocked at 500 and the AMD k2-500 roughly the same. 32 vs 31.2. Now, here comes a new Winstone and all a sudden the AMD-K6-2 at 450 is more than a few winstone % slower than a Celeron at 450. Same thing happened when they went from Winstone 97 to 98. Maybe they include more 32 bit apps every year? I wonder where Anand got his engineering same? Why not a production sample? Seems to me that it AMD should hand pick a chip for the board wizards to clock...yet the one Anand had would only clock to 450....not to mention Anand seems to be in love with the overclocked CeleronA-300. In perspective though, "Megahertz sells" (TM-McMannis) so the 400 will sell well...too well right now as it seems they can't make enough. Jim

To: Paul Engel who wrote (42309)	11/27/1998 10:33:00 PM
From: Cirruslvr	Read Replies (2) \| Respond to of 1572631

Paul- RE: "Scumbria & Intel Investors - Anand's Benchmarks of K6-2 400 MHz Anand has posted his report on WInstone 99 benchmarks fo the 400 MHz K6-2 and various other Intel & AMD chips at different speeds. The 400 MHz K6-2 is only about 4% faster than the 350 MHz K6-2 WHEN THE NEW BIOS with Write Allocate and Write Merge Buffer are implemented. Surprisingly , WITHOUT the new BIOS and Write Allocate and Write Merge Buffers implemented, the 400 MHz K6-2 is ACTUALLY SLOWER THAN THE 350 MHz K6-2 ! For comparison, the 400 MHz K6-2 is only 8% faster than a 300 MHz Celeron A! The 400 MHz K6-2 is 10% slower than a 400 MHz Pentium II and 15% slower than a 400 MHz Pentium II XEON." First of all, I will be the first to say I am disappointed with the chip's performance on THIS benchmark, and also that the non-bios-enhanced-CXT-core K6-2 performed lower than a non-CXT-core K6-2. What can be said of this? One thing to do is blame the processor for being an engineering sample. The more important thing to do is realize that the CXT-core 400MHz K6-2 will only be running on motherboards that support it and so that benchmark (the CXT-core K6-2 400MHz running w/o the bios support) becomes futile. Secondly, you should re-read this one part from Anand's comparison because you seem to have selective memory: "One interesting thing to note is that under Winstone 99, Intel processors score considerably higher than their Super7 counterparts. This is primarily because Winstone 99 centers itself around multitasking performance, where the Pentium II, Celeron, and Xeon processors excel. Since their L2 cache runs at a much higher speed than that of the K6-2, they allow the multiple applications, which fit almost entirely in the L2 cache, to perform much better than those that run on Super7 platforms where the L2 cache is limited to run at 100MHz (without overclocking that is)." Clearly, from what Anand wrote (which you just happened to overlook) the benchmark favors processors which have a fast L2 cache- something that puts the K6-2 at a hefty disadvantage that produces a lower benchmark. IF Anand had run Winstone 98, the numbers, would have been completely different and less in favor of chips with fast L2 caches. IF Anand had run Winstone 98, the K6-2 would have performed better and closer to its competitor at the same clock speed. IF another hardware site runs Winstone 98 on the CXT-core K6-2, I will have something to talk about. Obviously these IFs are just that- IFs. We must stick to what we have. Well, sticking to what we have will actually favor an AMD chip in the future- the K6-3. Since the benchmark favors chips with a fast L2 cache, I can easily assume the K6-3 will perform faster than the K6-2 and closer to, or maybe even better than, its competitors same clock speed processor. What can be concluded from Anand's benchmarks- -When running Winstone 99, chips that don't have a fast L2 cache will score noticeably slower than processors that have a fast L2 cache. -The K6-3 will perform very well on Winstone 99 compared to the K6-2 -You have something to brag about while the K6-2 is out -I will have something to brag about when the K6-3 comes out BTW, 10% Slower but 24% CHEAPER makes ($259 K6-2 400 vs. $323 PII 400 on Pricewatch) the 400 MHz K6-2 the Price Performance LEADER !

To: Paul Engel who wrote (42309)	11/28/1998 1:38:00 PM
From: Scumbria	Read Replies (1) \| Respond to of 1572631

Paul, Surprisingly , WITHOUT the new BIOS and Write Allocate and Write Merge Buffers implemented, the 400 MHz K6-2 is ACTUALLY SLOWER THAN THE 350 MHz K6-2 ! Generally speaking, the kinds of design changes required to achieve higher clock rates are contrary to architectural performance. The K6-3 designers were probably depending on the onboard L2 to compensate for the other tradeoffs. If the K6-3 was ready to ship with the cache enabled, it probably would be shipping. Scumbria