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: Tenchusatsu who wrote (42689)	12/3/1998 1:29:00 PM
From: Jim McMannis	Respond to of 1577985

RE:"Wasn't Winstone 97 more favorable toward AMD and Cyrix chips than Winstone 98? Seems like every new benchmark from ZD Benchmark Operations (ZDBOp) favors Intel processors a little more. Three possible explanations come to mind:" Yes, that is true... RE:"1) ZDBOp is conspiring with Intel to make AMD chips look bad. 2) Coincidence. 3) Intel chips are designed with the future in mind, while AMD chips are stuck in the past." ----- Maybe Not likely Who says what the future brings. Last year you could have loaded all those programs, and the year before that and before that... So now all of a sudden, loading all those programs is the future? Very fishy. RE:" Second, if Intel is cheating on Winstone 99, I trust that Herr Uberclockermeister will catch them in the act. He's done it before when he found a discrepancy in the Winbench Disk scores running under Intel's Ultra DMA drivers. He did it again when he showed that 3D Winbench 98 gives abnormally high scores to Intel's i740 graphics chip."... Yes it's happened before but not to such a widely used benchmark as Winstone...not to mention all the magazines that use Winstone. Some owned by ZD.... Let's draw a parallel. Intel monetarily supports boxmakers who advertise on Television if the play and show the Intel logo...you know, da dum da dum... We know how powerful the Media is particularly TV... And so the written word...Internet, PC magazines etc... Having a benchmark that favors one manufacturers chips is an extremely powerful tool. So maybe a chip maker "helps" a publishing company out with a benchmark? Humm...or maybe the makers of the benchmark have a stake in the chip maker and want to make the chip maker look good? Or maybe it's like a campaign contribution being paid back? ---- Has it happened before? You bet something fishy has...you mentioned 3D Winbench....a benchmark can be made to make any chip look good. ----

To: Tenchusatsu who wrote (42689)	12/3/1998 1:35:00 PM
From: Kevin K. Spurway	Read Replies (1) \| Respond to of 1577985

Re: "Second, if Intel is cheating on Winstone 99, I trust that Herr Uberclockermeister will catch them in the act. He's done it before...." Funny that you agree that Intel has been caught "cheating" on benchmarks before, given your position. I'm not too worried about Winstone 99. If AMD's shortcoming is the cache speed (and you make a pretty good argument that it isn't), then products arriving shortly from AMD will solve things. If the problem derives from Winstone 98 and now Winstone 99 devoting progressively more attention to 32 bit vs 16 bit code (the latter where the K6 core really excels), then fair enough. I have no doubt that K7 will correct this "shortcoming" in the K6. I also have no doubt that MHz and price are what sells--not benchmarks (even though they're good for bragging rights). Kevin

To: Tenchusatsu who wrote (42689)	12/3/1998 1:45:00 PM
From: Ali Chen	Read Replies (1) \| Respond to of 1577985

Tench, <Intel chips are designed with the future in mind, while AMD chips are stuck in the past.> This is your Intel propaganda. The reality is: 1. New Intel chips are designed for a better proprietary bus that was not made open to PC industry, with the goal to cement Intel's monopoly in PC business in the "future". 2. AMD must rely on the business/market reality. The NexGen already had their own advanced "dual bus", but we all know their business results. Therefore AMD had to design their chips for the available "Socket7" architecture, which incidentally was designed by Intel in a crippled way - with look-aside L2 cache that blocks memory accesses and is cheap but inefficient. Reason for the cheap is the same - to flood the PC market with x86 and dominate it. 3. The AMD K6 chips vastly outperform the corresponding Intel designs - Pentium MMX. These P-MMX chips from the "past" still accounted for about 50% of Intel revenues in 1998. Now, back to ZD benchmark. Your speculations about caches and context switching are irrelevant. The 16-bit argument is the history as well - both Winstone98 and 99 are fully 32-bit applications. Your Uberclockermeister also caught everything wrong. The ZD does not "cheat" nor favour any CPU. With each test generation they just follow the main software trend - more bloated applications, with many apps loaded simultaneously into memory. As I said before, S7 CPUs are handicapped in the bus area. With the new ZD method of "jumping" between several apps, the caches are trashed more frequently, which creates more bus traffic (where current K6 has the disadvantage). That's it. No conspiracy nor coincidence. - Ali