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.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: Road Walker who wrote (61976)	11/3/2001 8:54:12 AM
From: combjelly	Read Replies (1) \| Respond to of 275872

"That it is unilateral, contrived, and intentionally deceptive." In other words, you don't know. Thanks for being honest about it. Yes it is unilateral, although there is nothing stopping Intel from submitting machines to get rated. I doubt if they will, although it would be interesting if AMD did it for them. As far as contrived, all benchmarks are to some extent. The only question is exactly how close they match the way that end users actually use their computers when they are interested in performance. So far, Intel hasn't really provided that. As far as being intentionally deceptive, don't you think that running at a higher clockrate and delivering 60% of the IPC deceptive? Now true, pipelines are always being deepened to get greater headroom, but usually the increase in clockrate more than exceeds the drop in IPC, giving a net benefit. And the P4 does have enough clock rate advantage over the PIII so that the net increase is worth it. But that is not true for Athlons, it looks as if the clock rate advantage for the P4 over the Athlon will be a whole lot smaller, at least on 0.18 micron. Whether or not the Palomino can reach 2GHz is beside the point, it can reach 1.5GHz. And that is enough to erase the clock rate advantage of the P4 at it's apparent peak of 2GHz. So how to present it? Like many things in computing, this has happened before. The traditional solution has been to use benchmarks or metrics of some type. Like MIPS, FLOPS, Linpack and/or SPEC. In the past, the machines in question had different instruction sets, and that made things more difficult because you could always play Stupid Compiler tricks and do other things that clouded the issue. Now, the two architectures run the same code and can use the same hardware. If a decent suite of benchmarks can be chosen, then you can use the same binaries on more or less the same hardware (ok, the motherboards may differ, but VIA chipsets can level even that difference) and actually benchmark the CPU itself. Now you may argue that AMD chose software that favors their processor and the benchmarks need to be more balanced. But to argue that the concept itself is "intentionally deceptive" is either ignorant, stupid or deceptive itself. To hold that the speed of the processor is the only acceptable metric is attempting to re-write history, the industry has rejected that in the past, and there is no reason to change it.