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: wanna_bmw who wrote (52664)	8/27/2001 6:34:31 PM
From: fyodor_	Read Replies (2) \| Respond to of 275872

Wanna: This makes the Athlon a great graphics workstation, if AMD can market it well as that, but consumer level 3D realtime rendering performs much better on the Pentium 4. Well, that depends very much on the specifics. There are plenty of NEW games out there where the Athlon (1.4) beats the P4 (1.8). Additionally, virtually all the old games "consumers" still play would run better on the Athlon (but you could easily argue that performance on these games is already "high enough"). Additionally, the Athlon kicks P4 butt in AI benchmarks (searching, shortest path, chess etc.). I honestly don't think I'm being unfair to either side when I say that a 1.4GHz Athlon is on par with a 1.8GHz P4 in terms of performance. With the 2GHz P4, however, Intel has moved into the (performance) lead. While I haven't seen any direct benchmarks, extrapolation and other guesswork leads me to conclude that it would require a 1.5GHz Palomino to reach performance parity with the 2GHz P4. Btw, do you know if Intel's "consumer" P4 on .13&#181 will have 512kB L2 cache as well? I suspect the P4 would benefit more from the added cache than the A4, thus allowing the P4 to partially close the IPC gap. -fyo

To: wanna_bmw who wrote (52664)	8/27/2001 7:10:00 PM
From: jcholewa	Read Replies (1) \| Respond to of 275872

> consumer level 3D realtime rendering performs > much better on the Pentium 4. Are you sure about that? What do you classify as "consumer level 3D realtime rendering"? If you are talking about 3D games, then your comment is a blanket statement that is not always true and it therefore, well, erroneous. There are 3D games in which the Pentium 4 is stronger, and there are 3D games in which the Athlon is stronger. Two modern, popular games in which the Athlon outperforms the P4 would be Serious Sam and Max Payne. It also does better in (I think) that Aquanox thing, whereas P4 tends to win in Quake III, MTBR, and DroneZ. Er, if you weren't talking about 3D games, then never mind. ;) -JC

To: wanna_bmw who wrote (52664)	8/27/2001 7:13:18 PM
From: Dan3	Respond to of 275872

Re: When an optimized app is used You are confusing "an optimized app" with an optimized benchmark. There are almost no examples of actual applications, such as serving data, encoding video, compiling code, etc. where a P4 1.7 is faster than an Athlon 1.2. Meanwhile virtually the entire installed base of millions of programs (un-"optimized" code) is faster on an Athlon. What do exist are some artificial "benchmarks" that were carefully coded to avoid the many instructions that P4 can't handle well. These aren't optimized X86 applications, they are optimized P4 benchmarks. We need to see, for example, how Microsoft Media encoder (tuned to avoid P4's weaknesses) compares with some of the more standard media encoding software (that was compiled with the goal of encoding media, rather than working around P4's weaknesses). Maybe Microsoft Media encoder, using "optimized" code really is an optimized application. The interesting thing to do would be to compare its performance with other software with the same functionality that was neither aided nor crippled by the need to accommodate P4's various warts and failings. The hard thing, of course, is to to ensure that equal quality output comes from each program (when encoding anything, it's not hard to pump out ugly garbage at high speed, or high quality sound and images at low speed, what's tough is to do both at once). So far, at least, whenever the task is defined as a real application, doing real work - serving data, compiling code, etc. Athlon is faster. When the task is defined as completing an Intel optimized benchmark, sometimes the P4 comes out ahead.

To: wanna_bmw who wrote (52664)	8/27/2001 7:20:18 PM
From: pgerassi	Read Replies (1) \| Respond to of 275872

Wanna_bmw: Why is WME weighted 60% in the benchmark? Would you like it if the benchmarks used for comparisons use 3DNow and does not use SSE or SSE2, "just to be fair"? Why is P4 so slow on that common consumer task of rotating a scanned image 1 degree to make it straight? Would you like the majority of benchmarks to do lots of shifts, rotates, branch randomly and do lots of x87 and 3DNow code? That is the same as what you are saying for the P4 done for the Athlon. I'm sure that I could come up with a few dozen programs that absolutely hate the architecture of P4 and show it so poor at doing tasks that a 600MHz Duron beats a 2GHz P4. That would be as fair as the benchmark suites promulgated by you. I just do not think that is fair but, you do. Pete