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: Windsock who wrote (81640)	12/1/1999 1:05:00 PM
From: Jim McMannis	Read Replies (1) \| Respond to of 1572545

RE:"It will be interesting to see some independent comparisons of the AMD 750 and the Intel 733'... ------- Last I checked, 750>733. OTOH, Intel has taken benchmarks to a new level of "yeah right"..."sure, anything you say". <G>

To: Windsock who wrote (81640)	12/1/1999 2:18:00 PM
From: Ali Chen	Read Replies (2) \| Respond to of 1572545

Winsuck, <SPECfp95 -> AMD-750 24.3 Intel-733 28.1 Whoa, what happened here !! If you look at the AMD benchmark configurations you discover that an old driver was used, DX6.1A instead of the current DX7> If you would have any idea about what you are talking about, you would realize that Microsoft DirectX drivers deal primarily with video (and sound) application interfaces. If you would be remotely familiar with the SPEC95 benchmark suite, you would notice that THE WHOLE SET OF SPEC BENCHMARKS is simple CONSOLE applications, or just a black "DOS-box" if you don't know what the "console app" is. SPEC tests perform PURE CALCULATIONS and do not use any graphics calls to DX whatsoever. The SPEC binaries are PRECOMPILED and no DX can "enable" or "disable" any "fp extensions" at run time. If you want to know as why Intel posted 28.1 in SPECfp95, then that is because they employed NEW Intel Fortran/C compiler which uses Intel SSE instructions. This compiler is a) not available for general public until Feb-2000, and b) presumably has problems since their binaries sometimes give incorrect results in SPECs. Therefore it is still a waporware from practical standpoint. Secondly, no one uses Intel compilers when designing commercial software - most people use Microsoft VC++, for system compatibility reasons (remember your DXxx?). Therefore, due to above reasons, the results presented by AMD reflects practical performance differences (or lack of them) given current state of commercial software. Therefore calm down and relax, or you can burn down your last brain cell.

To: Windsock who wrote (81640)	12/1/1999 4:03:00 PM
From: Scot	Respond to of 1572545

If you look at the AMD benchmark configurations you discover that an old driver was used, DX6.1A instead of the current DX7 driver. The DX6 driver enables the AMD fp extensions but NOT the Intel fp extensions. The current DX7 driver enables BOTH the AMD and Intel extensions. This is a very nice way to rig the benchmark results. Generally, the benchmark comparison game is played with more subtle bias than this. Umm....ok. I don't represent that I'm even well-versed with benchmarking....but that sounds liks FUD to me. If I recall, DX allows certain programs (e.g., games) to directly access hardware (e.g., sound cards and video cards) in Win9X. What does that have to do with a synthetic processor benchmark? DirectX 7 extensions? Are you talking about some SSE or 3dnow extensions? If so, I don't see how that will affect fp and integer performance on a synthetic test. Graphic performance....yes..but we're not counting frames....are we? BTW, there was no mention of Direct X on the spec site. You'd think there would be some discussion if it could so radically change the outcome of the test. I did find the following: spec.org Q3: What does the "C" in CINT95 and CFP95 stand for? A3: In its product line, SPEC uses "C" to denote a "component-level" benchmark and "S" to denote a "system- level" benchmark. CINT95 and CFP95 are component-level benchmarks. Q4: What components do CINT95 and CFP95 measure? A4: Being compute-intensive benchmarks, these benchmarks emphasize the performance of the computer's processor, the memory architecture and the compiler. It is important to remember the contribution of the latter two components; performance is more than just the processor. Q5: What component performance is not measured by CINT95 and CFP95? A5: The CINT95 and CFP95 benchmarks do not stress other computer components such as I/O (disk drives), networking or graphics. It might be possible to configure a system in such a way that one or more of these components impact the performance of CINT95 and CFP95, but that is not the intent of the suites.