Юмор

Томас Пабст,  28 ноября 2000 Страница:   1 2 3 4

Another Recount?

My recent article 'Painting A New Picture Of Pentium 4' has caused a lot of commotion in the processor scene, unfortunately also resulting in several misunderstandings. Numerous people are comparing this third Pentium 4 article with the latest 'hand recount' issue in the State of Florida. I appreciate the analogy, but it should also make understand that evaluating Pentium 4 properly is just as meticulous as finding the next president. I hope that we can at least do without the supreme courts. I can see the need to clarify a few things, which seemingly weren't pointed out well enough in the initial article.

Intel's Involvement Unfair?

I received several emails that considered Intel's involvement into the MPEG4-benchmark process as unfair. The point was made that supplying Pentium 4 optimized executables was an unlawful way to show this CPU in a better light.

Although I can understand this point of view, it actually misses the point that I was trying to make. The latest FlasK-benchmarks were in no way intended to make Pentium 4 the winner of a benchmark in which it had initially been the winner and then the loser. My intentions to publish those new results with Intel-optimized executables were merely to show how much the performance of Pentium 4 can be improved once optimized code is being used.

I tried to remind you of the fact that AMD had not supplied me with Athlon-optimized executables so far, which is why the latest FlasK-benchmark results are in no way useful to compare the performance of Pentium 4 and Athlon right now. Once I have received those executables, it is fair to compare Athlon with Pentium 4 by using optimized software for each of the two.

I have tested the Athlon-optimized code from Alexander Goodrich last night (November 28, 2000, 12:35 AM) and so far Athlon is not able to surpass Pentium 4 with Athlon-optimized x87 FPU code and the 3DNow!-optimized code is way slower than the Pentium 4 double-precision SSE2-optimized one. This is not the last word yet, because Alexander is still working hard on further optimizations. However, even if Athlon should end up losing this contest it doesn't mean that Pentium 4 is the better processor. As I already said, the FlasK-issue was less of a comparison between Athlon and Pentium 4. It was more of an example to show what is actually possible. The actual software world out there looks a lot different. Sudden optimizations for certain processors simply don't happen in the real world. As long as current applications don't get replaced by new ones, Athlon is still ahead of Pentium 4 in the majority of situations.

I have also asked AMD itself to work on an Athlon-optimized FlasK. Once we received those executables from AMD we will supply you with valid and final Athlon-results.

feed_id: 6177 pattern_id: 2818

Страница:   1 2 3 4

СОДЕРЖАНИЕ |Одной страницей|Версия для печати

Another Recount? Intel's Involvement Unfair? FlasK Meaningless? General Opinion About P4 GPL-issue Pentium 4's Bus Clock Is NOT 400 MHz! 3D Studio Max Issue Last Words  Обсуждение в Клубе Экспертов THG