The main reason why this article appeared is simple: our lab got hold of a motherboard based on Intel 925XE, which allows manual FSB frequency control and CPU and memory voltage rise. That's why we managed to have Pentium 4 560 (Socket 775, Prescott core, Hyper-Threading support, standard FSB frequency – 800 MHz) operating stably at the FSB frequency of 1066 MHz retaining standard frequencies for the other devices. Thus, we obtained "almost original" Pentium 4 3.74 GHz, operating at 1066 MHz FSB.
We have chosen for comparison (which is logical) the only CPU from Intel operating at the same FSB frequency, being its standard frequency: Pentium 4 Extreme Edition 3.46 GHz. It would have been obviously much more interesting to compare two processors operating at the same frequency, but, alas, Prescott's multiplication factor cannot be less than 14, even in engineering samples. But two columns on the diagrams look too lonely, that's why we have additionally tested a new platform from ATI, which only recently appeared in our lab: Radeon Xpress 200 chipset - the first chipset for AMD64 platform with PCI Express x16 support. Unfortunately, the recently announced top AMD processors - Athlon 64 4000+ and Athlon 64 FX-55 - have not reached our lab yet, so we took Athlon 64 3500+. But we solved this problem easily: Athlon 64 3500+ system features as a guest star in this test, that is it can be considered only as a reference :). Obviously both the ATI platform and top CPUs from AMD are worthy of a separate article, and we are sincerely hoping to publish it soon enough...
However, this article has an inner bottom: it reflects some changes, which will be introduced to our standard method of testing CPU performance in the nearest future. The tests below are not yet finalized, but you can already evaluate the general trend of changes. The same issue explains the small set of results: all the programs either got upgraded or their reading methods changed, so we cannot use all the data from previous articles.
I think that the main idea of the article is clear even without extra explanations: sooner or later we are going to see Pentium 4 processors on Prescott core operating with the 1066 MHz FSB. Why not try to look to the future and estimate their performance now? It's always tempting to see test results of a processor, which is not yet officially released, admit it. Certainly, we cannot guarantee that our results will perfectly match the future tests. However, practice shows that they are not that different from those obtained by such non-standard ways.
CPU RightMark 2003 beta 5
From now on we completely switch to CPU RM 2003, as it's quite stable now. "Beta" status in this case means that our programmers have not yet added all the functions planned in the release. Note that the situation with performance has slightly changed - processors on NetBurst core demonstrate higher results now. But having compared the absolute values with the results of the previous tests, we didn't notice CPUs (including AMD) operating any slower, it's just that the P4 code has been optimized. And certainly, the improved SSE3 support has its effect in case of Prescott. Surely AMD processors have their own features to be supported (e.g. AMD64 architecture), however the protracted release of 64-bit Windows for AMD64 does not allow to work on this problem.
Math Solver (physical model)
Considerable breakaway of Prescott is the only issue worthy of a discussion here. It cannot be explained solely by the increased frequency: it is higher than in Pentium 4 XE only by 8%, while according to the test results Prescott @ 1066 FSB outscored it approximately by 16%. We are only to assume that those additional 8% of gain are due to the SSE3 optimization. However, we should also note the drastically improved hardware prefetch in Prescott core - it should have made its presence felt with fast memory.
In case of Intel processors, the renderer operated in two-threaded mode (as they both support Hyper-Threading), and basically brought no surprises - NetBurst core demonstrated the best results here, as it did before.
SPEC viewperf 8
We returned (or, one might say, tried to return) to CPU tests using SPEC viewperf, as the 8th version of this benchmark is considerably updated and contains a lot of potentially nice applications: CATIA, Maya, SolidWorks, Pro/ENGINEER. But we must make one important comment here: viewperf cannot be considered a test, which results fully reflect processor performance in the given applications, because it reflects only one aspect - rendering speed. Besides, viewperf results depend on a video card and its drivers to a considerable degree.
The highest rendering speed in 3ds max was demonstrated by Intel processors, but this is actually not that important (remember that the AMD platform acts as a guest star in this test and it cannot be directly compared with systems based on Intel CPU, not being a top solution). Interestingly, both Pentium 4 processors are on a par. Most likely, this subtest is limited by the performance of a video card.
The situation here resembles that of the previous test.
...The same picture again...
And here we can speak of the obvious heavy load on the processor and some architectural preferences. The latter are most likely expressed by the performance - pipe length relationship - the results are arranged in neat regular steps, the best result being demonstrated by Athlon 64, and the worst - by Prescott.
And the Maya subtest is evidently cache-hungry. Of course, you cannot draw conclusions only from a single test of three processors, but Prescott hardly managed to catch up with Athlon 64 3500+, having a cache twice as large and a huge frequency advantage. But P4 Extreme Edition outscored them all anyway, most likely due to its 2 MB L3.
...Video card is the limit here again... To all appearances, if the AMD processor had a higher clock and a larger cache, all the three columns would have been aligned in a row.
Approximately the same situation, only the gap is still smaller.
Actually, there is nothing to comment on.
Well, "return to SPEC veiwperf", despite its updated set of applications, brought no special surprises. To all appearances, we shall publish only the general mark for the entire test in the next articles: the majority of separate subtests are hardly worthy of rapt reviews.
SPECapc for 3ds max 6
On the whole, 3ds max 6 performance tests with the help of a script from www.spec.org should be acknowledged more informative than the set of scenes we used for rendering. So we decided to leave the above mentioned set of scenes for the expanded tests and use the package from SPEC in usual tests. The fact is that it reflects better the work process in 3ds max, because it combines the rendering process proper with other operations. You can even see how it happens by downloading a 10 MB video file in Windows Media Video format (certainly, it does not show the entire testing process, but you can get an idea). We provide the overall time of script execution as a resulting value, though more detailed statistics are available as well.
The "future" Prescott demonstrates the best result, though it can be written off due to higher frequency (the gain is actually comparable). The system based on the new ATI chipset and Athlon 64, unfortunately, didn't pass this test (it would constantly freeze somewhere in the middle). Nothing serious, we'll write it off for a "raw" sample.
Adobe Photoshop CS (8.0)
The new (for our method) Adobe Photoshop version and a new test script. It was considerably expanded and is still in the development stage (basically, the number of various operations is changed to obtain better balance), so these results must be considered interim. The script now uses the following set of actions and filters:
High memory operating speed, "optimization" of major used functions for multi-threaded operation (this was not an end in itself, they are actually used more often than the other!), relatively large L2 cache, and frequency of course - that's the list of key parameters that we think procured victory for Prescott. And even the huge L3 did not save P4 XE from defeat in this test...
Compressing Multimedia Data
Remember that our standard test settings for MP3 encoding with variable bitrate (VBR) do not allow direct bitrate control, that is it is completely up to the codec. But the use of q=0 parameter (the highest CBR quality) in this version significantly slows down the operation and makes the codec cache-hungry. That's why we used a more standard setting as an alternative --alt-preset standard.
LAME (standard iXBT.com test)
When q=0, a large cache is of critical importance, that's clear.
LAME (--alt-preset standard)
...But with --alt-preset standard all processors demonstrated approximately the same results. A paradox appears in this connection: none of the presets can be considered good from the point of view of its usage in the tests. In the first case we can see the difference, but this preset is somewhat "synthetic". In the second case, the preset is rather relevant from the point of users, but it shows practically no differences between processors...
Windows Media Audio 9
Another candidate to elimination: very quick and very nondiagnostic. But in this case audio data compression is endangered to drop out from our method in pleno...
Mainconcept MPEG Encoder
Updated MPEG Encoder supports SSE3, and one can assume that it procured Prescott's victory on a 1066 MHz bus together with high operating speed with memory and traditionally high gain from Hyper-Threading. There is a second option possible: in its competition with Pentium 4 XE everything was resolved by the frequency, and the large L3 of the latter was left out of job adding no points.
It's difficult to tell DivX preferences of this version so far: either to a large cache or to a short pipeline (remember that it's longer in Prescott core than in Pentium 4 XE).
Windows Media Video 9
A similar situation. But the fact that even non-top AMD processors successfully catch up with our "future Pentium 4" is alarming. By the way: Athlon 64 3500+ and the new ATI platform were "invited" into this test exactly to track such situations: it's natural when A64 is outscored by the other two contenders - it's weaker by all parameters. But when it goes on a par with one of them, it means that top AMD processors will probably ride high in this test.
Previously registered bugs were obviously fixed in the new XviD version, and now the situation with Pentium 4 performance changed cardinally, especially with respect to the new core. Well, it's nice when the new is faster than the old: at least it means that we are paying not for nothing...
Large cache in Pentium 4 XE obviously had its effect here. While Prescott looks not very good - it's again almost on a par with Athlon 64 3500+. Even 7-zip free-threading and the highest frequency among the contenders together with fast memory were of no help.
WinRAR would show preference to AMD before, so nothing has changed. And the large cache was of no help to P4 XE - Prescott looks quite advantageous against it.
No comments... DOOM 3 is hardly a good test for processors. Of course, it's not a top video card - but what's the point in testing a CPU with software, for which a graphics chip is much more critical?
It's a game, which, according to Andrey Vorobiev – hardware editor of the 3D Graphics Products section, heavily depends on a CPU. The latest version includes two built-in benchmarks, we decided to check both of them.
The new C5L2 is much better for CPU testing. And the results look quite optimistic for the Intel platform: ultraexpensive (and thus not very popular among common users) Extreme Edition is no good, while Prescott at 1066 MHz FSB demonstrates decent results.
Unreal Tournament 2004
Ons_dria demo on the ATI Radeon Xpress 200 + AMD Athlon 64 3500+ platform just showed blue screen of death (BSOD), that's why the diagram does not contain results for this testbed. However, both diagrams produce an impression that this game does not depend heavily on a processor. Still, let's wait for the results of top AMD processors. Only they can be shown at advantage in Unreal Tournament 2004 (according to the breakaway demonstrated by Athlon 64 3500+ in Primeval demo).
Let's precondition that a guest star is a guest star, it's used only to make a background. So we shall not touch the results of the new ATI platform and Athlon 64 3500+ processors in our conclusions. This topic will be reserved for separate testing (which is round the corner).
What concerns the comparison of results demonstrated by the first Intel processor officially supporting 1066 MHz FSB with the performance of an overclocked Prescott, on the whole it looks peachy for the latter. Of course there are some crevasses, but not many. And in many cases Pentium 4 Extreme Edition even lags behind, sometimes you cannot explain it even by the frequency difference. To all appearances (though it's too bold to draw conclusions based on a single test), increasing the core frequency and the memory system performance is strategically more correct than increasing the cache capacity. But on the other hand, store is no sore, so that could be combined...
Thus, on the whole, the future of Prescott looks good after the FSB is upgraded to 1066 MHz. All we have to do is wait for a CPU based on this core, which will officially support it. And if there appears Prescott with 2 MB L2 (there are some rumours...), it will have all chances to combine the best properties of both Intel processors represented in this test.
Stanislav Garmatiuk (firstname.lastname@example.org)
November 29, 2004
Write a comment below. No registration needed!