We already reviewed Intel Pentium M performance on the desktop platform not long ago. The only news since then is the official announcement of Pentium M 780 2.26 GHz. It's quite a commonplace event and performance of this processor it predictable: it's no different from the previous Pentium M 760/770 processors, but for its clock. But as this CPU found its way to our lab, we decided to test its performance anyway. At least because there are not so many Pentium M tests on the desktop platform on our web site :). But this time we decided to focus only on the maximum performance possible. So we carried out our tests only on a motherboard based on the i915 chipset with an adapter from ASUS (this configuration is described in detail in the previous article on a similar topic).

The adapter with a processor installed into Socket 478 on a motherboard.

The same components with an installed cooler
(it's a special cooler that comes shipped with the bundle of this adapter)

Thus, Pentium M 780 + i915 demonstrates the maximum performance we can squeeze from an Intel mobile processor on a desktop without overclocking. Besides, as Pentium M 780 is faster than the 770 model, we decided to add another two processors to Athlon 64 3700+ and Pentium 4 540 — Athlon 64 4000+ and Pentium 4 560. They are to mark the upper limit, that is to show what "truly desktop" processors cannot be outperformed by the mobile solution from Intel so far, even on a desktop motherboard with dual-channel memory.

Testbed configurations

• Processors
  • AMD Athlon 64 3700+ (2.2 GHz core, 1 MB L2)
  • AMD Athlon 64 4000+ (2.4 GHz core, 1 MB L2)
  • Intel Pentium 4 540 (1 MB L2, 800 MHz FSB, 3.2 GHz)
  • Intel Pentium 4 560 (1 MB L2, 800 MHz FSB, 3.6 GHz)
  • Intel Pentium M 760 (2.00 GHz core, 2 MB L2, 533 MHz FSB)
  • Intel Pentium M 770 (2.13 GHz core, 2 MB L2, 533 MHz FSB)
  • Intel Pentium M 780 (2.26 GHz core, 2 MB L2, 533 MHz FSB)

• Motherboards
  • Intel D955XBK Desktop Board (Socket 775, i955X)
  • EPoX EP-9NPA+ Ultra (Socket 939, nForce4 Ultra)
  • ASUS P4GPL-X (Socket 478, i915)

• Memory
  • 2x512 MB PC3200 (DDR400) DDR SDRAM DIMM 2-2-2-5 (Corsair TwinX)
  • 2x512 MB PC5400 (DDR2-533) DDR2 SDRAM DIMM 3-3-3-8 (Corsair)

• Video card: ATI Radeon X800 (256 MB)
• HDD: Samsung SP1614C (SATA), 7200 rpm, 8 MB Cache
• Windows XP Professional SP2, DirectX 9.0c
• ATI CATALYST 5.4 (Display Driver 6.14.10.6525)

Test results

Remember that diagrams with all test results (61 items!) are published on a separate page — without comments, just as is. The article provides only summary diagrams that calculate the results of entire test groups into average scores. This approach appeases curiosity of the most inquisitive readers, who are against cutting down the number of test results published in our articles, and still makes the article less motley and graphics-intense. What concerns our comments, real professionals (who are interested in details) are expected to need none of them.

SPECapc for 3ds max 6 (3ds max 7.0)

Detailed results

Pentium M features almost a linear performance growth, which proves that the 533-MHz bus is still sufficient at such frequencies. As always, Pentium 4 does not look very good in this test due to a serious defeat in the Interactive sub-test (see detailed results). Symptomatically, Pentium M fairs very well in this very sub-test as well as Athlon 64 — it means that their architectures are similar in terms of preferences of the program. That's understandable: both of them have a short pipeline.

SPECapc for Maya 6 (Maya 6.5)

Detailed results

And here the Athlon 64 stands off, the main fight is between Intel processors, Pentium M 780 being the fastest. Proceeding from this result, we can draw an interesting conclusion about the test itself: as the CPU with the lowest memory bus throughput managed to win, it would be logical to assume that SPECapc for Maya is indifferent to this parameter. On the other hand, Pentium M has the largest L2 Cache. Let's consider we have found an answer...

Lightwave 8.2, rendering

Pentium 4 560 turns out the best processor at rendering, the 540 model also acts well, having been outperformed only by the Athlon 64 4000+. In our opinion, Pentium 4 owes its good results to virtual multi-processing support — Hyper-Threading, as Lightwave supports multi-threading. Pentium M performance scales well with the clock, but there is nothing unusual about it for this task: render engines are usually indifferent to memory bandwidth.

SPECapc for SolidWorks 2003

Detailed results

Quite a predictable result for those who have read the previous article.

Adobe Photoshop CS (8)

Detailed results

Pentium 4 processors are the best here (probably again due to Hyper-Threading), Pentium M 780 is in between Athlon 64 3700+ and 4000+ in terms of performance. The "steps" of Pentium M processors is absolutely even, which proves their excellent scalability, at least within the reviewed clock range.

Adobe Acrobat 6.0

All-purpose data compression (archiving)

Detailed results

It's probably the only class of applications that shows memory bandwidth dependence. Pentium M still scales evenly — but the step size is much lower.

Multimedia lossy compression (MP3/MPEG2-4)

Detailed results

There is no point in commenting on the average score, but we'd like you to pay your attention to the detailed results. You can see well that AMD processors (traditionally) "failed" the LAME encoding test (Q=0 option). But these are general results, not as interesting as "drops" of the Pentium M. The latter didn't do well in MPEG4 / Windows Media Video sub-test and utterly failed the MPEG2 / Canopus ProCoder sub-test. As they say: "for your information..."

CPU RightMark 2004B

Pentium 4 processors set the fashion due to their Hyper-Threading support and good optimization of the CPU RightMark render engine to multi-processing.

3D games and graphics visualization
in professional packages

DOOM 3

Far Cry

Painkiller

Unreal Tournament 2004

Total score in games

Detailed results

The short-pipeline Pentium M crushes the Pentium 4 in games, but it still cannot compete with the Athlon 64. Quite unassuming step size in the line Pentium M 760 —> 770 —> 780 speaks of an important role played by the memory bandwidth in this class of software.

SPEC viewperf

Detailed results

The difference between our participants is beggarly here, which hints at an insufficiently powerful video card.

Conclusion

On the one hand, Pentium M 780 is not that bad. In most cases it fairs rather well against the fast desktop processors from Intel and AMD. On the other hand, let's not forget that we compare the fastest Pentium M processor with non-top desktop CPUs. In addition to Athlon 64 4000+, AMD offers Athlon 64 FX-55/57; Pentium 4 560 is outperformed at least by Pentium 4 570, to say nothing of the 660/670 and eXtreme Edition 3.73 GHz. So: "Not bad indeed, but nothing more". Of course, the Pentium M has its own trump cards — low heating, low power consumption. However, in the first place, users require performance from their desktops (why else you need a desktop?) If you take into account that building a desktop with a top Pentium M processor may be even more expensive than the same system on Athlon 64 FX-57, we can only establish a fact that desktop computers on Pentium M will most likely remain an interesting toy for "quiet PC" fans and other IT aesthetes. There is nothing good or bad about it, it's just a fact...