Linux and the Desktop Pentium M: Uncommon Performance
by Kristopher Kubicki on December 24, 2004 12:00 PM EST- Posted in
- Linux
Introduction
Dothan is something that both perplexes and intrigues us at the same time. Not quite a Pentium 3, not quite a Pentium 4, and not quite something that is entirely different either. Meanwhile, the NetBurst architecture has come under serious strain over the last few years, particularly since Intel's Prescott launch. Is Intel still capable of killer products? And more importantly, do they still dominate on Linux?As many who follow our Windows reviews know, Pentium M on the desktop is something a few years in the making. Even when the original 130nm Banias processor showed up in 2003, reviewers and customers alike were astonished with the technology. Intel received even more praise when their 90nm Dothan chips of the same product line showed up - utilizing less than 30W during peak operation and less than 5W on idle. Most of these advancements were due to Intel's controversial strategy to rethink the P6 architecture and refining a particularly interesting technology called Enhanced Speed Step. Enhanced Speed Step, also known as EIST, gives the operating system the ability to dynamically clock the processor. Typically, Windows will dedicate the full 100% of the Dothan's clock during intensive operation, but throttle the processor as far down as 10% of its capable speed when the computer is just idling. Thus, Pentium M has achieved incredible status among overclockers and HTPC enthusiasts - on Windows. Today, we will briefly explore the versatility of Pentium M on the Linux desktop. Lessons learned should also apply to the notebook market as well.
That being said, there are already a few fundamental flaws with the Pentium M architecture on Linux, the largest of these being compiler optimizations. While Opteron/Athlon 64 and Pentium M share substantial optimizations from every corner of the OSS universe, Pentium M receives very little regular attention. Dothan/Banias are slightly cursed, since most Linux OSes are built on the - mtune=i686 flag, which specifically tunes compilation to the P6 core (Pentium Pro), from which the Pentium M is derived. Why is that a curse and not a blessing? Although Dothan and Banias certainly share some key elements with the P6 architecture, they are far from it. Pentium M's Micro Ops Fusion, local branch prediction and general optimizations across integer division and register access are completely ignored by the compiler, even when setting - march=pentium-m, since most compilers (particularly anything before GCC 3.4.2) tend to just categorize Pentium M as a P6 processor with a higher clock.
Of course, the Intel C compiler, ICC, behaves very differently, but unfortunately, isn't very free either. We have a few tests today that include the non-commercial ICC as well and we see how they stack up against GCC 3.4.1. So, if it doesn't bother you that the majority of Linux sees your new Pentium M as a glorified Pentium Pro, without further ado, let's check out how it actually performs against other processors that we have looked at in the past.
Facebook
Twitter
RSS
47 Comments
View All Comments
bhtooefr - Saturday, December 25, 2004 - link
Problems on the John the Ripper section:DES: Where's the 755?
MD5: Where's the 400MHz FSB 765?
Blowfish: Where's the 533MHz FSB 765?
Also, for anyone who wants to know what the heatsink IS, x86-Secret reviewed this before the heatsink was available, and they used a MicroCool northbridge heatsink.
FWIW, I don't know why nobody's coupled this thing to an i865/875. It's definitely possible, as Shelton (0K L2 Banias) has been coupled to an i845, and Banias has been coupled to an E7501. And, the fact that Alviso is "i915GL" says a lot. Mobo makers should be able to simply rework the traces leading to the socket, and reuse their P4 board designs for a P-M board design. Or, if they're REALLY lazy, they could just make an adaptor - put the processor in it, and drop it in the socket.
ImSoHighRightNow - Saturday, December 25, 2004 - link
"Encode rate, more are better"? Some encoding rates can't be "more" than other encoding rates. They can be HIGHER, but not "more". I would suggest "Encoding rates, higher is better"You know what you need? A grammar handbook. Nothing annoys me more than someone who can't conjugate "to be" correctly. I learn conjugations for other languages, the least you can do is learn conjugations for one verb in English. Thanks
ImSoHighRightNow - Saturday, December 25, 2004 - link
"Rendering time in seconds, less are better"Shouldn't that read more like "Rendering time in seconds, fewer is better" or "Rendering time, a shorter amount of time is better"? At least something remotely grammatical would be preferable. Thanks
miketheidiot - Saturday, December 25, 2004 - link
dothan will need work before its ready for desktop. Simple as that. And make it cheaper two, that would be good. $500 + $250 for a mobo is quite a bit, even if it does OC well.DrMrLordX - Friday, December 24, 2004 - link
Looks like the Dothan falls flattest when it's faced with multiple concurrent threads. Dual-core Dothan solutions might alleviate some of that problem, but, perhaps this is one of the reasons why Intel has been rather shy about pushing multi-core Pentium Ms for the desktop?Googer - Friday, December 24, 2004 - link
It seemed to me that the extra performance 133mhz that the 533 bus provides is rather small. My suggestion is couple this with an 875 or 865 chipset @400mhz and let dual channel memory add the needed performance boost. Its probably the cheapest and most effective way to increase the performance of Pentium M.Pannenkoek - Friday, December 24, 2004 - link
I believe that the price of the processor tested should have a more prominent position in the whole test. It's, after all, about the price/performance ratio for most of the consumers.If you plan on testing GCC vs ICC then I recommend to visit http://www.coyotegulch.com (though the site is "temporarily unavailable"), where you can find comparisons between the compilers, compile settings and more.
The focus of the tests on the site is on scientific applications/algorythms which fit in cache, and is therefore more about how many micro-optimizations are not missed. Which explains why results can vary so much; and also why the ICC compiler, made by Intel for Intel processors, can be sometimes a lot faster than GCC, which does not share the intimate knowledge of the inner working of those processors and targets a zillion other architectures as well...
sprockkets - Friday, December 24, 2004 - link
I wish though you included pics of the fully assembled system. I would like to see that HSF, since it appears AOpen simply uses a 478 type heatsink bracket. But looking at the board at newegg.com, the Aopen board comes with the heatsink, and has DUAL Marvell Gbe, plus it has a SATA controller on it as well, and costs $14 less.MDme - Friday, December 24, 2004 - link
Anand and co: when will you guys do a full review of this chip against P4 and A64 across the different applications, games, etc? that would be a great gift to your readers at AT.JustAnAverageGuy - Friday, December 24, 2004 - link
Page 6:"We also took the same POV-Ray benchmark and ran it against the Pentium M clocked at speeds from 1.0GHz to 2.4GHz."
Should read 1.6GHz
[shake 3.5c]"Although it holds up well against an Athlon 64 3200+"
Doesn't seem like it. It took over 12 minutes longer.