Intel's Pentium M on the Desktop - A Viable Alternative?
by Anand Lal Shimpi on February 7, 2005 4:00 PM EST- Posted in
- CPUs
Floating Point Performance
Just about a year ago, our own Johan De Gelas made an extremely interesting point about one of the weaknesses of the Pentium M - floating point performance. The theory is this - the Pentium 4, Athlon 64 and Pentium M all have very different platforms, with equally different characteristics. Unfortunately, as we've already shown, the Pentium M is quite possibly the worst off with only a single channel 333MHz DDR memory bus. It's also widely known that most floating point intensive applications are highly memory bandwidth limited, meaning that the Pentium M already has an excuse for poor floating point performance - it doesn't have enough memory bandwidth.But what if we are able to take memory bandwidth out of the equation? This is where a little benchmark called "flops" comes into play. The beauty of flops is that it executes entirely within the L1 cache of the Pentium M, meaning that the benchmark is limited by two things: the performance of the Pentium M's L1 cache, and more importantly, the performance of the Pentium M's floating point and SSE units.
The actual tests that flops runs are a mixture of floating point add, subtract, multiply and divide operations. The mix of ADD/SUB, MUL and DIV operations is listed next to each test in the table below.
We compiled flops using the latest Intel C compilers to give the Pentium M as solid of a foundation as possible using the /O3 and architecture specific flags under Visual Studio .NET. All of the results are expressed in MFLOPs, higher scores being better:
| Test (% ADD, SUB, MUL, DIV) | AMD Athlon 64 3200+ (2.0GHz) | AMD Athlon 64 FX-55 (2.6GHz) | Intel Pentium 4 3.2GHz | Intel Pentium M 755 (2.0GHz) |
| 1 (50,0,43,7) | 1576 | 2057 | 1274 | 899 |
| 2 (43,29,14,14) | 856 | 1118 | 790 | 492 |
| 3 (35,12,53,0) | 1388 | 1802 | 2476 | 1470 |
| 4 (47,0,53,0) | 1244 | 1622 | 2792 | 1601 |
| 5 (45,0,52,3) | 1477 | 1923 | 2351 | 1019 |
| 6 (45,0,55,0) | 1466 | 1908 | 2762 | 1607 |
| 7 (25,25,25,25) | 458 | 595 | 365 | 252 |
| 8 (43,0,57,0) | 1585 | 2065 | 2566 | 1572 |
| Average | 1256 | 1636 | 1922 | 1114 |
The first comparison to look at is the Athlon 64 3000+ vs the Pentium M 755, since both CPUs run at the same clock speed. Despite the Pentium M's improvements to enhance IPC, the Athlon 64 is still able to outperform it at a core level (without the aid of its memory controller) by almost 13%. But here's where the next Athlon 64 score comes into play - while the Pentium M will hit 2.26GHz by the end of this year, the Athlon 64 will be at or above 3.0GHz. So, the headroom of the Athlon 64's architecture gives it a huge performance advantage here in flops as you can see by the Athlon 64 FX-55 results (remember that the larger L2 cache of the FX-55 has no effect on the flops results as the program runs entirely out of L1).
Next, we have one of the slower Pentium 4s vs. the Pentium M 755. Why not compare to a 3.6GHz or the new 3.8GHz Pentium 4? Well, look at how much the Pentium 4 3.2GHz outperforms the Pentium M 755 - 72% using Intel's 8.1 C++ compiler. When running optimized SSE2/3 code, the Pentium 4 is a much stronger FP performer than what the Pentium M ever could be, which is very important for the following reason: the future of desktop applications is in very floating-point intensive media transcoding tasks, and for those applications, the Pentium M just won't cut it. So, to those who feel that Intel will soon ditch Net Burst in favor of the Pentium M's architecture, the results speak for themselves. While elements of the Pentium M architecture will undoubtedly make an appearance in the Pentium 4's successor, its dated P6 execution core will not.
Facebook
Twitter
RSS
77 Comments
View All Comments
saratoga - Tuesday, February 8, 2005 - link
Overall this artical brings up a lot of the points missing in other Dothan reviews. Very nice work. Too many people have looked at a few benchmarks, bashed Intel for the P4, and missed the whole issue here.Intel isn't stupid. Its obvious they don't think Dothan will work in its current form as a desktop chip, and thats why they're still sticking with Prescott at the moment, and only bring the P-M over much later in a reworked form with Yohan. Assuming they ever do introduce a desktop chip based on the P-M.
Also, siginificant scaling out of Dothan seems unlikely. They'll probably get a few more speed grades out of it, but whoever was saying 3GHz was dreaming. Maybe at 65nm, but that sure as heck won't be dothan, and it won't be for a while yet.
PrinceGaz - Tuesday, February 8, 2005 - link
Well put classy, the P-M is a chip that at least in its current form can never be a desktop processor because of severe weaknesses in several areas.A faster dual-channel chipset will never make up for its poor FPU performance in heavy-duty applications, something I'd heard about many months ago but hadn't seen reliable benchmarks of until now.
If you want to do word-processing or browse the web, I'm sure the P-M will be very efficient. If you want to run the sort of spplications that seriously test a processor and are the reason you'd buy it in the first place for a desktop PC, then the P-M falls far short of the mark, in fact it is so far behind at times that it is embarrasing.
But you don't get anything for nothing, the P-M is great at doing easy stuff very quickly which is what laptops are used ofr mainly; but when the going gets tough, you want a real desktop processor like the A64 to keep things moving.
classy - Tuesday, February 8, 2005 - link
T8000What part don't you understand? The Pentium M has been reviewed all over the net. Out of all the reviews only one reviewer hit 2.8. Everyone else, was similar to Anandtech's results. 2nd I don't no where you been, but every review of an FX55 I have seen it routinely hits 2.8 with no problem. And almost all the lower speeds hit the 2.6-2.7 ballpark. Not mention that a small increase with A64 is much more signifcant than even a modest Intel OC because of the architecture of the A64 cpus. Hey everyone has a favorite cpu, video card, or motherboard maker. But when something is better, its just better. And for anyone to even remotely argue the Pentium M as a challenge to the A64 cpus is a bit silly. This chip reminds me a lot of the old 366@550 celery chips. IF you got a 366 to do 550 it was a great chip because it gave you nice performance for the price. The Pentium M doesn't have a price advantage and is on a platform that is outdated. IF you can overclock it to decent levels it performs pretty good in some aspects but still sucks in many others. The problem is IF. But as I stated ealier IF is out for the evening with MAYBE.
LackofVision - Tuesday, February 8, 2005 - link
I couldn't disagree more with the conclusions in this article.Anyone who can't see the promise of a desktop processor design based on the banias in't going beyond just looking at the numbers. Especially when you start thinking down the road about dual core's and the heat and performance bottlenecks associated with them.
So because the banias can't outperform the p4 or athlon64 in every benchmark, when hamstrung by an outdated chipset, and designed primarily for low power usage, the processor won't be competitive when running on a modern subsystem with a re tuning of the core design to make it more suited to the desktop?
Nothing like comparing apples to oranges and then drawing a conclusion on what a pear tastes like.
jamawass - Tuesday, February 8, 2005 - link
I doubt it, Intel makes huge profits by putting a price premium for mobile processors. They won't jeopardize this for a few enthusiasts.KristopherKubicki - Tuesday, February 8, 2005 - link
FrostAWOL, #51: What's your point? Those HP blade servers run Pentium Ms and there is no mention of Pentium 4 anywhere.Kristopher
HardwareD00d - Tuesday, February 8, 2005 - link
Pentium M = YawnT8000 - Tuesday, February 8, 2005 - link
#52Since it is very rare to see an A64 CPU overclock above the available speeds without subzero cooling, the comparision would likely be between a 2.4 or maybe 2.6 Ghz A64 and a 2.8 Ghz P-M.
Also, P-M CPU's with higher multipliers usually overclock better due to the limited FSB possibilities of the i855 chipset. This could explain why Anand did not reach 2.8 Ghz in this review.
dobwal - Tuesday, February 8, 2005 - link
While i think that this is a good article. Allowing us to see the performance of the dothan in its current state against desktop cpus. Some of the conclusions that are made by the author don't take account of alot factors.1. "The problem is that in the transition to the desktop world, its competitors get much more powerful, while the Pentium M is forced to live within its mobile constraints."
How can this statement be valid. The mobile constraints on the Dothan is never really removed. Nothing is really done to try to make the mobile dothan mimick a (possible) desktop variation of itself. Do you really think there is a chance for a official desktop dothan running at 2.4 with DDR 333 in single channel with 533 FSB. How about re-running these benchmarks along with a 3.2Ghz P4 with DDR 333 (single channel) and a FSB speed of 533.
2. "The fundamental issue is that although the Pentium M is surprisingly competitive with the Athlon 64 on a clock for clock basis, the Pentium M's architecture can't scale to the same clock speeds that the Athlon 64 can. The fact of the matter is that while the Pentium M will hit 2.26GHz by the end of 2005, the Athlon 64 will be on its way to 3.0GHz and beyond."
The fact of the matter is you are comparing the scalability of the king of mobile chips vs. the scalability of the king of desktop chip and making an assumption without taking account of all the factors involved. The fact is we do not know the scalability of the dothan without its mobile constraints. Even more so, we don't know the true scalability of the mobile Dothan. What other mobile cpu offers the same level of performance vs. battery life.
Its more profitable for a company to retard performance increases of its cpu if there is no other cpu that can offer the same level of performance currently or in the near future. Revisions or new steppings increase cost.
AMD is in the same boat with the A64.
How long has A64 been stuck on 2.4Ghz. Most of the latest PR number increases with relation to A64 have come from HT increases, dual channel and moving from 754 to 939. Imagine the scenario of where the Prescott worked as intended and the Tejas was around the corner. Do you think that the A64 would be still at 2.4Ghz or more like 3.0 or 3.2Ghz.
While some of the conclusions could be seen as true under the circumstances of Intel never officially introducing the dothan to the desktop world. Where all we get are mobile Dothans on chipsets with desktop features.
However, these benchmarks can't prove or disprove the viability of a dothan that was devised by Intel to be a desktop competitor.
classy - Tuesday, February 8, 2005 - link
#55IF Intel does this. IF Intel does that. Unfortunately IF left with MAYBE and they went to the movies to see the new #1 movie from Intel, Could Have, But Didn't, starring Mr Dothan CPU. :)