Slashdot Mirror
Centrino Mobile Equals Desktop Pentium 4 in Speed
Spy Handler writes "On Wednesday during the launch of its new Sonoma Centrino Mobile, Intel put on a demonstration running a video game on a laptop. It matched the performance of a high-end Pentium 4 desktop running the same game, declared Intel. The contenders were a laptop sporting a 2.13 GHz Pentium M processor, 1GB RAM, and the Alviso chipset versus a desktop with a 3.6 GHz Pentium 4 with hyperthreading, 1GB RAM, and the Grantsdale chipset. Is this a testament to how far the Pentium Mobile architecture has come, or a sad comment on the clockspeed-pushing design of the Pentium 4?"
Intel's finally learning the lesson everyone else knew about 5 years ago. Too little, too late? Or can Centrino save them?
hmm I would like to know which video game it ran to get equal performance. Also, was the game software rendered or was there a graphics chipset involved?
I am sure that they got together with NVidia and came up with some crazy optimized drivers (read: cheating).
Sounds like Intel is ready to write off the P4 as done for, and is putting all of thier eggs in the Centrino basket until the launch of their dual-core chips...
As an earlier poster mentioned, most newer games depend more on the GPU than the CPU; anything over 2Ghz is almost overkill.
Intel and AMD are in the awkward position of needing to create a market for new processors in a world where a 1Ghz processor will do most office tasks brilliantly. They pushed speed, speed and more speed for so long that the average consumer doesn't give a whit about HyperThreading or anything else. Tech heads and researchers and universities are different, but is that enough to support to very large chip manufacturers forever?
Don't be a looter...and yes, I know that it's spelled with an "A" instead of an "E".
Is this a testament to how far the Pentium Mobile architecture has come, or a sad comment on the clockspeed-pushing design of the Pentium 4?
Or perhaps a testament to how fill-rate limited the game was? Honestly, what was the game? Doom 3? Or Monkey Isnald 3? It makes a difference.
Refuse to make a statement in your sig!
What about "both"?
Or, of course, as the parent points out, perhaps the bottleneck is somewhere completely different. Maybe the processor speed is less important than something else (gasp!).
Personally, I'm much more excited by increases in network, disk and bus bandwidth than CPU. I don't spend much time waiting for my CPU; I spend time waiting for data to get to my CPU. YMMV.
Disclaimer: I work for a company, but I don't speak for them.
So what.. if Intel is going to compete with AMD, they need to make a 64 bit version of the P-M chip. AMD already has a mobile Athlon 64.
What about comparing them in 64 bit envrionments?
Oh, the P-M can't handle 64 bit?
That's too bad..
Not a Twitter sockpuppet... but I wish I was.
In the good old days, when the P-III was being replaced by the P-IV, you could roughly multiply the P-IV speed by 0.6 to get the comparable P-III speed. (Specifically, a 2Ghz P-IV was about the same as a 1.2Ghz P-III.)
So a P-IV at 3.6 Ghz, is roughly equal to a 2.16 Ghz P-III.
And guess what? The Pentium-M is a repackaged version of the Pentium-III core.
Maybe Intel and AMD and the rest of the world will start using something useful, like SPEC results to market their processors, instead of Hz ratings?
the front side bus speed is the biggest limiting factor on Pentium M processors.
Thank you... So far, I consider this the only "insightful" comment in this entire topic.
In terms of raw performance, though, Anand and Tom (of which you mention the latter) have both done "real world" tests that don't include the GPU as the bottlenext, and found that, for heavily CPU-bound tasks (such as compression, which also eats memory but mostly just CPU), the Pentium-M (Dothan, in particular) holds its own against both the Prescott (P4) and the Athlon 64. On some tasks any of those three would take the lead, though the Dothan does only take 2nd or 3rd most of the time (but still beats the Athlon XP and the Northwood P4).
For second best, and less than a quarter of the power consumption (less than a tenth when idle) for comparable performance, I fully plan to get a Pentium M as my next desktop upgrade. I care about raw performance, but I also care about my electric bill and about having something that sounds like a jet engine three feet from my head (lower power = less cooling needed = quieter).
PLUS they do the stupid thing here and put in DDR-2 which does little for performance but increases system costs.
Strange opinion... Yes, it increases the system cost a tad, but consider it from two POVs...First, since the Centrino line primarily targets laptops, 2.5V vs 1.8V means significantly lower power consumption (and correspondingly less need for active cooling, making battery life even better). And second - DDR2 picks up where DDR stops, FSB-wise... You could just as well say the original P4s did nothing for performance over the best-of-breed PIIIs, but after three core gens and a doubling of the clock speed, no one would now claim a "modern" PIII will outperform a modern P4.
Intel recently dumped a major portion of their roadmap citing heat and power issues. What most expect to see in the future are new chips built on the Pentium M architecture. I think this demo is nothing more than to get "us" all talking about how amazing it is how the M can stand up to the standard P4. Then when Intel makes the transition, the GHz shock is not so dramatic. This corresponds to Intel's move to start code numbering their chips - why in the world would Intel dump their marketing lead of using GHz ratings for marketing and start using obscure codes like AMD? Of course AMD did it because they did not want customers focusing on GHz. When Intel starts using the M as the core, their new systems will take a step backwards in GHz, and that's much easier to hide if all your chips are known only by their codes and everyone already thinks the M core is similar to P4 in performance.
Pentium-M has 2MB of on-die cache.
Athlon 64 has a cache size of 512kb or 1MB; interestingly 3800+ and 4000+ both run at 2.4Ghz but the latter gains in speed by having a double sized cache... so what if we add more cache?
To effectively compare Pentium-M architecture (by architecture NOT price!) to Athlon 64 we would need a 2MB cache version of it...
Isn't the Pentium-M based off the PIII core?
No. Or at least, not proveably so.
Intel has released very few architectural details of the Centrino line. From what little the public actually knows about it, it does seem more similar to a PIII than a P4, but by all (credible) accounts, it uses a complete core redesign, optimized based on different criteria than most desktop CPUs. As a result, it consumes a reasonable amount of power, and the performance seems like almost an unintended perk.
Google cannot tell me which tools are good. An experienced user, can.
Is this a testament to how far the Pentium Mobile architecture has come, or a sad comment on the clockspeed-pushing design of the Pentium 4?
The Pentium III has veen embarassing the Pentium 4 as long as the Pentium 4 has been shipping. This is merely another act in the continuing Greek Tragedy that is the Pentium 4.