Pentium M 1.6 at 3.1 GHz

[ANova]

SuperPi 1M - 23.766 seconds

Though what's even more impressive is they managed to get it to 2.8 GHz on air.

http://www.madshrimps.be/?action=getarticle&number=1&artpage=1359&articID=342

 

[RussSchultz]

That's pretty impressive.

Now if Intel can get the Pentium-M with a dual core/dual memory bus and running at those speeds retail...

 

[Moloch]

That's pretty impressive.

Now if Intel can get the Pentium-M with a dual core/dual memory bus and running at those speeds retail...

QFT, that would be truely fast!

 

[ANova]

Well a dual core version codenamed Yonah will be coming next year with a higher fsb (667 MHz) but it won't have a dual memory bus.

 

[Blazkowicz]

this is depressing, my PC does 70 seconds

 

[suryad]

Details on the SuperPi test...please...if anyone has got any. How fast does AMDs finest run that test at anyway anyone know?

 

[Karma Police]

Dothan at 3.1GHz isn't bad. Not bad at all.............unless..............

You're the famous Japanese overclocker Kaz-n and have a Dothan at 4GHz!!!:

http://www.xtremesystems.org/forums/showthread.php?t=66246

SuperPi 1M in 19 seconds, too! 8)

 

[ANova]

Details on the SuperPi test...please...if anyone has got any. How fast does AMDs finest run that test at anyway anyone know?

SuperPi calculates the number of decimal places after Pi (3.14...). This test was set to 1M which means 1 million decimal places.

A stock FX-55 can do it in ~32 seconds. Overclocked to 3.2 GHz 25 seconds.

 

[pascal]

That's pretty impressive.

Now if Intel can get the Pentium-M with a dual core/dual memory bus and running at those speeds retail...

Add an onchip memory controller for very fast, low latency access 8)

 

[suryad]

Its kind of unbelievable to think that the Dothan is slight bit faster than an FX processor!!!

 

[AlNom]

Tom did some benchmarks awhile back. It's about on par when you consider a bunch of different programs. (Sometimes it's faster, sometimes not, but always in the same ballpark)

 

[Fox5]

That's pretty impressive.

Now if Intel can get the Pentium-M with a dual core/dual memory bus and running at those speeds retail...

Add an onchip memory controller for very fast, low latency access 8)

And a pipeline as wide as the athlon's so it can outperform it in all programs and not just some.(though if it can clock higher, then it doesn't need the wider pipeline, but I believe intel has a wider pipeline planned for the desktop p-m)

Btw, 48.390s on my high end XP .

If the P-M clocks so well, how come Intel limits it to 2.1ghz? Does power consumption suddenly skyrocket, making it unsuitable for laptops? It'd be interesting if P-Ms pulled a Prescott at higher speeds.(supposendly the later model P3s did)

 

[RussSchultz]

Considering the Dothan can match Athlon 64 performance at a lower frequency, I'm not sure its pipeline needs much of anything, assuming it can run at these higher speeds.

 

[Fox5]

Considering the Dothan can match Athlon 64 performance at a lower frequency, I'm not sure its pipeline needs much of anything, assuming it can run at these higher speeds.

Not in all programs, sure in games and some other programs it does, but there are some where it's horribly destroyed. Won't matter for most home users, but if Intel wants to position the P-M for the workstation/server market, it's something that Intel needs to work on.

 

[RussSchultz]

It doesn't do so well in content creation because its memory bus is slower. It also lacks full SSE2, if I'm not mistaken.

 

[Fox5]

It doesn't do so well in content creation because its memory bus is slower. It also lacks full SSE2, if I'm not mistaken.

I believe the single channel Athlon 64s also whoop it in content creation, and I believe it does have full SSE2, it just lacks SSE3.

 

[RussSchultz]

When I say "full support", I mean a comprehensive implementation that is an accelerant, not just being capable of executing the instructions.

 

[Fox5]

When I say "full support", I mean a comprehensive implementation that is an accelerant, not just being capable of executing the instructions.

Well in that case, don't the AMD chips lack full SSE support as well?
BTW, the Pentium 3 fully supported SSE I believe, the P-M is an upgraded P3, what would have changed from SSE to SSE2 that would limit performance?

 

[l15741]

I can't add much to the discussion of Pi results and the Pentium-M...

However I can say my Athlon 64 3300+ with 512mb of ram does the 1 million in 43 seconds and the 32 million in 39 minutes 8 seconds...

 

[DudeMiester]

Just so you know, while the Dothan is great in 1m, when you start going up to 16m and 32m it's performance is bad. Apparently because the A64's caching sub-system is far superior.