The Pentium-M Crushes the Athlon and Pentium-4 in Seti@home

[VTboy]

Aparently the Pentium-M is the fastest consumer processor when it comes to Seti@home, even a 1.6 GHZ Pentium-M beats a 2.4GHZ Barton. Why do you think the Pentium-M is able to crunch numbers so quickly.

 

[Biggs]

Why.

 

[dnuggett]

Maybe because it is common knowledge that the chip is about 160% faster than a P4 at a given clock rate.

 

[Derango]

I believe the 1MB of full speed cache help with seti. I remember the p3-m used to be able to churn out workunits in 4 hours at 866mhz

 

[alexruiz]

Originally posted by: VTboy
Aparently the Pentium-M is the fastest consumer processor when it comes to Seti@home, even a 1.6 GHZ Pentium-M beats a 2.4GHZ Barton. Why do you think the Pentium-M is able to crunch numbers so quickly.

A 2.4 GHz barton, or a 2400+?

 

[VTboy]

A 2.4GHZ Barton. A 1.6GHZ Pentium-M beats the P-4 even at 3.2GHZ.

 

[alexruiz]

Originally posted by: VTboy
A 2.4GHZ Barton. A 1.6GHZ Pentium-M beats the P-4 even at 3.2GHZ.

Unlikely given tha fact their IPC is close..... unless you have the barton running in a KT133 with PC100 😉

 

[sciencewhiz]

I'm averaging 3.5 hours per work unit on my 1.3 ghz pentium-M. I thought the fastest processors were doing work units in the low 2 hour range.

What about a 2mb cache Xeon?

 

[dnuggett]

I would like to see where your information is coming from. The Pentium M is a conservatively rated proc, but is by no means the "fastest" one out there. Where are you getting this from?

 

[VTboy]

Here are some WU times for the Pentium-M 1.6GHZ. WU time for 1.6GHZ

Here is another times list for a 1.6GHZ Pentium-M More Times for 1.6GHZ

Here are times for other chips and Barton. Other CPU times

Clearly the 1.6GHz Pentium M is about as fast as the 2.4GHZ Barton.

Also the Pentium-M has a much higher IPC than Bartson does.

 

[VTboy]

Alex its numbers are almost identical when it comes to running Seti@home. I am only talking about seti@home.

 

[DAPUNISHER]

Highly impressive performance to say the least :beer: Too bad they can't market it's strong performance to clockspeed ratio without partially validating AMD's PR system.

 

[dnuggett]

These #'s are what we should expect and should not come as a shock. A 1.6 M chip is equivalent to a 2.56 P4 chip. The # are validating what is already known.

 

[alexruiz]

Wow, I must admit it is impressive. Can anyone confirm the IPC for both proccesors?

 

[sciencewhiz]

Originally posted by: dnuggett
These #'s are what we should expect and should not come as a shock. A 1.6 M chip is equivalent to a 2.56 P4 chip. The # are validating what is already known.

If you use only this benchmark, then a 1.6 M is equivilant to a 3.4 - 3.8 ghz P4

 

[dnuggett]

Originally posted by: sciencewhiz

Originally posted by: dnuggett
These #'s are what we should expect and should not come as a shock. A 1.6 M chip is equivalent to a 2.56 P4 chip. The # are validating what is already known.

If you use only this benchmark, then a 1.6 M is equivilant to a 3.4 - 3.8 ghz P4

How so?

 

[sciencewhiz]

Originally posted by: VTboy
Here are some WU times for the Pentium-M 1.6GHZ. WU time for 1.6GHZ

Here is another times list for a 1.6GHZ Pentium-M More Times for 1.6GHZ

Here are times for other chips and Barton. Other CPU times

Using the first two links, you see that a 1.6 M does a seti workunit in about 2 hours. Looking at the third link, a 3481 p4 does one in 1:59 and a 3840 p4 does one in 2:00

 

[dnuggett]

Using the first two links, you see that a 1.6 M does a seti workunit in about 2 hours. Looking at the third link, a 3481 p4 does one in 1:59 and a 3840 p4 does one in 2:00

Right, but I also see a 2.4 Tbred and Barton in @ 2.06. The M chip was @ 2.08 (approx avg)

I am guessing that there were other resources committed when those P4's were crunching.

 

[dexvx]

Originally posted by: dnuggett

Using the first two links, you see that a 1.6 M does a seti workunit in about 2 hours. Looking at the third link, a 3481 p4 does one in 1:59 and a 3840 p4 does one in 2:00

Right, but I also see a 2.4 Tbred and Barton in @ 2.06. The M chip was @ 2.08 (approx avg)

I am guessing that there were other resources committed when those P4's were crunching.

Can't you just accept that the Pentium-M is a SETI monster?

Its obvious from looking at the initial benchmarks that Pentium-M is very suitable for SETI.

 

[dnuggett]

Yes I sure can. I believe it is very suitable. But I ? the degree of suitablity over the other P4' s in this discussion. Care to explain why the Tbred and Barton were so close?

I love the fact that the M does such a great job, I think it is a great chip and it also furthers my belief of having just a laptop for a Windows machine and using a Mac for a desktop.

 

[dexvx]

Originally posted by: dnuggett
Yes I sure can. I believe it is very suitable. But I ? the degree of suitablity over the other P4' s in this discussion. Care to explain why the Tbred and Barton were so close?

More cache in a CPU doesnt mean it scales linearly to more cache in another different type of CPU.

P4 scaled much better than the AXP going from 256 -> 512.