Does anyone know how many IPC a Pentium M has?

[MDme]

"If the a64 and the dothan were matched in clock speed the dothan would win, in games of course. I have no proof handy, but I have seen this on many occasions on different webistes.
So at the very least we know that pentium M IPCs > Athlon 64 IPCs
Note: I don't know how many either of the processors has exactly. "

you can't conclude that. because each architecture may be better suited for one task better than the other. example in video encoding A64 is better than P-M clock for clock, do you therefore say that A64 IPC is > than P-M???

The P-M has a very good architecture (in games) albeit not as good over all as the A64.

 

[Duvie]

Originally posted by: Sentential

Originally posted by: carlosd

Originally posted by: Sentential
I *believe* it is 12. My friend has a Dothan setup and his at mid 2.6 can cripple a FX55 in the 3.4ghz range.

Picture proof is here:

http://home.comcast.net/~gautamb/cpucomparo.jpg

OK , meaningless synthetic benchmarks. In real APPS dothan won't cripple an FX55 at 3400 GHZ , specially if we talk about FPU intensive apps, the dothan performance will be laughable.

Name one "real app"... btw is this "real" enough for you?

http://home.comcast.net/~gautamb/2745pi.JPG

Oh the idiot statement again...OK Beavis...tell him to run 32mb...The fact is or the scam when used by Intel fanboys is that the 1mb test is not an accurate reflection of Dothan prowess cause the entire test can be loaded up into the cache and L2 low latency will shine...Now run the same test at 4mb, 8mb and so on....When you run 32mb compare it to my score of 27min 28sec.....You beat my 30sec at 2.7ghz but lets see if it holds true once you raise the size of the test...

 

[MDme]

yup, duvie is right. the only reason the P-M wins superpi with the 1mb test is the entire test runs in the low latency L2 cache. but once you set superpi to let's say, 32mb, the P-M won't beat the A64, again, this proves the fact that different architectures shine differently in different tests. sometimes even in the same program with different settings!!

 

[Duvie]

Intel fanboys of late have relegated themselves to spinning the data to make thinks look rosier then they are....I am used to it from sentential...he has been doing that since the 6xx series of prescotts came out....

 

[IntelUser2000]

Ooookay, I see some people need corrections here.

Hi folks, those results are mine. I believe that the Dothan has a pipeline stage length of either 14 or 15, compared to the Athlon64's 12. It's raw IPC, as a consequence, is actually lower than the Athlon64. It's performance in raw integer calculation, as tested by the Sandra Arithmetic test and CPU Renderbench suffer as a result. In the Sandra test, it slots between an AthlonXP and Athlon64 at the same clock speeds.

Oookay people, some corrections must be taken here I see.

No no no, you are seriously wrong. Even though the integer pipeline numbers are 14-15(which is what we use for pipeline #'s), we have to look at the branch misprediction stages. Even though Itanium 2 based CPUs have 8 stage pipeline, the branch misprediction stages are 7. First Itanium has 10 stage pipelines, but the branch misprediction stages are 9.

We can't compare that way. THE ONLY WAY TO LOOK AT PERFORMANCE IS BY LOOKING AT VARIOUS BENCHMARKS, NOT COMPARING IPC, BECAUSE IPC FOR EVERY X86 CPU FOR PC MARKET IS SAME!!! Its how efficient the CPU takes advantage of the IPC that matters, plus, two CPUs with same IPC but with different FSB speeds will have different real life performance.

Its like how you can go at 150km/h in highway say, but then because of so many toll gates, it slows you down, IPC is like the km/h in this case.

WHAT MATTERS IS THE REAL LIFE BENCHMARKS.




What's funny is that people get the Athlon IPC=9, Pentium 4=6 from AMD's site, because their site states it that way, but the REAL measure of IPC is different than "IPC" that AMD claims.

 

[deveraux]

Originally posted by: IntelUser2000
Ooookay, I see some people need corrections here.

Hi folks, those results are mine. I believe that the Dothan has a pipeline stage length of either 14 or 15, compared to the Athlon64's 12. It's raw IPC, as a consequence, is actually lower than the Athlon64. It's performance in raw integer calculation, as tested by the Sandra Arithmetic test and CPU Renderbench suffer as a result. In the Sandra test, it slots between an AthlonXP and Athlon64 at the same clock speeds.

Oookay people, some corrections must be taken here I see.

No no no, you are seriously wrong. Even though the integer pipeline numbers are 14-15(which is what we use for pipeline #'s), we have to look at the branch misprediction stages. Even though Itanium 2 based CPUs have 8 stage pipeline, the branch misprediction stages are 7. First Itanium has 10 stage pipelines, but the branch misprediction stages are 9.

We can't compare that way. THE ONLY WAY TO LOOK AT PERFORMANCE IS BY LOOKING AT VARIOUS BENCHMARKS, NOT COMPARING IPC, BECAUSE IPC FOR EVERY X86 CPU FOR PC MARKET IS SAME!!! Its how efficient the CPU takes advantage of the IPC that matters, plus, two CPUs with same IPC but with different FSB speeds will have different real life performance.

Its like how you can go at 150km/h in highway say, but then because of so many toll gates, it slows you down, IPC is like the km/h in this case.

WHAT MATTERS IS THE REAL LIFE BENCHMARKS.




What's funny is that people get the Athlon IPC=9, Pentium 4=6 from AMD's site, because their site states it that way, but the REAL measure of IPC is different than "IPC" that AMD claims.

Dude, calm down. He was just giving us his opinion on what he believes it does. You don't really have to jump on his argument like that.

 

[miketheidiot]

Originally posted by: IntelUser2000


What's funny is that people get the Athlon IPC=9, Pentium 4=6 from AMD's site, because their site states it that way, but the REAL measure of IPC is different than "IPC" that AMD claims.

that is actually in reference to execution untis i thought?

 

[sangyup81]

so then according to AMD's IPC, how many does the Pentium M have?

 

[Fox5]

Originally posted by: miketheidiot
i believe p-m has the same number of decode units as a64. I think it may be lacking an FPU unit or two (k7/k8 have 3 fpu and 3 integer if i recall) While a64 benefits from having lots of low latency memory bandwidth and dedicated HT link to the chipset, P-m benefits from a better branch predictor and a large very low latency l2 cache. I expect the next generation for each will improve on each processors weaknesses.

EDIT: i can't tpye

Except the next gen p-m is coming out rather soon(what will it be called on the desktop?), whereas the K9 is just quad core K8s isn't it? And then K10 will be a new design.

If the a64 and the dothan were matched in clock speed the dothan would win, in games of course. I have no proof handy, but I have seen this on many occasions on different webistes.

Anandtech has an article of PM desktop performance against athlon 64 s939 performance, the a64s win handidly in games at the same mhz.
And I since the Dothan is P3 based, I thought it had the same theoretical IPC as the P3.(heck, the early Celeron Ms had the same cache size as the later mobile P3s, and performed almost exactly the same mhz per mhz)

 

[BitByBit]

Dothan was designed from the ground up to be a low-power processor, and it acheives this by completing instruction execution in as few cycles as possible, resulting in very high efficiency. It was also designed to limit memory access to as little as possible (also contributing to the aforementioned efficiency), since memory access is costly not only in terms of performance, but also in terms of power usage.
This is why Dothan has such a large L2 - to avoid memory access, and this is why it does not respond as well to increases in memory speed as the P4.
The Athlon 64, while having much faster memory access, has a smaller, slower L2 cache than Dothan, which gives the latter an edge with smaller datasets (gaming).
This, combined with its superior branch prediction ability, ensures that its pipeline suffers from bubbles less than the A64.
Dothan's execution engine is narrower than the Athlon's, and its decode bandwdith is significantly less, but thanks to its efficiency, this lower theoretical performance peak is more than offset.
In answer to the original question:
The Athlon's maximum IPC is 3, and Dothan's is 2.
Dothan is, however, more efficient clock for clock, and it acheives an average IPC far closer to its maximum than the Athlon 64.

 

[Dothan]

Originally posted by: Duvie

Oh the idiot statement again...OK Beavis...tell him to run 32mb...The fact is or the scam when used by Intel fanboys is that the 1mb test is not an accurate reflection of Dothan prowess cause the entire test can be loaded up into the cache and L2 low latency will shine...Now run the same test at 4mb, 8mb and so on....When you run 32mb compare it to my score of 27min 28sec.....You beat my 30sec at 2.7ghz but lets see if it holds true once you raise the size of the test...

Oh please, give up your scam Duvie !!!

You must be the biggest AMD zealot in history, maybe you are on the payroll ???

Don't get mad that your PoS AMD machine gets owned by a mobile chip !!!

We know Dothan is the fastest clock - for - clock processor right now and it is improving as Intel is beefing up the FPU and moving to 65nm process !!!

Once dual core Dothan become available all you AMD fanatic will weep !!!

 

[Markfw]

Originally posted by: Dothan

Originally posted by: Duvie

Oh the idiot statement again...OK Beavis...tell him to run 32mb...The fact is or the scam when used by Intel fanboys is that the 1mb test is not an accurate reflection of Dothan prowess cause the entire test can be loaded up into the cache and L2 low latency will shine...Now run the same test at 4mb, 8mb and so on....When you run 32mb compare it to my score of 27min 28sec.....You beat my 30sec at 2.7ghz but lets see if it holds true once you raise the size of the test...

Oh please, give up your scam Duvie !!!

You must be the biggest AMD zealot in history, maybe you are on the payroll ???

Don't get mad that your PoS AMD machine gets owned by a mobile chip !!!

We know Dothan is the fastest clock - for - clock processor right now and it is improving as Intel is beefing up the FPU and moving to 65nm process !!!

Once dual core Dothan become available all you AMD fanatic will weep !!!

Why don't you fire up your Dothan CPU, run the test like he suggests ? Because it won;t win ??

 

[Dothan]

Originally posted by: Markfw900
Why don't you fire up your Dothan CPU, run the test like he suggests ? Because it won;t win ??

I already have !!!

I have beaten Duvie's AMD machine with my lowly Dothan mobile chip !!!

 

[Markfw]

Originally posted by: Dothan

Originally posted by: Markfw900
Why don't you fire up your Dothan CPU, run the test like he suggests ? Because it won;t win ??

I already have !!!

I have beaten Duvie's AMD machine with my lowly Dothan mobile chip !!!

The 32m test ? where ?

 

[Gautam B]

Originally posted by: IntelUser2000

No no no, you are seriously wrong. Even though the integer pipeline numbers are 14-15(which is what we use for pipeline #'s), we have to look at the branch misprediction stages. Even though Itanium 2 based CPUs have 8 stage pipeline, the branch misprediction stages are 7. First Itanium has 10 stage pipelines, but the branch misprediction stages are 9.

We can't compare that way. THE ONLY WAY TO LOOK AT PERFORMANCE IS BY LOOKING AT VARIOUS BENCHMARKS, NOT COMPARING IPC, BECAUSE IPC FOR EVERY X86 CPU FOR PC MARKET IS SAME!!! Its how efficient the CPU takes advantage of the IPC that matters, plus, two CPUs with same IPC but with different FSB speeds will have different real life performance.

Its like how you can go at 150km/h in highway say, but then because of so many toll gates, it slows you down, IPC is like the km/h in this case.

WHAT MATTERS IS THE REAL LIFE BENCHMARKS.

What's funny is that people get the Athlon IPC=9, Pentium 4=6 from AMD's site, because their site states it that way, but the REAL measure of IPC is different than "IPC" that AMD claims.

Wasn't aware of this but thanks for pointing it out...no need to get so jumpy about it. Here's my 32m test; still compares to an A64 San Diego at about 2.9GHz.

 

[Sentential]

Nice. Was that with a Gig or 512?

 

[Duvie]

Originally posted by: Dothan

Originally posted by: Markfw900
Why don't you fire up your Dothan CPU, run the test like he suggests ? Because it won;t win ??

I already have !!!

I have beaten Duvie's AMD machine with my lowly Dothan mobile chip !!!

Post pic then...I can believe you will be along side me but to say liek Sentential that it will be alongside a 3.4ghz A64 is pure fiction...I am only at 2.66ghz so if you can run 2.6ghz Dothan you may be right there with me if you are on an Asus mobo and running some high speed ram.....A 3.4ghz according to the test thread by Zakee kills me so you need to be in the 23 minute range for sure...Again show link to picture...LIAR!!!

 

[Gautam B]

Originally posted by: Duvie

Originally posted by: Dothan

Originally posted by: Markfw900
Why don't you fire up your Dothan CPU, run the test like he suggests ? Because it won;t win ??

I already have !!!

I have beaten Duvie's AMD machine with my lowly Dothan mobile chip !!!

Post pic then...I can believe you will be along side me but to say liek Sentential that it will be alongside a 3.4ghz A64 is pure fiction...I am only at 2.66ghz so if you can run 2.6ghz Dothan you may be right there with me if you are on an Asus mobo and running some high speed ram.....A 3.4ghz according to the test thread by Zakee kills me so you need to be in the 23 minute range for sure...Again show link to picture...LIAR!!!

I don't know where the 3.4 gigs came from exactly...Aquamark3 is just weird like that.

But I did show that my 2.67 gig Dothan is equivalent to a 2.9 gig San Diego.

 

[Duvie]

In a few things I can definitley sse that...In an more overall computing comparison throughout differing areas I think it is closer clok for clock like most have said....

Have you run the 32m test of superpi??? If so I would appreciate if you would post a pic....

 

[Gautam B]

I did two posts up link is here again, 25m 29 sec. My A64 3000+ 754 Clawhammer did it in 27' 29" at 2.75GHz.

Both with 512 megs of memory. (nearly a 30 second penalty)

btw, just about every number in MBM5 is way out of whack.

 

[Duvie]

AMD SYSTEMS
1. dnavarro- 24m 06s FX-55 @ 3.14GHz
2. Whizzz- 25m 05s FX-55 @ 3.15GHz (Patched)
3. Zebo- 27m 0s A64 3000+ @ 2.655GHz (Patched)
4. billwinkle- 27m 14s A64 3000+ @ 2.66GHz (Patched)
5. Duvie- 27m 28s A64 3000+ @ 2.66GHz (Patched)

So it appears in that test you would be in the 3.05ghz range....I can buy that....

 

[Gautam B]

If I patched, I'd be in the 23 or 24 sec range....I don't do patched scores. SuperPI mod 1.4, uncheatable version. That's all I run, and IMHO, it's all anyone should run, instead of all of these versions that tamper with the original program. It's point is to be free of software optimizations, and gauge the power of the architecture at a low level.

sorry about the rant.

Here is my A64 score using this unpatched version.

 

[sangyup81]

Dothan = Troll

 

[bunnyfubbles]

Originally posted by: Dothan
Oh please, give up your scam Duvie !!!

You must be the biggest AMD zealot in history, maybe you are on the payroll ???

Right, considering you've been around here for what, 6 days, you conveniently missed all the threads from Duvie about his Northwoods and the praise he gave Intel then... !!!

Don't get mad that your PoS AMD machine gets owned by a mobile chip !!!

How could he get mad? He paid less for his AMD system than a similar P-M desktop setup. Even if it is faster in a synthetic benchmark, it still doesn't make it a better buy. Ofcourse there's also poor FPU performance which completely kills any killer deal the Dothan might provide, which brings us too... !!!

We know Dothan is the fastest clock - for - clock processor right now and it is improving as Intel is beefing up the FPU and moving to 65nm process !!!

Right, and in order to fix the FPU performance Intel might have to lengthen the pipeline which would most likely hurt its clock for clock advantage. Let's hope they can do it though it would be a great processor to have with no glaring weaknesses. !!!

Once dual core Dothan become available all you AMD fanatic will weep !!!

Yeah, the fanboys might weep, but guys like Duvie and I will most likely have these dualcore Dothans in our systems if they are offered at reasonable prices. !!!



edit - forgot all my !!!

 

[MDme]

i can't believe how blind some people could be. The P-M is a great chip, BUT it is not as good in other programs as the A64 or even the P4. Let us all remember that they are all different architectures and as such, do some tasks better than other architectures.

overclocked P-M benches on a desktop scenario
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2382&p=10
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2382&p=11
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2382&p=12

Encoding for example is a task where the P-M can't hold a candle to the A64 and the P4.
Games pretty much neck and neck with the A64
in other tests the P-M varies from great to poor.

so really, you can't say that the P-M is better than A64. In fact, over all the A64 is a better all around chip. The P4 is great in media encoding, but poor in games. The P-M is the other way around. The A64 pwns the P4 in games, and while not as fast as the P4 in encoding it still is much better than the P-M in encoding.

by the way....do you think dothan and porskter is one and the same person??
or maybe father and son???