P M 780 review

[Intelia]

http://www.xbitlabs.com/articles/cpu/display/pentiumm-780.html

After this review It looks to me like yonah is going to be one hell of a cpu and merom well wait and see .all i can say is holy moly. I think these are going to be faster than we dared dream.

 

[Duvie]

I think the 780 at stock looked like a POS on several of those benches...so you spend a lot for the chip and better OC it cause you get marginal performance in the stuff that matters to me.....

http://www.xbitlabs.com/articles/cpu/display/pentiumm-780_16.html
http://www.xbitlabs.com/articles/cpu/display/pentiumm-780_17.html
http://www.xbitlabs.com/articles/cpu/display/pentiumm-780_18.html
http://www.xbitlabs.com/articles/cpu/display/pentiumm-780_19.html


All I have to say is ouch and thank god for its cache....

I think at $670 for a 770 I would get a 4400+ an dlcock it to 2.6-2.7ghz as most are seeming to get right now and that 20% OC wont help it out....

So basically I am not as impressed as you....

based on what I say clock for clock to a AMD in apps that would even take advanatge of a dual core it seems at current it would not be enough to eclipse the X2....

 

[n7]

The review shows that the Dothan certainly is not ready for the desktop at all.

It does well in games...but that's all.
In everything else, it gets trampled on.

That being said, i certainly am looking forward to seeing what they can do with it in the future, but it looks like a lot of work is still needed.

 

[Duvie]

I agree yonah is bringing more to the table then what the current P4 M is now.....This current incarnation would franky just not cut it....

 

[Intelia]

Originally posted by: Duvie
I agree yonah is bringing more to the table then what the current P4 M is now.....This current incarnation would franky just not cut it....

I agree 100% these cpu's can't compete were FFU is important and hyper threading.But right now in gaming the 2.26 has no equal. IF Intel did improve it by 30% as they have stated>Yonah well definitly be very fast indeed and I would say will definitly be a bad ass cpu and as far as O/C goes the 65 nano yonah should easily do 3.0 GHz right off the bat.
6 more months and the tables will be reversed . Quit a bit out on the web right now about intel bringing yonah to the desk top right away. In any case someone will.

http://www.xbitlabs.com/articles/cpu/display/pentiumm-780_15.html

http://www.xbitlabs.com/articles/cpu/display/pentiumm-780_17.html

http://www.xbitlabs.com/articles/cpu/display/pentiumm-780_18.html

http://www.xbitlabs.com/articles/cpu/display/pentiumm-780_9.html

http://www.xbitlabs.com/articles/cpu/display/pentiumm-780_10.html

http://www.xbitlabs.com/articles/cpu/display/pentiumm-780_12.html

Any body find any G70 reviews yet??????????????????

 

[Acanthus]

so... you are predicting a 25% increase from dothan to yonah in clockspeed, based on a completely unproven die shrink, on a chip that will be highly modified from the current dothan design.

Considering that 90nm from 130nm bought Intel about 16% clockspeed over the course of a year now, and thats with extending the pipeline and increasing the cache latency dramatically, i think 25% clockspeed increases may be setting bar a little high.

 

[Duvie]

IF Intel did improve it by 30% as they have stated

Yeah and I would believe those fvckers??? that is marketing spin which is 40% hype, 50% flat out lie and 10% truth...

They told us HT would give us 25% before HT arrived and the best I ever found was 22% and that was in like 2-3 apps out of the dozens and dozens I tested...avreage was more like 10%....

Sorry words are words, and from INtel they usually are quite overhyped.

I would wait until actual silicon arrives in the labs of real testers and we see for ourselves.

 

[Intelia]

Originally posted by: Acanthus
so... you are predicting a 25% increase from dothan to yonah in clockspeed, based on a completely unproven die shrink, on a chip that will be highly modified from the current dothan design.

Considering that 90nm from 130nm bought Intel about 16% clockspeed over the course of a year now, and thats with extending the pipeline and increasing the cache latency dramatically, i think 25% clockspeed increases may be setting bar a little high.

No actually Intel has stated yonah well be 30% clock for clock than dothan .

Not 30% increase in clock speed . 30% faster than current dothans at same clock

Ya Brian I know Intel can hype . I am thinking this time with SSE 3 Micro opps.fusion
and improved FFU 30% is a good possiability

 

[Duvie]

Originally posted by: Intelia

Originally posted by: Acanthus
so... you are predicting a 25% increase from dothan to yonah in clockspeed, based on a completely unproven die shrink, on a chip that will be highly modified from the current dothan design.

Considering that 90nm from 130nm bought Intel about 16% clockspeed over the course of a year now, and thats with extending the pipeline and increasing the cache latency dramatically, i think 25% clockspeed increases may be setting bar a little high.

No actually Intel has stated yonah well be 30% clock for clock than dothan .

Not 30% increase in clock speed . 30% faster than current dothans at same clock

Yeah right...They lie flat out...Again it is best case scenario...it is like their blatant spin of the truth with their TDP numbers....

I say I book mark this and when they deliver this and it doesn't deliver 30% I will laugh my arse off....

I cant believe you believe what they say....

 

[Intelia]

Originally posted by: Intelia

Originally posted by: Acanthus
so... you are predicting a 25% increase from dothan to yonah in clockspeed, based on a completely unproven die shrink, on a chip that will be highly modified from the current dothan design.

Considering that 90nm from 130nm bought Intel about 16% clockspeed over the course of a year now, and thats with extending the pipeline and increasing the cache latency dramatically, i think 25% clockspeed increases may be setting bar a little high.

No actually Intel has stated yonah well be 30% clock for clock than dothan .

Not 30% increase in clock speed . 30% faster than current dothans at same clock

Ya Brian I know Intel can hype . I am thinking this time with SSE 3 Micro opps.fusion
and improved FFU 30% is a good possiability
[/q

Hay 10% would be a lot and all they really need from what I can see ! With 2cores a lot of test scores would fly high. and in this review the only 2.2 AMD was dual core all others were higher clocks.
I can't help but wonder how far Conroe is going to clock to . I was think 3GHz was going to be tops but its looking like thats going to be easy.

 

[fatty4ksu]

A chip designed for notebooks...killing the AMD processors.

Gotta love it.:thumbsup:

 

[n7]

Originally posted by: fatty4ksu
A chip designed for notebooks...killing the AMD processors.

Gotta love it.:thumbsup:

A mentally challenged poster who cannot comprehend reality.

Gotta love it :thumbsup:


BTW, in case you get confused about what i meant...the Pentium Ms are awesome notebook CPUs. Best out there.
Your posts are another story

 

[Lonyo]

Originally posted by: n7

Originally posted by: fatty4ksu
A chip designed for notebooks...killing the AMD processors.

Gotta love it.:thumbsup:

A mentally challenged poster who cannot comprehend reality.

Gotta love it :thumbsup:


BTW, in case you get confused about what i meant...the Pentium Ms are awesome notebook CPUs. Best out there.
Your posts are another story

No, he's fairly right.
CLOCK FOR CLOCK the Dothan is pretty f-ing close to the AMD64's in the main, and if Yonah improves the weak areas (not 30% across the board, but maybe 30% in its very weak areas), it will be pretty much on par/above AMD64 clock for clock.
Remember the 3800+ is something like 2.4GHz compared to 2.26GHz for the Dothan, and the Dothan is fairly close to it in 90% of the tests.
The only thing is that its clock speeds are lower.
When you overclock though, it then becomes quite a lot faster (25% oncrease in clock speed), which makes it even more competetive. But then again, the AMD procs can overclock as well.

But as a potential desktop platform for the future with improved areas and higher clocks for desktop where heat and power are smaller issues, it could whip AMD's ass.
Although AMD by the time we get desktop Banias/Dothan/Yonah/Conroe etc will have probably moved to something different (be it quad core or something else).

 

[n7]

I posted what i did because i am sick of reading his crap in so many threads.
Do a search if you don't know what i mean.

I do agree though, the Yonahs look like they will have a lot of potential...i'm just not a fan of speculation, since alot can happen before they ever come out

 

[Viditor]

Originally posted by: Intelia

Originally posted by: Intelia

Originally posted by: Acanthus
so... you are predicting a 25% increase from dothan to yonah in clockspeed, based on a completely unproven die shrink, on a chip that will be highly modified from the current dothan design.

Considering that 90nm from 130nm bought Intel about 16% clockspeed over the course of a year now, and thats with extending the pipeline and increasing the cache latency dramatically, i think 25% clockspeed increases may be setting bar a little high.

No actually Intel has stated yonah well be 30% clock for clock than dothan .

Not 30% increase in clock speed . 30% faster than current dothans at same clock

Ya Brain I know Intel can hype . I am thinking this time with SSE 3 Micro opps.fusion
and improved FFU 30% is a good possiability
[/q

Hay 10% would be a lot and all they really need from what I can see ! With 2cores a lot of test scores would fly high. and in this review the only 2.2 AMD was dual core all others were higher clocks.
I can't help but wonder how far Conroe is going to clock to . I was think 3GHz was going to be tops but its looking like thats going to be easy.

I'm sure that they WILL be 30% faster, but only on 1 or 2 apps...however they will certainly be better at multitasking when not on battery. (remember that the 30% is predicated on plugged in speed only, when on battery they will be locked to only one core)

Some other things to note...
1. The AMD roadmap only goes to the end of this year (though there have been comments and announcements that aren't listed on the roadmap).
2. It is quite likely that we will see dual core Semprons at the end of the year (if you produce an X2 with a cache defect, you can still sell it as a Sempron)
3. It is also quite possible that we will see dual core Turions early next year.
4. Comparing a dual core Turion and Merom would be pretty damned close in performance with one exception. The platform for Turion contains HTransport and the Merom does not. Until Intel release their CSI platform in 2007, AMD should continue to have the performance edge...after that though, who knows!

 

[Intelia]

Originally posted by: Viditor

Originally posted by: Intelia

Originally posted by: Intelia

Originally posted by: Acanthus
so... you are predicting a 25% increase from dothan to yonah in clockspeed, based on a completely unproven die shrink, on a chip that will be highly modified from the current dothan design.

Considering that 90nm from 130nm bought Intel about 16% clockspeed over the course of a year now, and thats with extending the pipeline and increasing the cache latency dramatically, i think 25% clockspeed increases may be setting bar a little high.

No actually Intel has stated yonah well be 30% clock for clock than dothan .

Not 30% increase in clock speed . 30% faster than current dothans at same clock

Ya Brain I know Intel can hype . I am thinking this time with SSE 3 Micro opps.fusion
and improved FFU 30% is a good possiability
[/q

Hay 10% would be a lot and all they really need from what I can see ! With 2cores a lot of test scores would fly high. and in this review the only 2.2 AMD was dual core all others were higher clocks.
I can't help but wonder how far Conroe is going to clock to . I was think 3GHz was going to be tops but its looking like thats going to be easy.

I'm sure that they WILL be 30% faster, but only on 1 or 2 apps...however they will certainly be better at multitasking when not on battery. (remember that the 30% is predicated on plugged in speed only, when on battery they will be locked to only one core)

Some other things to note...
1. The AMD roadmap only goes to the end of this year (though there have been comments and announcements that aren't listed on the roadmap).
2. It is quite likely that we will see dual core Semprons at the end of the year (if you produce an X2 with a cache defect, you can still sell it as a Sempron)
3. It is also quite possible that we will see dual core Turions early next year.
4. Comparing a dual core Turion and Merom would be pretty damned close in performance with one exception. The platform for Turion contains HTransport and the Merom does not. Until Intel release their CSI platform in 2007, AMD should continue to have the performance edge...after that though, who knows!

Another thing to keep in mind is AMD is already ondie memory and the dothan memory controller is at the moment on the desk top is handycapped and still its hanging in there. CSI will be welcomed but the dothan don't seem to mind the onboard controller much do it . Lets say ondie CSI will bring 5% improvement (low est.)

Amd is coming out with a new socket and memory controller DDR2 or 3 If I were AMD I would go right to XDR and than know upgrades needed for a long while .

As I understand it the Intel CSI is going to built around a token ring type arrangement.I don't know much about this tech. It seems to me how ever that a point to point would be better . Any ideas on this?

Ya I am really curious about the Quad cores it seems meroms 4 core are going to consume 115 watts. and underload AMD best consums way more power than a dothan . This really is shaping up into some kind of battle. Intel Intel vs. AMD IBM I don't know who well come out on top but its going to be fun ride.

 

[stevty2889]

Well I am certainly happy with my pentium-m running at 2.4. Stuffed in a cramped antec aria, with an overclocked 6600GT..my load temps while gaming in an air conditioned room...38c. The loudest part of the system..the dvd drive spinning up..makes for a nice portable gaming system at least, or if you want dead silence. A-64 is still the best over all, since the dothan can't keep up in much else.

 

[Pabster]

Jeez, talk about early speculation

No point in comparing Dothan right now to A64 when you can clearly see that Dothan is severely hampered by single-channel DDR and a relatively low FSB.

When Dothan at 2.26 (with the limitations above) is able to compete in ANYTHING with a modern A64 running at 2.4, I think that is enough said.

 

[Viditor]

Originally posted by: Intelia
Another thing to keep in mind is AMD is already ondie memory and the dothan memory controller is at the moment on the desk top is handycapped and still its hanging in there. CSI will be welcomed but the dothan don't seem to mind the onboard controller much do it . Lets say ondie CSI will bring 5% improvement (low est.)

Amd is coming out with a new socket and memory controller DDR2 or 3 If I were AMD I would go right to XDR and than know upgrades needed for a long while .

As I understand it the Intel CSI is going to built around a token ring type arrangement.I don't know much about this tech. It seems to me how ever that a point to point would be better . Any ideas on this?

Ya I am really curious about the Quad cores it seems meroms 4 core are going to consume 115 watts. and underload AMD best consums way more power than a dothan . This really is shaping up into some kind of battle. Intel Intel vs. AMD IBM I don't know who well come out on top but its going to be fun ride.

I really don't think that AMD will ever go XDR, and I doubt that Intel will either...
The reason being that as both companies move away from the FSB model and into a distributed model, memory bandwidth becomes less of an issue...
The quad cores might have an issue, but the problem would be better solved by having a second dual channel controller.
It appears that the next generation of AMD server chips will be going with FB-DIMMs rather than XDR, and I suspect Intel will do the same.

As to CSI vs P2P, both are excellent but the latency of CSI will be (very slightly) higher from what I can tell (which isn't much until the specs are released). If it follows along the same stats as a Token ring, then it will be...
Of course that will be MUCH better than the FSB model.
With the CSI on-die for Merom, I expect a good 20-30% improvement and an overall reduction in power draw...

One last thing...I have a feeling that the test Xbit did was flawed...you mentioned that the AMD draws more power under load, but that just isn't the case. Lost circuits tested the chips themselves, and they matched the temps almost exactly. Venice under load drew 30w max...the xbit article had it in the 60w+ range!
The Hammer and Newcastle chips do indeed run at that range, but the Venice and San Diego chips run at less than half the power of the older chips...

As a form of anecdotal evidence, I did a test using a program called "Toast" (available here). It's a great program for stressing out the CPU to absolute max...
I have a 3200+ Venice core system as my backup, so I decided to try it.
The max temp (after 1 hour) was 35C, standard temp is 24C, case temp is 28C. I have only 1 case fan and am using only the standard "boxed" HSF...

 

[Viditor]

Originally posted by: stevty2889
Well I am certainly happy with my pentium-m running at 2.4. Stuffed in a cramped antec aria, with an overclocked 6600GT..my load temps while gaming in an air conditioned room...38c. The loudest part of the system..the dvd drive spinning up..makes for a nice portable gaming system at least, or if you want dead silence. A-64 is still the best over all, since the dothan can't keep up in much else.

Stevty, if you get a chance, could you download that Toast program I linked above and run it on your system? I'd like to see how they compare...

 

[Acanthus]

Originally posted by: Intelia

Originally posted by: Acanthus
so... you are predicting a 25% increase from dothan to yonah in clockspeed, based on a completely unproven die shrink, on a chip that will be highly modified from the current dothan design.

Considering that 90nm from 130nm bought Intel about 16% clockspeed over the course of a year now, and thats with extending the pipeline and increasing the cache latency dramatically, i think 25% clockspeed increases may be setting bar a little high.

No actually Intel has stated yonah well be 30% clock for clock than dothan .

Not 30% increase in clock speed . 30% faster than current dothans at same clock

Ya Brian I know Intel can hype . I am thinking this time with SSE 3 Micro opps.fusion
and improved FFU 30% is a good possiability

You said 3.0ghz, thats where i got 25%, i wasnt talking about IPC performance at all.

 

[Vegitto]

This is pure troll bait.
If you want Intel, buy it, and don't bother us (me) with it.
If you want AMD, buy it, and don't bother us (me) with it.
Buy what you want, and be happy. What does it matter who manufactured it? All I want is the best chip/performance for my money, and I don't care if the label on the processor says "AMD", "Intel", "IBM" or for all I care "VIA". It just doesn't matter.

 

[imported_michaelpatrick33]

Originally posted by: Intelia

Originally posted by: Viditor

Originally posted by: Intelia

Originally posted by: Intelia

Originally posted by: Acanthus
so... you are predicting a 25% increase from dothan to yonah in clockspeed, based on a completely unproven die shrink, on a chip that will be highly modified from the current dothan design.

Considering that 90nm from 130nm bought Intel about 16% clockspeed over the course of a year now, and thats with extending the pipeline and increasing the cache latency dramatically, i think 25% clockspeed increases may be setting bar a little high.

No actually Intel has stated yonah well be 30% clock for clock than dothan .

Not 30% increase in clock speed . 30% faster than current dothans at same clock

Ya Brain I know Intel can hype . I am thinking this time with SSE 3 Micro opps.fusion
and improved FFU 30% is a good possiability
[/q

Hay 10% would be a lot and all they really need from what I can see ! With 2cores a lot of test scores would fly high. and in this review the only 2.2 AMD was dual core all others were higher clocks.
I can't help but wonder how far Conroe is going to clock to . I was think 3GHz was going to be tops but its looking like thats going to be easy.

I'm sure that they WILL be 30% faster, but only on 1 or 2 apps...however they will certainly be better at multitasking when not on battery. (remember that the 30% is predicated on plugged in speed only, when on battery they will be locked to only one core)

Some other things to note...
1. The AMD roadmap only goes to the end of this year (though there have been comments and announcements that aren't listed on the roadmap).
2. It is quite likely that we will see dual core Semprons at the end of the year (if you produce an X2 with a cache defect, you can still sell it as a Sempron)
3. It is also quite possible that we will see dual core Turions early next year.
4. Comparing a dual core Turion and Merom would be pretty damned close in performance with one exception. The platform for Turion contains HTransport and the Merom does not. Until Intel release their CSI platform in 2007, AMD should continue to have the performance edge...after that though, who knows!

Another thing to keep in mind is AMD is already ondie memory and the dothan memory controller is at the moment on the desk top is handycapped and still its hanging in there. CSI will be welcomed but the dothan don't seem to mind the onboard controller much do it . Lets say ondie CSI will bring 5% improvement (low est.)

Amd is coming out with a new socket and memory controller DDR2 or 3 If I were AMD I would go right to XDR and than know upgrades needed for a long while .

As I understand it the Intel CSI is going to built around a token ring type arrangement.I don't know much about this tech. It seems to me how ever that a point to point would be better . Any ideas on this?

Ya I am really curious about the Quad cores it seems meroms 4 core are going to consume 115 watts. and underload AMD best consums way more power than a dothan . This really is shaping up into some kind of battle. Intel Intel vs. AMD IBM I don't know who well come out on top but its going to be fun ride.

AMD's best mobile the Turion at 2.2 is a 35W processor. That is not using way more power than a Dothan. I would like to see a comparison between the two.
Turion at 35W at 2.2

The Dothan does not blow away AMD's best in power consumption per processor. 10W is a lot but not blowing away like it is implied. Other issues such as DDR1 and DDR2 etc. can begin to make a difference.

 

[Viditor]

michael - the 35w number is the TDP, and TDP is not the amount of power a chip actually draws. TDP is a thermal guide for OEMs....
Turion actually draws quite a bit less than 35w, in fact the Venice 2.4 GHz only draws 30w under load...

I should add that the P-M is very good as well. It is safe to assume that the P-M draws about the same power as the Turion. Remember that for the PM, you have to add the Northbridge power as well...

 

[imported_michaelpatrick33]

Originally posted by: Viditor
michael - the 35w number is the TDP, and TDP is not the amount of power a chip actually draws. TDP is a thermal guide for OEMs....
Turion actually draws quite a bit less than 35w, in fact the Venice 2.4 GHz only draws 30w under load...

I should add that the P-M is very good as well. It is safe to assume that the P-M draws about the same power as the Turion. Remember that for the PM, you have to add the Northbridge power as well...

I agree. I was just using that as a counterpoint to the Dothan blowing away the Turion in power usage. Once the Turion is more fully utilized it will be a great competitor to the Dothan. People talk about overclocking the Dothan but you can also overclock the Turion and get overall dominant performance, not just games.