P M 780 review

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

stevty2889

Diamond Member
Dec 13, 2003
7,036
8
81
Originally posted by: 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...

Will do, running on my prescott and northwood systems right now. Doesn't seem to raise the temps nearly as much as prime95 does. Prime95 had my northwood at 62c a few minutes ago, so far the highest temp toast has given me is 53c..

Hey Viditor, I can't seem to find much info on this toast program, I see a the top that it's switching between time @ xxxxmhz and 23 instructions in xxxx cpu clocks. If thats how long it takes the cpu to complete 23 instuctions, which is what it appears to be..my 3.06ghz northwood is completeing them in 1000 less cycles than my prescott @ 3.82ghz..still hasn't raised my temps much either...
 

Intelia

Banned
May 12, 2005
832
0
0
Originally posted by: 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...

Very very good you even new my source nice post. But could you say a little more about the 20 to 30% increase with CSI . That would be great but Holy moly we are really pushing the power of this Merom Cpu up there . With the other (?) improvements Intel has made over dothan with the Yonah . We are talking in the area of 50% faster at the same clock. Thats a lot! But I like how sweat it sounds.
 

Intelia

Banned
May 12, 2005
832
0
0
Originally posted by: 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.

Oh! please we are just having a discussion . Go to another thread if you don't like it . You seem to be the name calling troll that broke in . Even got me to reply to your Smart Arse remarks . Shame on you.

 

Intelia

Banned
May 12, 2005
832
0
0
Originally posted by: 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.

Nice post but remember we are just speculating about what Merom is going to be like . Also in referance to the power load I was only tring to show that a 4 core Merom is going to draw at least 115 watts at load. So that would put Turion at 155 watts at load . So whats my point ? Point is were going right back to high thermal loads. Intel, AMD, who ever.
It is really nice discussion though.
 

imported_michaelpatrick33

Platinum Member
Jun 19, 2004
2,364
0
0
Originally posted by: Intelia
Originally posted by: 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.

Nice post but remember we are just speculating about what Merom is going to be like . Also in referance to the power load I was only tring to show that a 4 core Merom is going to draw at least 115 watts at load. So that would put Turion at 155 watts at load . So whats my point ? Point is were going right back to high thermal loads. Intel, AMD, who ever.
It is really nice discussion though.


Ahh, but you are assuming the AMD 65nm heat and power dissipation will be the same 30W. Since the dualcore AMD X2's are already less than the one 130nm we can hope that the thermals of the 65nm will be even less. However, I hope both companies are successful because that would mean an AMD for me (my preference) and an Intel for you, your preference. We are both happy, LOL.