Why don't AMD and Intel seem to get that the process is dead?

[Penti]

Well welll.... looks like AMD is leaking more power than a Prescott. LoL It's official. Now what do you fanbois have to say for yourself?

Socket 939:


http://www.x86-secret.com/pics/cpu/s939/tablew.png
http://www.x86-secret.com/pics/cpu/s939/tablew2.png
http://www.x86-secret.com/pics/cpu/s939/hot.jpg


http://www.x86-secret.com

Hahaa.
First AMD's numbers are not EQUAL to INTELS TDP. Amd's numbers are maximum with even the memorycontrollers power included. Intel's are about 75-80% of max power consumption, they do not calculate the power as P = IV, P = IMax * V.
What AMD include is "1. Thermal Design Power (TDP) is measured under the conditions of TCASE Max, IDD Max, and VDD=VID_VDD,
and include all power dissipated on-die from VDD, VDDIO, VLDT, VTT, and VDDA."
Wheres TCASE Max = Max CPU Temp, VDD = Vcore, IDD Max = Maximum current, VLDT = Hypertransport voltage, VDDIO/VTT = DDR Voltage/DDR VTT Voltage.

Intel clearly states this in there datasheet
"1. Thermal Design Power (TDP) should be used for processor thermal solution design targets. The TDP is not the maximum power that the processor can dissipate.
2. This table shows the maximum TDP for a given frequency range. Individual processors may have a lower
TDP. Therefore, the maximum TC will vary depending on the TDP of the individual processor. Refer to
Figure 12 and Table 27 for the allowed combinations of power and TC."
If you check the maximum numbers they are about 20% higher.
Prescott is still hotter I'm a bit dissapointed at x86-secret over a such simple fault.


Penti

 

[Sunburn]

i have to say you dont know what your talking about

AMD gives the theoretical maximum power output for the entire line of chips. so while if every single transistor was on it would be 104watts.(an impossible situation basically)

intel gives a normal operating wattage on a per chip basis. so while prescotts normal operating range is 103watts it could very well be MORE then that.

in short. AMD's TDP != Intel's TDP.

LoL, if it makes you feel better to deny a fact published by AMD then so be it. Your theoretical argument is kinda funny though. Why don't you just look at the temp. 58c There is nothing theoreticall about that, of course as someone who denys facts I'm sure you could argue that it really isn't 58c. Anyway, hope I didn't hurt your feelings.

 

[Sunburn]

go go fanbois!!! YEAH!!! I admit I am one too

 

[Sunburn]

oh, btw enough with the technical jibber jabber.. Fact is 939's are going to be as HOT as a Prescott! LoL, Sorry, not my fault, just a fact of life. I mean heat has been rubbed in the face of prescott for a long time, now it's your turn .

 

[xtrmeocr]

your the one in denial.

btw that 58 degrees was at 1.8 volts with the STOCK heatsink.

 

[Penti]

The Facts says it's not gonna be as hot, you much know what the facts mean wich you don't know.

Please stop spamming about something you won't even lookup by yourself.

 

[Sunburn]

Suprise! According to very confession of amd, the CPU Socket 939 are likely to consume more than the... Prescott! With 104W of TDP against 103W, the CPUs Socket 939 undergo an increase of 20% of their consumption in comparison with the Socket 754.

 

[PunkRawk911]

I'm almost sorry I started this thread, it has just turned into an intel vs. AMD fight. How about we stop bashing each other, and stop talking about the thermal outputs of processors that aren't even out yet (socket 939). All I really meant this thread to be was a discussion about future technologies and whatnot. Not a stupid AMD vs. Intel thread.

 

[{Epsilon}Stalin]

sorry Punkrawk,
but if u think about it this is a thread discussing future technologies.....

as you said yourself, the guy who posted b4 u posted on socket 939 which is a future tech.
and as to the power consumptions, more power needs to be consumed for the CPU to make up for the leaked power caused by the thin walls of the 90nm process.

i think this thread is doing exactly what you intended it to do

 

[deathBOB]

Lol.... YOu just had to open this can of worms... SHAME ON YOU!!!

lol no... I think most of the people (me included ) CAN'T talk about future processor technology because:

1) We dont have enough hard information on what AMD or Intel will do.

2) We dont have the technical knowledge/background to discuss processor technology in-depth.

 

[{Epsilon}Stalin]

hahahhah
deathBOB you hit that one right on the head.....

yea i mean all we can really do is talk aboutthe "what if's"

but that brings us to the lusted after THz CPUS

 

[dippy_skoodlez]

you are the ones comparing a stock AMD heatsink to a stock P4 heatsink and saying they run hotter. If I use dry ice, and you use Ln2, you still gonna say mine runs hotter? You cannot take actual temperatures, unless its the SAME HEATSINK, SAME ROOM. PERIOD. If you do that, then comparing temps are fine. But, you really wanna compare your P4 to my 2.5ghz mobile at 38C? I think not. j00 be pwned when comparing that. But, oh... I have an SLK and a tornado... but you have stock heatsink. Hey.... its a temperature.. WE CAN COMPARE! right? Maybe you should look it up.

 

[dippy_skoodlez]

Suprise! According to very confession of amd, the CPU Socket 939 are likely to consume more than the... Prescott! With 104W of TDP against 103W, the CPUs Socket 939 undergo an increase of 20% of their consumption in comparison with the Socket 754.

And also be a lot faster If you wanna compare heat, atleast use an equivilant speed.

 

[Gautam]

Talk to TC(whose owned an FX-51 and a couple of A64's) and me, and we'll tell you these procs don't run cool at all. TC's owned a Prescott too, AFAIK, and tested each using a real thermistor. He said that the Prescott and FX lined up pretty evenly, in fact. Look at my last post here, and you'll see why this makes sense. The FX51 is just about identical to the 939's.

 

[deathBOB]

I believe it... Any time you add transistors (the A64s and FXs have huge transistor counts right? Same with Prescott) your gonna have more heat...

Dippy, you might have to compare similar steppings also... My 2.4c is a lot hotter than other 2.4cs with the stock sink because its particular stepping puts out 5-10 more watts of energy... I bet different chips all produce different heat, even from the same batch...

 

[TC]

Correct - I started out with an FX51 last fall. When I first got it I couldn't get my hands on a water block, so I had to use air cooling for a while. I had to order a taisol unit since nobody else was making anything for the A64's at the time. Now granted it's difficult to make perfect comparisons if you can't use the same heatsink, so all I'm talking about is a rough estimate. The taisol unit was a hefty hunk of aluminum - far larger than the retail unit that comes with prescotts. Having had a bad experience with mobo temp readings I usually use Joe's method of boring a small hole in the block and inserting a thermal diode. That gives me something to make a judgement with when looking at what the mobo says. I can tell you that the taisol unit with an odd size fan (I think it was 70mm) turning 3000+ rpm was getting up to 65-70C with the FX under full load and 1.7 volts. People laughed when I posted this saying that the heatsink must not have been making good contact with the cpu. Uh, well how does it get hot if it's not? Anyway it was my experience that the FX51 overclocked to 2400 on air was running about the same temp as my 2.8 prescott was at 3500 on 1.6 volts with the retail box cooler. Like I said this was no scientificly perfect comparison, but a pretty good ball park figure. The FX at 130nm is not a cool chip, and if AMD is having the same problems as everyone else going to 90nm, then I don't doubt those numbers that have been leaked. Of course another thing to consider is performance. The FX was hot, but it was largely spanking the prescott in everything I threw at it, except for seti. So you get more for your heat than the prescott. More recently I've run a laptop version of the A64, and it was noticebly cooler than the first FX I had. I had to use water to get anything out of the FX, but the mobile is tame enough to stick with air. Anyway that's my take on it for what it's worth.

 

[OC Detective]

Why are people badly translating (using babelfish) and using it as fact that the 939 pin will output 104W of heat? (I suspect X-86 are looking at the 90 nanometre figures not the 939 at 130nm). Also why is this thread continually being dragged back into an Intel v AMD situation and not about the topic itself. It may as well be locked!
My initial comments on this topic still stand.

 

[PunkRawk911]

Well, people are using the 90nm heat figures because they're trying to make a comparison between prescott at 90nm and socket 939 at 90nm, it's just a direct comparison of how each company is dealing with the smaller manufactuing process.

 

[Dukemurmur]

IDK all i have to say is that this is turning into an all out war....any way that said if it is "all is fair in love and war" ppl love AMD otherd love Intel and as stated it is a "war" we are battlin about the process shrink and the 90nm comparrison is in "general" on topic, man....so i would have to say as long as they are using "facts" to back it up let them keep at it if they want to.......

 

[k_r_i_p_l_e_r]

heh