Overclocking Sandbox: Tbred B DLT3C 1700+ and Beyond

damn thats some nice figuring how you know it uses 136 watt power? how you figure that?

I get 33 C idle temp to. its not that unbelievable. gotta luv my SK7..

EspElement said:

damn thats some nice figuring how you know it uses 136 watt power? how you figure that?

Click to expand...

Power of Tbred B 1700+ DLT3C at 2.5GHz 1.9V

The power of a CPU has two components: leakage power and active power.

leakage power = V^2/R
active power (for computation) = C V^2 f

From AMD spec,
1700+ 1.5V 1.467 GHz, power = 44.9 W (typical)
1700+ 1.5V 1.467 GHz, power = 49.4 W (max)

active power at 1.9V 2.5 GHz = 44.9 x (1.9 x 1.9 x 2.5) / (1.5 x 1.5 x 1.467) = 128 W (typical)
active power at 1.9V 2.5 GHz = 49.4 x (1.9 x 1.9 x 2.5) / (1.5 x 1.5 x 1.467) = 135 W (max)

For more details,
1. go to page 2 of this thread and look for this post:
Relationship of clock, die temperature and Vcore

2. Links for power and current estimate, PSU requiremetns and discussions:
Question: How much power is increased ...
how many watts does your cpu take up?

u know what the amd spec of a 2500 is?


PS: you are like the smartest person i have known, you know alot of ****ing math and science ****..... **** i wish i was smart like that, of course im only 16......let me guess your like 18 lol

element I was just about to say the same thing. are u in some computer class? how did u gain all this knowledge lol? seriously...

lololololol thats great, man it would be fun to know that much ****.... lol he must be like a DR. or something cause they have done that **** for years and know a **** load about math and science lol. anyways i have been searching for that out put for the 2500..... but is there a way i can configure it?

Comparing power of Barton w/ 1700+ DLT3C

Some useful links for AMD CPU's:

AMD Athlon? XP Processor Tech Docs

Data sheet for Model 10 (Barton)

Data sheet for Model 8 (Tbred)


If you go to page 33 of the data sheet for model 10, you shall see

2500+
rated frequency 1833 MHz (FSB 333)
rated voltage 1.65V
current 41.4 A max, 32.5 A typical
power 68.3 W max, 53.7W typical
max die temp 85 C

etc, etc, plus many many other things, ...

So if you can oc the Barton to 2.5 GHz at 2V,
active power (typical) = 53.7 x (2 x 2 x 2.5) / (1.65 x 1.65 x 1.833) = 108 W
active power (max) = 68.3 x (2 x 2 x 2.5) / (1.65 x 1.65 x 1.833) = 137 W


Observation:

- The active power (typical) of Barton 108 W at 2V is less than that of 1700+ 128 W at 1.9V, assume both oc'ed to 2.5 GHz.
- The max power per Vcore of Barton 137 W at 2V is less than the 135 W at 1.9V for the 1700+, assume both oc'ed to 2.5GHz.
- I used 2V for Barton and 1.9V for the 1700+.
- I think it is because the 1700+ is a high current CPU to begin with, even it is rated at lower Vcore 1.5V.

So in this sense, the Barton is much better, especially if Barton can be oc'ed to 2.4+ GHz and when the price drops lower. The Barton above 2.38 MHz will then be faster than a 1700+ at 2.5 GHz (assuming the 512KB L2 has a 5% average performance advantage), and cooler (less power and power density, bigger die area) !!!

that doesnt add up correctly to what it is tho..... i got 41c somethin out of that number.. at 2.5? well i get 50c at 2.3 at 2.0 vcore... very unstable
hmmm what about the 68.3 number? = 138w(max)
(.38)(138)=52.44 sounds more right

EspElement said:

that doesnt add up correctly to what it is tho..... i got 41c somethin out of that number.. at 2.5? well i get 50c at 2.3 at 2.0 vcore... very unstable
hmmm what about the 68.3 number? = 138w(max)
(.38)(138)=52.44 sounds more right

Click to expand...

I think you may need to change to a SK7 (economical) or SLK-800U, it cools the CPU on air much better than a volcano 9. It gives better chance to oc higher with more stability.

Your Barton now at 2.3 GHz is good oc, it is faster than a Tbred B 1700+ DLT3C at 2.4 GHz on system performance (assuming 512KB L2 has 5% av system performance advantage)..

PS: I updated the post (maybe you were browsing while I was editing).

I'll be ugrading to a Barton when the prices come down later this year. I've pretty much maxed out my L due partly to cooling issues. I don't want a screaming CPU fan in my living room while we watch TV or listen to music.

yea it runs at 2.3 with a **** high vcore 1.975..... but yea i tried running my cpu at 2.200 vcore last night with my cpu @ 600mhz
it seems to be faster then my friends 2.4 intel....lol
i think its cause his is a store bought one.... lol

yea the cpu does have around a 5% advantage....
@2.3 it runs fast then a 3.06(133bus/533) intel..... according to sandra

EspElement said:

yea it runs at 2.3 with a **** high vcore 1.975..... but yea i tried running my cpu at 2.200 vcore last night with my cpu @ 600mhz
it seems to be faster then my friends 2.4 intel....lol
i think its cause his is a store bought one.... lol

yea the cpu does have around a 5% advantage....
@2.3 it runs fast then a 3.06(133bus/533) intel..... according to sandra

Click to expand...

There is a magic number, called IPC (intruction per cycle) ratio, of 1.28 to meausre the relative raw CPU integer computation power (as DMIPS in Sandra) between an AMD Tbred (and Barton) CPU and a P4 CPU for a given clock frequency. That is for the same clock frequency, the DMIPS of Tbred is 1.28 times that of a P4. For floating point, that ratio is around 1.22 in reverse order, P4 to Tbred.

E.g. if your Barton is running at 2.3 GHz, it should equate to the raw CPU power a P4 running at 2.3 x 1.28 = 2.94 GHz. This is what you have described as above and measured by Sandra. The measurement does not account for the system performance which includes memory, cache, FSB, ...

yea true its at 200fsb so thats why its around teh 3.06 speeds

but i heard a 2.3 tbred would run at a 3.06 intel.....
700 mhz less, i read this at tomshardware.com or somethin

if i had a SLK900U i would have more clocking room i think
GOAL: 2.4 Barton @200*12

that is a really good overclock man...

i have an 1800xp dlt3c which does 200*11 at 1.75 not as good as yours doh...

i dont what week code though

kamz said:

i dont what week code though

Click to expand...

Take the heatsink off and check it out.

kamz said:

that is a really good overclock man...

i have an 1800xp dlt3c which does 200*11 at 1.75 not as good as yours doh...

i dont what week code though

Click to expand...

It should go higher if you can raise the Vcore, but you may need a Vcore mod or socket wiring trick to accomphish that to allow you to set higher Vcore in the bios. Look around for info regarding to that. Hope it can go 100-200 MHz more. What motherboard and HSF do you use?

It should be a Tbred B 1800+ DLT3C (rated 1.5V). Since a Tbred A would max out at 1.9-2.0 GHz, and you are getting 2.2 GHz.

So the interesting question is, do the Tbred B 1800+ DLT3C oc as well as the 1700+ DLT3C. From what I've seen in this forum, and they seem oc not as high as the 1700+ DLT3C (but that may be due to not many tested and/or reported). Even though both of them should overclock equally well, since they come from the same 0.13u process, rated as 1.5V, both have higher rated current, ... That is they should have very similar oc characteristics.

Tbred B
1700+ DLT3C 1467MHz 1.5V 32.9A max 29.9A typ 49.4W max 44.9W typ
1800+ DLT3C 1533MHz 1.5V 34.0A max 30.9A typ 51.0W max 46.3W typ

i have a dlt3c 1800xp on an nforce 2 nfs v2, 2*512 twinmoss with winbond, slk900 with 92mm ystec running at 2800 at 48 cfm...
Week code was around 03 sumting, dont want to take heat sink off clean AC£ and then re-aplly etc...

kamz said:

i have a dlt3c 1800xp on an nforce 2 nfs v2, 2*512 twinmoss with winbond, slk900 with 92mm ystec running at 2800 at 48 cfm...
Week code was around 03 sumting, dont want to take heat sink off clean AC£ and then re-aplly etc...

Click to expand...

Week code is not important now since you already have the CPU, just for you own record.

You have all the good hardware to do max overclocking on air. Only thing is you may need a fan with higher CFM, in case you need to do oc on air to 2.4-2.5 GHz, such as a TT SFII or even a Tornado (with VR for noise control).

Do you mean the mb a ABIT NF7-S V2, which should have higher adjustable Vcore than 1.75V, true?

Have you tried increase the Vcore to more than 1.75V? Hope you can get to 2.4 GHz

the code i got is 0307, i have tried higher volts even 1.8 but the system crashes or sumtimes wont boot after exiting bios... mayb some of my settings are wrong in the bios...

ps what is socket wiring trick???

Socket wiring tricks are a way to physically change the circuitry on the motherboard's CPU socket with a thin wire to raise the voltage for better overclocking. This is old school as most of the decent late model overclocking mobos like Abit's NF7 can give you all the voltage your components can handle through BIOS settings.