My New AMD system

still_runnin · Sep 7, 2003

First off, I would like to say thankyou to everyone for all of their knowledge about computers. I would be stuck with a compaq POS if you wernt for you guys (and girls). So now the unvailing of my system. I didn't know whether to go with the 1700+ or the 2500+. At the last minute, I decided to go with the 1700+ for xcal just save a few bucks and brag my system is massively overclocked. The mobo was a no-brainer. The NF7-S rev2.0 also from xcal. Wanted it for the overclocking and the sound. Wanting only the sound from my speakers and not the sound of fans, I bought the Zalman 7000A to cool my cpu along with 3 panaflos. The Zalman probably explains my high temps though. I will get to that later

I bought the Twinmos 3200 with the Winbond CH-5 chips. I'm running 2x256 in dual channel. The memory seems to be fine though. Im running it at 210 right now. I havn't messed with the timings yet. I was looking at getting a 9500/9600pro for my system, but I couldnt pass up the deal on newegg for a refurbed 9700Pro AIW for 235 so I picked that up instead.

For thd drives, I have a Maxtor 160Gb 7200RPM, ata133, 8mb hard drive. Great for all the tv I record with my AIW. I also have a generic dvd player. I'm waiting for bestbuy to have the liteon for cheap after rebates. Besides that I have a silver x-dreamer, a 19in flat CRT, NO FLOPPY, a cold cathode, a few rounded cables, mx500, logitech elite keyboard, and the logitech z-340 speakers. Yes, I am a logitech fanboy.

As for my scores, temps, and speeds.... My CPU is running at 11.5x210=2415 right now at a 1.725V core. I havn't tried any higher multipliers yet, but I will. The temps are a little high though. My idle is 44C and I dont want to know what load is. I don't think I will be able to push the CPU much higher though because of the high temps, but the 1700+ is crazy how it overclocks so we'll see. I knew when I bought the Zalman I was going to be sacrificing Mhz, but thats ok. My 3dmark 2001se is only about 16500 (haha I make that sound like a bad thing). Again, I did no tweaking to increase my score. I'm pretty sure I can break 17000 or even 17500 in the future no problem. Not a bad system for a little over $900.

I don't know how to post pics so I can't get a picture for you all. Thanks again for the help and feel free to ask questions of give suggestions.

hitechjb1 · Sep 7, 2003

Nice setup for your objective of low noise with air cooling. If you can get your system stable to run at 210 MHz FSB and 2.4 GHz CPU, a 5-10% lower in FSB and CPU from that by pushing them to their limit (using voltage) would not bear too much practical impact on your system.

One thing I'd like to point out is that, if CPU MHz or raw power is not the only avenue, for achieving the same system performance which including CPU raw power, cache and memory performance, memory bandwidth, the Barton 2500+ would consume less power than the Tbred B 1700+ DLT3C. The main argument is that Barton 2500+ can achieve the same level of system performance as the 1700+ by virtue of the larger L2 cache, even the Barton 2500+ is running 5% lower in CPU MHz and raw power compared to a 1700+. As a result, Barton 2500+ consumes lower CPU power (~9% lower) and die temperature (for the same cooling).

From the following overclocking characteristics of a 2500+ and 1700+, at 1.8-1.9 V, by reading from the graph, a 2500+ running about 7% lower in MHz, would consume about 12.5% lower power and about 3 C lower in die temperature.

Overclocking characteristic (voltage, frequency, power, current, temperature) of Tred B 1700+ DLT3C and a Barton 2500+
(to get to the graph, may have to scoll the page up and down for the right post).

Further, for setup with a less aggressive HSF to obtain the same system performance, lower power is less likely running into stability problem due to temperature than the case by simply pushing voltage and CPU MHz.

hitechjb1 · Sep 7, 2003

L337 M33P said:
Hitechjb1, are those data points from the graph calculated or empirically determined?

I would imagine certain chips hitting a "wall" or "flat spot" at a certain voltage above which does not yield any more MHz for the temperature and instability increase.

The characteristics are created based on calculation using a CPU model on voltage, frequency, power, temperature, faster/slower chips, cooling thermal resistance, ... The details are in the various posts for those characteristics. The model gives what temperature would be for a CPU frequency and voltage, it does not model instability due to high die temperature. But one can read and determine the what-if relationship between frequency, temperature, voltage for a given cooling, and faster/slower chips, ...

For example, some quotes from
Overclocking characteristic (voltage, frequency, power, current, temperature) of a Tbred B 1700+ DLT3C

hitechjb1 said:
Overclocking characteristic (voltage, frequency, power, current, temperature) of a Tbred B 1700+ DLT3C

This is an overclocking characteristic on CPU voltage, frequency, active power, leakage power, total power, die temperature
of a Tbred B 1700+ DLT3C generated based on calculations.

Conditions:
Tbred B 1700+ DLT3C, raw data from AMD Model 8 data sheet
Power based on rated worst case data, with a 90% active power model (described earlier)
Vcore 1.50 - 2.2 V
Cooling thermal resistance 0.22 C/W (such as SLK-900U w/ Vantec Tornado)
ambient system temperature 30 C
CPU was stable (full load) at 2200 MHz at 1.5 V rated Vcore.

The rest of the data are generated from calculation.
It turns out to be accurate and consistent with the actual measurement on my 1700+ overclocking
on frequency, Vcore, temperature, ...

If one has a slighter faster or slower chip, such characteristic may vary slightly.
If one knows about
- the full load overclocking frequency at rated voltage,
- another data point such as the full load overclocling frequency at rated voltage + 50 mV (or 100 mV)
- the cooling used and the ambient system temperature
an overclocking characteristic for that particular CPU can be generated also.

The characteristic for a given CPU is useful to project how the CPU varies with Vcore subject to heat constraint,
and what temperature, power, current would be for a given cooling, ambient system temperature, ...

Other data and graphs such as
- overclocking characteristics for other CPU's, such as Tbred B/Barton, P4, ...
- overclocking characteristics for slower, faster chips of a given CPU type
- overclocking characteristics to compare multiple CPU's
- overclocking characteristics of a CPU with different cooling, ...
- effect of cooling on overclocking for other HSF, water, extreme cooling (based on C/W)
- effect of ambient system temperature on overclocking characteristic
- ...
(will post some of these graphs).

Observations:
The calculation and characteristic show that this particular 1700+,
- at 1.95 V, it can do 2.59 GHz, 61 C (w/ 30 C ambient system temperature),
- at 1.95 V, it dissipates 142 W power, draws 72 A current
- from 1.5 V to 1.95 V, the power rises at an increasing rate (3rd power),
from 72 W (2.2 GHz) to 142 W (2.59 GHz)
- from 1.5 V to 1.95 V, the temperature rises at an increasing rate,
from 46 C (2.2 GHz) to 61 C (2.59 GHz)

- it could do 2.63 GHz at 2.0 V, 63 C, 30 C ambient system
(potentially, but may fail due to other constraints, stability, ...)
- it could do 2.70 GHz at 2.1 V, 68 C, 30 C ambient system
(this seems impossible due to heat constraint and stability, as we believe)
...

hitechjb1 · Sep 8, 2003

L337 M33P said:
OK, I will probably compile a list of the minimum voltages and maximum speeds for my processor as soon as I get the time - I have to prime95 each one for 1/2 an hour =x

Then you'll have another data set to work with, and possibly refine the model from it as I don't think that the straight-line theory holds out at the top end.

If you have some actual data points including
- CPU type
- stock voltage and its max stable load frequency, plus one or two more such points
(the model needs only very few points (two or three) to characterize the overclocking characteristic)
- cooling used (if possible w/ C/W number)

I can generate a overclocking characteristic for you.

The model is not strictly linear, it takes into account the slow down of the CPU due to heat and temperature of CPU.

I suggest, we move back to the original thread in which these characteristics posts were started, than threadjacking here.

Tbred B 1700+ DLT3C 2.5+ GHz air results

L337 M33P · Sep 8, 2003

Agreed. I'll transfer my posts

Edit - now in the 2.5GHz Tbred B results thread, and I quoted your posts to maintain consistency.

My New AMD system

still_runnin

Registered

hitechjb1

Senior Member

hitechjb1

Senior Member

hitechjb1

Senior Member

L337 M33P

Member

Similar threads