[NEWS] AMD Athlon64 4000+ Underclocking

[Mr.Guvernment]

WHy do these lame sites use video card based testing for a CPU!!!!!!

+--------------------------------------------------------------------+
| AMD Athlon64 4000+ Underclocking |
| from the slowing-things-down dept. |
| posted by CowboyNeal on Saturday June 04, @10:29 (AMD) |
| http://hardware.slashdot.org/article.pl?sid=05/06/04/1341241 |
+--------------------------------------------------------------------+

Bios_Hakr writes "PC Stats is running an article on their [0]experiences underclocking an AMD 4000+ processor. Their goal was to try and reduce the voltage requirements and lower the heat output. They benchmark using [1]3dMark01, [2]3dMark05, as well as [3]SuperPi. From the article: 'This got us thinking though; what about under-clocking? Most modern processors and motherboards can just as easily run under a rated speed as it can run over... but is there a point to this? Well possibly.'"

Discuss this story at:
http://hardware.slashdot.org/comments.pl?sid=05/06/04/1341241

Links:
0. http://www.pcstats.com/articleview.cfm?articleid=1798&page=1
1. http://www.futuremark.com/products/3dmark2001/
2. http://www.futuremark.com/products/3dmark05/
3. http://www.google.com/search?q=supe...ient=firefox-a&rls=org.mozilla:en-US:official

 

[felinusz]

3DMark2001 is quite CPU-Intensive on a modern machine, SuperPi is CPU and Memory intensive, and 3DMark2005 is GPU based. With memory speeds remaining the same, SuperPi becomes a CPU-specific benchmark.

Only one of the three is "video card based testing", I don't see why you're complaining.

I'll bet that their results will reflect these benchmark choices, with a small difference in 3DM05, and large differences in 3DM01/SPi.

 

[yeha]

i wish the emphasis wouldn't be on underclocking. it should be on undervolting.

voltage plays a much larger role in overall power consumption than frequency - a venice at 2.7 ghz and 1.1 volts (reached prime-stable with an 0516 chip over at coolaler) uses fewer watts than when run at 1.8 ghz and 1.4 volts. plus with many 90nm chips reacting poorly to voltage, sacrificing frequency by 7% to reduce heat by a full 50% (2900/1.5v -> 2700/1.1v) seems worth it for me.

 

[Mr.Guvernment]

3dmark01 i thought was more "overall" or was that 03' ? SuperPI yes - but i dont see the point in using any app that is not CPU intensive directly when testing a CPU.

SuperPI
Prime95
f@h
seti

and god knows how many other apps out there that are made for testing just the CPU.

 

[gvblake22]

I thought that PCStats "Underclocking" Article was very interesting; especially where you can see that even the same processor running at double speed doesn't translate to double performance!