Phenom II On Ice: AM3 Overclocked With LN2

Status
Not open for further replies.

accessgranted

Distinguished
Feb 10, 2009
9
0
18,510
0
[
At 5.2 GHz, we were able to run this particular bench at 16 minutes and 36 seconds, compared to over 28 minutes at stock. This is an increase of over 40% in performance.
It's a 75% (28/16 = 1,75) in performance, not 40%.
 

JustPlainJef

Splendid
Feb 20, 2001
9,699
0
30,780
0
No, AccessGranted, you are wrong. If the first one benched at 28 minutes, and it ran OC'ed in 14 minutes, that's a 50% increase. Since it ran in 16:36, that's less than a 50% performance increase. Same with the second one. 18.8 / 2 is 9.4. Benchmark finished in 10.6, which is less than a 50% increase. You are doing the math backwards as you always compare to the original number. 10.6/18.8 = 56% of the original or a 44% decrease in time.
;)
 

macer1

Distinguished
Mar 6, 2006
424
0
18,810
5
Toms editor should really get his glasses checked. that of find a new job with all the mistakes in this artical..


but good to see from AMD.
 

accessgranted

Distinguished
Feb 10, 2009
9
0
18,510
0
If a benchmark takes 2 minutes at default and then 1 minute overclocked that's a 100% increase in performance. 100% increase in performance = 50% decrease in time. It's twice as fast thus taking half the time. It can't be any other way.
 

accessgranted

Distinguished
Feb 10, 2009
9
0
18,510
0
Ok, lets say a CPU scales perfect with clockspeed.

X Mhz:
Benchmark 1: 20 fps
Benchmark 2: 20 sec

2X Mhz
Benchmark 1: 40 fps
Benchmark 2: 10 sec

It's 100% faster -> twice as many fps, half as long rendering time.
 

cangelini

Contributing Editor
Editor
Jul 4, 2008
1,878
9
19,795
4
[citation][nom]macer1[/nom]Toms editor should really get his glasses checked. that of find a new job with all the mistakes in this artical..but good to see from AMD.[/citation]

Irony?
 

Lozil

Distinguished
Feb 1, 2008
39
0
18,530
0
It looks good to see AMD back... But What AMD needs now Is some winner CPU's which can take the crown away from Intel....
 

accessgranted

Distinguished
Feb 10, 2009
9
0
18,510
0
cangelini, the formula you wrote is only good for benchmarks where an increase in performance equals an increase in the unit that is measured (fps for example).

The correct formula to use when an increase in performance equals a decrease in the unit measured is (Old/new)-1 = performance increase in percent
 

StupidRabbit

Distinguished
Jan 27, 2009
60
0
18,630
0
are we going to see something similar with the i7? i mean, is it possible or is there some crazy restriction.. (or something i missed in a different article)
 

foreignalien

Distinguished
Feb 10, 2009
1
0
18,510
0
Quote:
accessgranted 02/10/2009 10:16 AM
It's a 75% (28/16 = 1,75) in performance, not 40%.

The operative word is performance. It is a ~75% increase in performance and a ~40% decrease in time. I have to agree, TH got it wrong.

Numbers aside, awesome test!
 

cangelini

Contributing Editor
Editor
Jul 4, 2008
1,878
9
19,795
4
[citation][nom]accessgranted[/nom]cangelini, the formula you wrote is only good for benchmarks where an increase in performance equals an increase in the unit that is measured (fps for example).The correct formula to use when an increase in performance equals a decrease in the unit measured is (Old/new)-1 = performance increase in percent[/citation]

I'm sorry, you're absolutely right Access. I've edited all of the author's original numbers to reflect the proper percentages. Thanks for the correction!
 

ravenware

Distinguished
May 17, 2005
617
0
18,980
0
[citation][nom]macer1[/nom]Toms editor should really get his glasses checked. that of find a new job with all the mistakes in this artical..but good to see from AMD.[/citation]

You should have your English checked.
 

jcknouse

Distinguished
Oct 23, 2008
447
0
18,780
0
This article has me excited to see all of the AM3s when they come out.

This is making me want to build my first liquid-cooled gaming machine in the fall, which is hopefully when all the manufacturers should have their MBs on the market.

Nice article, guys.
 

accessgranted

Distinguished
Feb 10, 2009
9
0
18,510
0
During the baseline run, AMD's Phenom II X3 720 Black Edition earned a score of 3463. With the processor running at -190 C at 4.5 GHz, we obtained a CPU score of 5262, which is over 34% faster.
Another error ;). It certainly is more than 34% faster but how much faster? 5262/3463 = 1,52 = 52% faster.

 

mapesdhs

Distinguished
Jan 22, 2007
2,507
0
21,160
111
No, it's 40% (40.72 to be exact), ie. (1 - ((16 * 60) + 36) / (28 * 60)) * 100.
Your math is the wrong way round.

Think about it; a 50% improvement would mean a time of 14 minutes,
so how can 16.5 minutes be a 75% improvement?

A 75% improvement would be 7 minutes.

Ian.

 

accessgranted

Distinguished
Feb 10, 2009
9
0
18,510
0
No, it's 40% (40.72 to be exact), ie. (1 - ((16 * 60) + 36) / (28 * 60)) * 100.
Your math is the wrong way round.

Think about it; a 50% improvement would mean a time of 14 minutes,
so how can 16.5 minutes be a 75% improvement?

A 75% improvement would be 7 minutes.

Ian.
They already edited it the way I said it was. Your math is wrong.

Is it difficult to understand that if something is 100% faster it can finish a job in 50% time? I don't think so. It's intuitive to say the least.

If it takes you 10 minutes to get to the store and you double your speed (i.e. 100% performance improvement), how long would it take you to get there? Come on, this can't be that hard :)
 

JumpKickJoe

Distinguished
Jan 15, 2009
49
0
18,530
0
Lol! I feel like I am in math and english class here guys! @ accessgranted, thanks for the sharp eyes. You made the tests a whole lot more sweeter, with your error correction.

AMD is finally showing up that muscle ppl....Who say Intel prices are gonna fall like whoa?
 
Status
Not open for further replies.

ASK THE COMMUNITY