AMD CPU Efficiency Compared

Well when you look back at the first A64's(754) 89(as far as i know) watts for the 3200+(single core 2.2GHz 512k cache) things have improved quite a bit over the time the A64 has been around....Maybe i missed it(i do that sometimes) were the power tests with or without the power supply losses?

I know it starts a fight, but you should try to do a Intel vs AMD on again now...
 

zenmaster

Splendid
Feb 21, 2006
3,867
0
22,790
The other problem with the article is that the tests used generally did not scale to 4 cores. The raw performance data between the dual and quads did not give the quad core its full due.

The other problem is that the Phenom is not designed to be a low power solution for geneal computing tasks. Which is what they sort of tested.

It can be low power for heavy duty tasks.

Until AMD releases a Dual-Core version of the Phenom might that qualify.

I also would have been very curious to see the 5000+ BE.
That tends to run quite cool with some oomph.

That may have done better on the charts.
 

krisia2006

Distinguished
Aug 1, 2006
9
0
18,510
"revealed performance per watt improvements of over 400% if you compare the a 3 GHz Pentium 4 630 and a 3 GHz Core 2 Duo E6850"

Isn't that outdated? I have an E8400 in my case...
 

teldar

Distinguished
May 8, 2006
58
0
18,630
Not to mention that they compared a P4 3GHz to a core 2 duo, which are SIGNIFICANTLY different architectures while the AMD cores are all basically the same architecture over just two generations of processes. I think that AMD's 60% reduction in power over the same architecture is fairly impressive.

I think they could have drawn another conclusion about the strength of the Athlon architecture vs the craptacular mArch of the P4, which was a mistake from Day 1, if we're going to be truthful. It was Intel saying all that mattered was GHz, not performance.

T
 

endyen

Splendid
I really liked the power consumtion graphic. It was so pretty will those lines going mostly across, with the odd squigle up and down.
Then again, to be a valid cpu benchmark, you would expect to see the straight across part at the top wouldn't you?
It brings up a good point. Why do you not show cpu usage in your benchmarks?
It would certainly show floptimization by one company or the other.
 

Hellboy

Distinguished
Jun 1, 2007
1,842
0
19,810
You mean AMD processors are now efficient.


At what !

Some one tell me cos I think Thundermans rotted me brain......


Come back Sempron ( Ooops Phenom ) All is forgotten..

I had a tri core once - when i was 3 and it fell over round corners when i went to fast !!!!! Now where have i heard that one before.....
 

rodney_ws

Splendid
Dec 29, 2005
3,819
0
22,810
I felt the % comparisons (400 to 60) were a bit unfair... mainly because Intel's Netburst CPUs were about the worst (i.e. most inefficient) CPUs ever made... and of course whatever Intel comes out with next (C2D) is going to be light years ahead of that.
 

bloodem

Distinguished
Sep 10, 2007
14
0
18,510
of course intel is now 400% more efficient because they compared a Core 2 Duo with a Prescott (which if we all remember, would requiere a 500W PSU, and it's idle temp would be somewhere between 50 - 60 degrees) Now I`m no Intel/AMD fanboy but please... that comparison is flawed.
 

bloodem

Distinguished
Sep 10, 2007
14
0
18,510
of course intel is now 400% more efficient because they compared a Core 2 Duo with a Prescott (which if we all remember, would require a 500W PSU, and it's idle temp would be somewhere between 50 - 60 degrees) Now I`m no Intel/AMD fanboy but please... that comparison is flawed.
 


I thought the same thing!

It's easy to improve when what you're comparing something to a train wreck.

Speaking of train wrecks, anyone buy a Phenom yet?
 

Hellboy

Distinguished
Jun 1, 2007
1,842
0
19,810



Lol, I like they way you left out the Mrs bit......

OOOOOhhh thats a hot response :)

Keep it up TC
 

zenmaster

Splendid
Feb 21, 2006
3,867
0
22,790


So how does and article that uses the latest and greatest AMD Chips but uses older Intel Chips that are quite inferior to the current chips in both raw performance and power requirements work as an ad for Intel?