Athlon 64 X2 5600 Windsor2.8ghz vs Athlon 64 X2 5400 Brisbane 2.8GHz

[elicgonzales]

The windsor has 2x1mb l2cache the brisbane only has 512x2mb l2 cache

Which one is better for oc'ing,

which one is better for "stock use"


Thanks for the input.

 

[cusideabelincoln]

Depends on your definition of better. The Windsor will have better performance at stock, but it also uses more power. The Brisbane is potentially better for overclocking because it comes with an unlocked multiplayer (for the black edition). But realistically, don't expect either chip to reach or go much beyond 3.2 GHz.

 

[bryanW1995]

hint: higher number = better performance.

 

[geokilla]

Originally posted by: bryanW1995
hint: higher number = better performance.

You forgot to mention that only applies if it's stock speeds...and of the same CPU models and revision and crap... How could you Bryan?

The Windsor should max out at more or less the same speeds as the Brisbane, so depending on the price vs performance, go with what you're comfortable with. What are your other computer parts?

Oh, and welcome to the forums.

 

[elicgonzales]

Originally posted by: bryanW1995
hint: higher number = better performance.

You must be implying that there are a rather LARGE number of morons that browse this website, first of all. You are as sarcastic as you can be and of no help at all. I would rather not hear any of your commentary.

I think it is fairly simple to note that, for the same processor, the next in the lineage will be better (higher number), a no-brainer. But I am looking at two different processors.

Second, I am looking to build a quasi-budget computer and would like to see the best performance, not necessarily the best power consumption. With one having larger L2cache, that is all I noticed that is different aside from the power consumption. I will spend under 250 bucks on cpu/mobo/ram/video card.

So basically, either processor will be about the same with mild overclocking eventhough one has larger L2 cache? or will the brisbane be a better choice with more OC'ing possible even with less cache?

Thanks geokilla, appreciate the help.

 

[error8]

I'd go for the Brisbane anyway, I just couldn't feel good in 2008, buying a 90nm cpu like the 5600+ Windsor. Like others have said, the 5600+ is a bit faster at stock ( a performance that you won't feel really) and they will both overclock at about the same level, although the Brisbane might have a little more headroom in it. Now, the 90nm "elephant" would suck a lot of power at 3-3.2 ghz, opposed to the 5400+ at the same clock and that matters sometimes, because you could have issues in keeping the cpu's temperature at a normal level.

I had a Brisbane once and it was a good cpu. You can't fail with it.

 

[elicgonzales]

thanks for all of the good advice. I think i might as well go for the lower power / ~same performance cpu. Thanks for the input. I am just getting into pc building. Thanks again

 

[cusideabelincoln]

Originally posted by: error8
I'd go for the Brisbane anyway, I just couldn't fell good in 2008, buying a 90nm cpu like the 5600+ Windsor. Like others have said, the 5600+ is a bit faster at stock ( a performance that you won't feel really) and they will both overclock at about the same level, although the Brisbane might have a little more headroom in it. Now, the 90nm "elephant" would suck a lot of power at 3-3.2 ghz, opposed to the 5400+ at the same clock and that matters sometimes, because you could have issues in keeping the cpu's temperature at a normal level.

I had a Brisbane once and it was a good cpu. You can't fail with it.

....

The whole "buying a processor in 2008 that was made in 2006" is the worst ****ing argument I keep hearing on these forums. Just please, stop using it. The 5600+ is, bar none, FASTER THAN THE 5400+. Honestly, its only disadvantage is that it consumes more power, and that's the only logical, rational, argument people should be using.

But the 5400+ holds two disadvantages to the 5600+: It has half the L2 cache AND the L2 cache runs at a higher latency. So a Brisbane core needs a 1-200 MHz clockspeed advantage over a Windsor core to even match its performance, and it may not even match it in applications that heavily use the L2 cache.

 

[A554SS1N]

Me agrees with the guy above ^

 

[error8]

Originally posted by: cusideabelincoln

Originally posted by: error8
I'd go for the Brisbane anyway, I just couldn't feel good in 2008, buying a 90nm cpu like the 5600+ Windsor. Like others have said, the 5600+ is a bit faster at stock ( a performance that you won't feel really) and they will both overclock at about the same level, although the Brisbane might have a little more headroom in it. Now, the 90nm "elephant" would suck a lot of power at 3-3.2 ghz, opposed to the 5400+ at the same clock and that matters sometimes, because you could have issues in keeping the cpu's temperature at a normal level.

I had a Brisbane once and it was a good cpu. You can't fail with it.

....

The whole "buying a processor in 2008 that was made in 2006" is the worst ****ing argument I keep hearing on these forums. Just please, stop using it. The 5600+ is, bar none, FASTER THAN THE 5400+. Honestly, its only disadvantage is that it consumes more power, and that's the only logical, rational, argument people should be using.

Well that is basically what I have said, but you seem to have a problem with the way I've put it on this thread. Buying a cpu made in 2006, on 90nm, in 2008, is synonymous with "sucking " more power then a 65nm cpu. Being made on 90nm, in 2006 or not, is the cause that creates the "more power sucking" effect. Doesn't this seems logical to you?

Ups, just noticed that I've wrote "fell" instead of "feel".

 

[edplayer]

Originally posted by: elicgonzales

Second, I am looking to build a quasi-budget computer and would like to see the best performance, not necessarily the best power consumption.

apart from having more L2 cache, the Windsor also has lower L2 cache latency compared to the Brisbane:

http://www.tomshardware.com/re...fight-back,1455-5.html

http://www.anandtech.com/cpuch...howdoc.aspx?i=2893&p=3


Edit: Didn't notice that cusideabelincoln already posted this. Oh well, here are some benchmarks