New dual-core Phenoms 6300+, 6400+, 6500+

[coolpurplefan]

http://xtreview.com/addcomment...w-kuma-processors.html

Here's a little info on the 6300+ and 6400+. And some on the 6500+ here:
AD6500WCGHBOX
http://www.xtremesystems.org/F...howthread.php?t=200291

lol, it's getting weirder every day.

Apart from overclockability, what's the advantage of those compared to an Athlon at 3.1 GHz? They're only at 2.1, 2.2 and 2.3 GHz.

 

[cusideabelincoln]

Well these new parts should provide much better peformance per clockspeed than the Athlon X2s. I'd say you can expect these chips to perform very similarly to Intel's first lineup of Core 2 Duos at similar clockspeeds, since the Phenom core seems to be about as fast as the first Core 2s per clockspeed.

So I would expect a dual core Phenom at 2.3 GHz to match or exceed the performance of a 3.1 GHz Athlon X2.

 

[harpoon84]

I've seen a few benchmarks and I doubt a Phenom X2 @ 2.3GHz can outperform a 3.1GHz X2, especially if its a Windsor core. A Brisbane @ 3.1GHz would be closer, but still slightly faster considering a 2.8GHz X2 5400+ competes well with the 2.3GHz Phenom 6500:

http://www.xtremesystems.org/F...=3322002&postcount=190

 

[Bateluer]

I think they'll be offering these as a well to sell off inventory that failed to meet quality control for X3s and X4. They will have half their cores disabled for failing validation, and AMD still wants to sell them to someone to make it less of a loss.

Speculating on performance without knowing what their end price will be is a bit premature, IMO.

 

[Idontcare]

Originally posted by: Bateluer
Speculating on performance without knowing what their end price will be is a bit premature, IMO.

Couldn't agree more. In fact since we never know the end price, as it varies week-to-week after release, it makes your statement all the more valid and we should not be discussing any CPU's performance, ever.

We should simply never discuss performance or speculate on anything until absolutely everything is known...although no doubt you will encounter some naysayers who would be so pithy as to argue that at that point you aren't speculating once you know everything.

Off topic, but I wonder why people bother coming to this CPU forum anyways. Its like they all want to talk about stuff and ask folks quesitons about things which they themselves don't know everything about yet or something. Weird huh. Wonder why such an idea got into their heads.

 

[coolpurplefan]

Originally posted by: harpoon84
I've seen a few benchmarks and I doubt a Phenom X2 @ 2.3GHz can outperform a 3.1GHz X2, especially if its a Windsor core. A Brisbane @ 3.1GHz would be closer, but still slightly faster considering a 2.8GHz X2 5400+ competes well with the 2.3GHz Phenom 6500:

http://www.xtremesystems.org/F...=3322002&postcount=190

OMG, thanks for the link. Now I realize my choice of 5400+ was a little odd since the e5200 outperforms it. lol Except it came out right after I bought my CPU if I'm correct.

I guess I was right they might be bought by people who want to overclock. I'm surprised 2.3 GHz = 2.8 GHz when you change cores.

 

[Idontcare]

Originally posted by: coolpurplefan

Originally posted by: harpoon84
I've seen a few benchmarks and I doubt a Phenom X2 @ 2.3GHz can outperform a 3.1GHz X2, especially if its a Windsor core. A Brisbane @ 3.1GHz would be closer, but still slightly faster considering a 2.8GHz X2 5400+ competes well with the 2.3GHz Phenom 6500:

http://www.xtremesystems.org/F...=3322002&postcount=190

OMG, thanks for the link. Now I realize my choice of 5400+ was a little odd since the e5200 outperforms it. lol Except it came out right after I bought my CPU if I'm correct.

I guess I was right they might be bought by people who want to overclock. I'm surprised 2.3 GHz = 2.8 GHz when you change cores.

I would like to see the power consumption numbers for the new Kuma processors. The TDP on these parts should be sub-45W given the clockspeed and tha fact there are half the cores active compared to an X4.

 

[harpoon84]

Originally posted by: IdontcareI would like to see the power consumption numbers for the new Kuma processors. The TDP on these parts should be sub-45W given the clockspeed and tha fact there are half the cores active compared to an X4.

Unfortunately it looks like they took the worst bins (in terms of thermals). Compared to an X2 5000+ it consumed ~20W more despite being clocked 300MHz lower: http://en.expreview.com/2008/0...500-kuma-review.html/3

This actually makes it worse than Brisbane in performance/watt, and its definitely not 'half' the power consumption of a Phenom 9650, in fact both chips share a 95W TDP... go figure.

 

[DrBombcrater]

Originally posted by: coolpurplefan
Apart from overclockability, what's the advantage of those compared to an Athlon at 3.1 GHz? They're only at 2.1, 2.2 and 2.3 GHz.

The Phenoms are a better choice if you're building a system with integrated graphics because their faster HyperTransport links give the IGP more memory bandwidth. I recall the 780G chipset benchmarking about 50% quicker with a Phenom than a K8-based Athlon X2.

 

[coolpurplefan]

Originally posted by: DrBombcrater

Originally posted by: coolpurplefan
Apart from overclockability, what's the advantage of those compared to an Athlon at 3.1 GHz? They're only at 2.1, 2.2 and 2.3 GHz.

The Phenoms are a better choice if you're building a system with integrated graphics because their faster HyperTransport links give the IGP more memory bandwidth. I recall the 780G chipset benchmarking about 50% quicker with a Phenom than a K8-based Athlon X2.

Gee, that's very interesting to know.

 

[edplayer]

Originally posted by: DrBombcrater
The Phenoms are a better choice if you're building a system with integrated graphics because their faster HyperTransport links give the IGP more memory bandwidth. I recall the 780G chipset benchmarking about 50% quicker with a Phenom than a K8-based Athlon X2.

They do give the IGP more memory bandwidth but I doubt they average 50% quicker. Maybe on certain tests they can achieve that but gamers won't see 50% better fps on average

 

[SolMiester]

Originally posted by: Bateluer

Speculating on performance without knowing what their end price will be is a bit premature, IMO.

While I understand your reasoning here, you are talking about performance per\cost!

Price has nothing to do with performance, the car will reach 100mph or not.

 

[zach0624]

This could turn out to be interesting, its good to see that AMD hasn't abandoned dual cores quite yet, because the truth is that most people don't need quads or quad thread processors and to cores is usually enough since probably 99%(just an estimate, don't freak out) of apps are single threaded any way.