I must admit, Intel's Core Duo is quite possibly the coolest name to come out of Intel in at least a decade. As you can probably guess, part 2 is up and I think it's a pretty complete look at what Yonah is and what Conroe/Merom could be.
At this point, I'm fully expecting Conroe to be competitive with the Athlon 64 but not a significant improvement in performance. There is obviously a lot of time between now and 2H 2006, so there may be things to change my opinion, but I don't think AMD is necessary making the wrong move by keeping their architecture fairly stagnant next year. The only negative aspect of AMD's policy of standing still in 2006 is that it will allow Intel to play catch up, rather than allow AMD to maintain their performance lead.
I do wonder if AMD is just betting on Intel not embracing an on-die memory controller anytime soon because of the aftermath of the Rambus situation, as you will see, Yonah can pretty much do anything the Athlon 64 X2 can do but without an on-die memory controller. If Intel were to integrate one, they may be able to do to the Athlon 64 what it did to the Pentium 4 (to a lesser extent obviously).
I am also curious about how much of an impact the wider 4-issue Conroe core will actually have on performance. Intel has been talking about how ILP is dead and TLP is the future of extracting processor performance, but I would say Conroe's existence is proof of the opposite. Or maybe it's just that Conroe's improvements are the last major ones we'll see to ILP for the foreseeable future.
This week ATI will hopefully be releasing their Catalyst 5.13 drivers and there may be another launch or two from the folks in Canada. Despite what you'd think, we're actually told that they will have product available at these launches before the end of the year. Only time will tell, but at least ATI is admitting to and talking about the availability problem in the sense of fixing it these days.
I've got to run and tackle a lot of errands around the house and town today before we leave tomorrow. Enjoy the day and I may chime in later today about some Xbox 360 stuff.
Take care.
At this point, I'm fully expecting Conroe to be competitive with the Athlon 64 but not a significant improvement in performance. There is obviously a lot of time between now and 2H 2006, so there may be things to change my opinion, but I don't think AMD is necessary making the wrong move by keeping their architecture fairly stagnant next year. The only negative aspect of AMD's policy of standing still in 2006 is that it will allow Intel to play catch up, rather than allow AMD to maintain their performance lead.
I do wonder if AMD is just betting on Intel not embracing an on-die memory controller anytime soon because of the aftermath of the Rambus situation, as you will see, Yonah can pretty much do anything the Athlon 64 X2 can do but without an on-die memory controller. If Intel were to integrate one, they may be able to do to the Athlon 64 what it did to the Pentium 4 (to a lesser extent obviously).
I am also curious about how much of an impact the wider 4-issue Conroe core will actually have on performance. Intel has been talking about how ILP is dead and TLP is the future of extracting processor performance, but I would say Conroe's existence is proof of the opposite. Or maybe it's just that Conroe's improvements are the last major ones we'll see to ILP for the foreseeable future.
This week ATI will hopefully be releasing their Catalyst 5.13 drivers and there may be another launch or two from the folks in Canada. Despite what you'd think, we're actually told that they will have product available at these launches before the end of the year. Only time will tell, but at least ATI is admitting to and talking about the availability problem in the sense of fixing it these days.
I've got to run and tackle a lot of errands around the house and town today before we leave tomorrow. Enjoy the day and I may chime in later today about some Xbox 360 stuff.
Take care.
12 Comments
View All Comments
vijay333 - Tuesday, December 20, 2005 - link
fyi: http://anandtech.com">http://anandtech.com doesn't seem to link to the main site anymore...ksherman - Saturday, December 24, 2005 - link
seems to be working now ;-)Wahsapa - Friday, December 23, 2005 - link
BUMPmonsoon - Monday, December 19, 2005 - link
hey ANAND, thanks for the review !=)...do you think this CPU will make it in the coming ( January ) new Apple computers ?
i'm really hoping for a Mac Mini with one of these inside...
=P
Quiksel - Monday, December 19, 2005 - link
at least it doesn't have "PENTIUM" in the title anymore.we all loved the P-M naming because it didn't continue the Pentium/II/!!!/4 convention.
This is good news. Too bad AMD didn't come up with it; they are always in use of a good name for a proc...
Remember how much everyone hated the name "Athlon"???
mlittl3 - Monday, December 19, 2005 - link
I really loved Part II of the review. Great job!Competition is really helping the consumer nowadays. Go AMD and Intel!
IntelUser2000 - Thursday, December 22, 2005 - link
Why does Anandtech use reference timing for memory for new products and use lowest latency RAM for older products?? That kinda skews the benchmarks in favor of the older products.Plus, I have seen how Pentium M likes single channel DDR2-533 better than dual channel DDR2-400. Weird, but it takes no advantage of dual channel at all. It didn't even get more bandwidth.
What I am saying is Yonah should use DDR2-667 since that's what the top-end Yonah notebooks would use.
And the comment about clock speeds. There will be Yonahs with near 2.6GHz clock speeds. They will also consume more power but since they will be DTR chips, it will be pretty good at ~50W TDP. Its called "E" series and they will be clocked 1-2 speed grades higher than the fastest "T", which means we'll see Yonah with 2.5-2.67GHz. That will keep it very competitive with X2's.
Viditor - Tuesday, December 27, 2005 - link
Have you seen this anywhere other than the Digitimes story? With Merom coming out at the end of 2006, I doubt they will be adding that many speed grades...
Also, with Vista being released at 64 bit in October, it seems to me that Yonah may have the shortest lifespan of any chip we've ever seen...
ksherman - Monday, December 19, 2005 - link
"Coolest name" ?!?!?! I tink you need to get some rest Anand ;-)Furen - Monday, December 19, 2005 - link
Hah, seriously. Core Solo/Duo is the lamest name ever, though I suppose it fits the whole "Centrino" thing.