Intel also make chipsets so prehaps its not in there benifit to go with a on board memory controller. Prehaps doing so would make it easier for competitors (like SIS, Nvidia and VIA) to make chipsets performing as well as/better than Intel's.
Trying to hype AMD when the subject of this thread is a new gate process simply amazes me since if the C2D was on the drawing boards as AMD finds itself with the K8L perhaps the guy that brought up this subject would have a point in speculating about the amount of transistors needed to create 2, 4, 6, or 8MBs of L2 cache but the C2D has been around now for 6 months and from all I have read at every comparable clock speed the C2D has a performance advantage and then when you add the E6700 and E6800 AMD has no reply for those.
Intel has shared L2 cache meaning that single threaded application benefit from one cores ability to use the L2 cache of the secondary core. then on top of that Intel uses their new memory disambiguation to allow loads in front of stores in main memory all this plus the extended front, execution core, and back end make the C2D a better performer. More L2 Cache makes the chances of a miss and the need to go to main memory slighter than a small L2 cache and a faster Memory interface that AMD possesses now.
There is nothing left to say, other than articles have come out about the K8L and it is a fact that it will be the K8L that will imulate the C2D, using all of the technology of the core while remaining with the onboard memory controller.
I do not see why not since it seems that Intel is playing nice, nice ever since the market came to the realization that AMD was a force to be reckoned with.
As Tim said Intel is sticking with its wattage numbers for the Penryn and raising clock speeds so why would they need another socket other than if things change dramatically from now till then such as chipsets using PCI-Express 2 and manufactures want everyone to upgrade. Since Dell was a Intel house solely up to a few months ago I believe Dell had more to do with Intel continuously changing sockets than Intel themselves, Dell needs something new always to move market share.
-Morres law ends in seven years (1985)
-Intel cant continue to shrink its chips to increase speed and computing power anymore (1995)
-45 nm technology allows us to maintain moore's law (2005)
20 years after, moore's law is still there... I wonder how much longer that would last.
Funny point, Intel can get 100,000,000 transistors working 100% of the time. In LCD technology, in just 2,000,000 errors occur... I know this is more a quirk of the technology, but the accuracy which some things can be done to..
Edit: That site is sticking some nasty viruses on my pc... 2 .exe's just apeared....
Originally Posted by sl1xx arrr man this look soo good i was eyeing my conroe system up look like i got more waiting to do !
core2duo is good... but waiting for this will be even abigger step.. so hold on to those pennies... and upgrade once they release em new cores. And oh, just in time for r600 and the lower end and mid range dx10 cards from nvdia. What a happy year this is gonna be
Originally Posted by ikra core2duo is good... but waiting for this will be even abigger step.. so hold on to those pennies... and upgrade once they release em new cores. And oh, just in time for r600 and the lower end and mid range dx10 cards from nvdia. What a happy year this is gonna be
Ikra, R600 is coming out Q2 where as penryn is going to be H2 (Q3 or Q4).
Even though this looks mightily tempting I have waited long enough, I'll get the Q6600 after the price drops, and then if the penryn is that good I will sell my Q6600 afterwards (although I have to get it first lol)
Tim, good stuff, I actually understood (some of) it this time :)
Nice article, but one sentence at the bottom of page 2 has me stumped. Can someone explain what the bolded part means. To me, reducing something by 5 times makes it -4 times what it was (1 times - 5 times = -4 times), and that's plainly ridiculous...
Quote:
This, in conjunction with the high-k gate dielectric, has given Intel more than a 20% increase in transistor performance while reducing the source-to-drain leakage by more than five times, improving energy efficiency inside each transistor.
Perhaps it could be changed to a percentage or a percentage reduction or something a little easier (for me at least :D ) to understand...
I haven't posted here in years and had to come back to comment on this. I keep getting ready to upgrade and speccing out a system, and every time, Intel goes and launches something brand new like this.
As far as mobile processors go, I'm less interested in seeing what they do with laptop designs. I'd love to see them integrate this tech into pocket devices. They could probably squeeze P3/P4 level performance into PDAs and smartphones.
With the technology of 45nm construction, do you think Intel now build a quad core processor that is one unit :? or will they just be lazy and do nothing about it :?
Comments 26 to 44 of 44
Just a guess but you never know
sorry if I confused you
Intel has shared L2 cache meaning that single threaded application benefit from one cores ability to use the L2 cache of the secondary core. then on top of that Intel uses their new memory disambiguation to allow loads in front of stores in main memory all this plus the extended front, execution core, and back end make the C2D a better performer. More L2 Cache makes the chances of a miss and the need to go to main memory slighter than a small L2 cache and a faster Memory interface that AMD possesses now.
There is nothing left to say, other than articles have come out about the K8L and it is a fact that it will be the K8L that will imulate the C2D, using all of the technology of the core while remaining with the onboard memory controller.
As Tim said Intel is sticking with its wattage numbers for the Penryn and raising clock speeds so why would they need another socket other than if things change dramatically from now till then such as chipsets using PCI-Express 2 and manufactures want everyone to upgrade. Since Dell was a Intel house solely up to a few months ago I believe Dell had more to do with Intel continuously changing sockets than Intel themselves, Dell needs something new always to move market share.
-Morres law ends in seven years (1985)
-Intel cant continue to shrink its chips to increase speed and computing power anymore (1995)
-45 nm technology allows us to maintain moore's law (2005)
20 years after, moore's law is still there... I wonder how much longer that would last.
Funny point, Intel can get 100,000,000 transistors working 100% of the time. In LCD technology, in just 2,000,000 errors occur... I know this is more a quirk of the technology, but the accuracy which some things can be done to..
Edit: That site is sticking some nasty viruses on my pc... 2 .exe's just apeared....
Ikra, R600 is coming out Q2 where as penryn is going to be H2 (Q3 or Q4).
Even though this looks mightily tempting I have waited long enough, I'll get the Q6600 after the price drops, and then if the penryn is that good I will sell my Q6600 afterwards (although I have to get it first lol)
Tim, good stuff, I actually understood (some of) it this time :)
As far as mobile processors go, I'm less interested in seeing what they do with laptop designs. I'd love to see them integrate this tech into pocket devices. They could probably squeeze P3/P4 level performance into PDAs and smartphones.