I have to agree with Thrax on this. I think that K8L might get AMD back in the ballpark, but I don't think that it will eclipse Conroe or it's successors; at most level the playing field. Intel got caught with their pants down when netburst puked itself with the 90nm and 65nm die shrinks and they've been trying to play catch up since. But Intel does have formidable design teams and they did definitely come up with a winner for the single and dual socket market with Conroe and Woodcrest. And yes, their off-die memory controller does limit them in the future exandability as far as multi-cores per socket, but they are also working on their own version of on-die memory controller too.
Another concern I have about AMD is the state of their 65nm process. They seem to be lagging Intel quite a bit with transitioning to 65nm and I don't believe that they are dragging their feet because they have had the clear performance advantage over Intel for the last few years. After all, a smaller process = smaller core dies, which would give AMD more profit per processor and more cores per wafer. So it seems to me that AMD must be having problems with something on the die shrink, whether it's yields or something worse. I would love to see AMD come out with another performance king with K8L but I'm not holding my breath for it.
0
LeonardoWake up and smell the glaciersEagle River, AlaskaIcrontian
edited July 2006
Has anybody seen the NF590-SLI Intel Edition? It's supposed to support Conroe and SLI. I'm hoping these boards come out with Conroe!
Intel has for a long time made the best motherboard chipsets to support Intel processors, bar none. If AMD were making chipsets, I would assume theirs would also be the best for AMD processors. Now, perhaps one of the second parties makes chips for better dual graphics operations. I don't keep up with that. But for pure CPU and memory bus performance, I would take Intel chipsets over Nvidia without hesitation. But, with all that rah rah said, apparently the first series Conroes have RAID implementation problems. This seems to be a CPU problem and not a chipset problem.
Yeh, but its only having problems with raid 5 isnt it?
0
LeonardoWake up and smell the glaciersEagle River, AlaskaIcrontian
edited July 2006
Yes, I think it is just RAID 5 problems. As for me, this would not be a drawback at all. My experience with desktop RAID is that it's more hype than actual utility. But with any tech advancement, CPUs, video, motherboards, etc. I'll wait for all the brave early adopters to:
proof the early production parts
establish a user knowledge base for troubleshooting and performance
and to drive down prices
Well, I dont bother with onboard raid controllers, because they are just all softare raids, i prefer an actuall card with a controller, processor, and ram. Really, you dont need raid for a desktop rig unless you want mirroring for data integrety
That's good. I was hearing rumors of it going to an AM3 socket, which would be horrible! Hopefully AMD sticks with the AM2 socket for a while and keeps things progressing on that socket.
That's interesting. This makes it sound more like AMD still has something up their sleeve, to me.
0
LeonardoWake up and smell the glaciersEagle River, AlaskaIcrontian
edited July 2006
I don't know if AMD has anything up their sleeve or not, but Intel, once again, is shooting itself in the foot. I hope AMD will capitalize on every misstep Intel makes.
Comments
Another concern I have about AMD is the state of their 65nm process. They seem to be lagging Intel quite a bit with transitioning to 65nm and I don't believe that they are dragging their feet because they have had the clear performance advantage over Intel for the last few years. After all, a smaller process = smaller core dies, which would give AMD more profit per processor and more cores per wafer. So it seems to me that AMD must be having problems with something on the die shrink, whether it's yields or something worse. I would love to see AMD come out with another performance king with K8L but I'm not holding my breath for it.
proof the early production parts
establish a user knowledge base for troubleshooting and performance
and to drive down prices
http://www.dailytech.com/article.aspx?newsid=3169 :headbange
That's interesting. This makes it sound more like AMD still has something up their sleeve, to me.
US government unit throws Intel out over RAID problems