O.K. Sander Sassen,

Sounds good. Thank you for the link.

Take Care,

I just received some updated info regarding the performance of AMD's M2-socket. Fitted with two low-latency DDR2-800 modules the M2-socket should perform on par with a similarly configured socket-939 using generic DDR400 memory. Needless to say that low-latency DDR2-800 is far from standard in OEM boxes, so don't expect the M2-socket to bring you any performance benefits at launch.

Best regards,

I'd be a little careful about predicting doom'n'gloom regarding AM2. The info I've seen floating around suggests these points...

The version toms hardware tested was very preliminary, and the mem controller was very broken.
The samples currently out mostly fixed the totally broken nature of the earlier one.
There are new samples shipping to partners in late april that will be further enhanced from what's out now.

I'm not saying it's going to be the second coming of the almighty, but believing that there won't be an improvement over s939 will turn out to be fantasy.

Well, I have to agree.

According to a very reliable souce of mine, he too is stumped with what is going on in the AM2 situation. It looks like AMD is giving everybody a spin for some reason, and "HE" thinks it all has to do with keeping everything low key just until Intel releases there new CPU's.

He does not think that the Socket 939 CPU's will match the Conroe in Clock for Clock performance, but he also was not hesident to say that the New AM2's have something special with them when you run a DDR2-1000+ Ram, there performance increases quite abit from the DDR2-800.

And he also states unofficially that AMD has done something special with there integrated memory controller. He hasent seen any proof of DDR3 support, but says that it is quite possible via bios update, but depending what chipset is running on the mobo.

I don't know if it's good to have brand loyalty in the CPU market. If Intel makes the fastest processor, why is it so bad if AMD doesn't? Does that mean AMD can do better but doesn't want to? Or does that mean that Intel has made a fast processor, and that's all? So if a lot of people want AMD to succeed, but if Intel does, instead of being happy about a new processor that has come out, why do people sulk because it's not AMD? Prices go down when new processors come out, and that's always good. Also, those who can afford the new processors can buy them. Not AMD? Oh, I'm sorry. They'll just have to suffer with the fastest consumer-level processor without an AMD sticker.

"If Intel makes the fastest processor, why is it so bad if AMD doesn't?"

It is not bad at all. The problem is when AMD released there Athlon 64, even though it outperformed anything Intel had to offer at the time, the IT Industry still did not give credit to AMD the way it should have.

AMD’s Athlon 64 “IS” an innovative CPU design. The true innovations were the 64-Bit & there Integrated Memory Controller, followed closely by there Hyper Transport & there SOI.
1) Did they get credit for these innovations?
2) And did they get credit for future proofing there CPU Design so that they can transition into Dual Core & Multi-Core anytime they like?
I would say NO, not at the time, it took AMD quite some time for the IT Industry to really realize what a phenomenal architecture AMD’s 64-Bit CPU’s were.

And now we get really early benchmark results from Intel resulting in the new Conroe CPU beating AMD’s finest in almost 90% of all benchmarks.

To me, this is good. It is healthy for the IT Industry to have this type of competition. And these results are unofficial, anyway, so anything can happen when they are finally released.

All I say is to give credit to innovation & new technologies.

I completely understand and have changed my mind. Thanks.

No problem Bro, I wish both companies great fortune, but may they not cheat each other IMO

Nice job with the latencies, by the way. 2-2-2-7 at 510MHz. I hope I'll be able to do the same when I get my OCZ PC-4800 ram.

With my current setup (4200+ at 2.81gigs and ram being at 533mhz 2.5-3-2-5 3.0 V) my PC totally smokes DDR2 CPU/Memory benchmarks posted on Toms hardware. I dont have the best comp, not even close to it. DDR2 looks to be totally useless until they start running DDR2-1066. Now THATS hot. But even when they do guess who is gonna run it? Yeah, 2x1gb of those sticks nicely pair up for 500$ margin JUST for the memory. I don't think the majority of the population will be able to afford it. Of course it will outperform current S939 if you're using FX-62 and DDR2-1066 memory but if you're not, S939 still offers similar if not better in most cases perfomance.

Also you see, at a certain point the higher operating frequency of the RAM doe not matter too much. I mean look at those DDR2-800 modules running at 5-5-5-15 CAS. Even though theyre much faster as far as "mhz" go, their timings go sky high basically making them almost Identical to say OCZ VX RAM OC'ed to 550 mhz at 2-2-2-7

I was wondering about that, too. There's plenty of really fast DDR ram out there. How many remember the time when PC-2700 was decidedly the final speed for DDR? I have a 1GB set of PC-4000 Patriot ram right now. It's great. Officially at 3-4-4-8 at 500MHz, it can hit much higher speeds with lower latencies. Some reviews reported this ram not being able to hit more than 490MHz, but mine runs just fine. Is it because of cost that manufacturers are turning to DDR2 for faster FSBs?

DDR2 consumes a good load less power than DDR. Additionally, it does provide a lot more bandwidth (theoretically) which will be good for the next generation processor designs, which look to be eliminating front side buses altogether in favor of high-speed interconnects.

-Pikl

Tom’s hardware? No, I don't find them at all credible.

Interconnect? Yes, I too think so. With AMD's HTT paving the way & Intel's future CSI interconnect coming anytime, this looks to be the future of CPU's.

I'd like to see single core processors come out that are faster than dual and quad-core CPUs of today. That would be even more futuristic for me. I prefer innovation over quantity.

Well, I think they are? IBM's CELL porcessor which has multi-cores in one CELL. I guess that is close enough.

But never the less, as long as companies are designing there games & app's & program's to use the Dual Core & Multi-Core, they will perform amazing.

@sovietdoc: You cannot compare an oced system to a system, which runs at default clock.

"@sovietdoc: You cannot compare an oced system to a system, which runs at default clock."
Of course you can. Wasn't it Hardware Analysis that overclocked a P4 to find out how fast the Presscott needed to be to beat an AMD FX-55? As I remember it, the verdict was that a 5.2GHz P4 was necessary to beat a stock 2.6GHz FX-55.

Ok, you can compare an overclocked system w/ a system clock@default, but it is not very intelligent, to say "socket M2 sucks, because ist is slower w/ standard clock than my overcloked system".

Yeah, but that wasn't the case. Intel said, "Our mainstream CPU at stock beat your enthusiast CPU overclocked."

-Pikl

Pickle, weren't both at stock?

Also, AM2 doesn't suck if it's slower than an overclocked processor. Obviously, overclocking the AM2 processors is the only way to go. There is no other choice. I must hit an 8GHz HT speed.