Enlighten me, what's the difference between this, and just getting a multi cpu board and shoving some opterons on it? Obviously AMD will 'tweak' some things, or at least claim to.
There's an old saying, that goes "you can't polish a turd" the turd being the FX cpu's which are not going to match intel's core2's on raw speed.
I just can't figure out who 4x4 is aimed at...
It's aimed at AMD fans who refuse to switch to Intel no matter how superior Intel's offering is at the time. It allows AMD fans to claim to be running a Quad Core CPU and claim some level of eliteness.
Honestly.. 4x4 right now is sh!t. It will be good in time with the arrival of K8L, but as of right now.. it's total and utter crap. Useless crap, like a waste of hard earned cash.
So you're saying they've turned in to HeatBurst just because Core 2 is faster?
Right, AMD is naturally progressing the FX platform. They can't clock 1GHz higher so the thing to do is make FX dual socket. If the 80% reported at the demo is correct that means Intel will HAVE TO HAVE Kentsfield to keep the Extreme Gamer title.
X6800 would probably need to be clocked to 4.0GHz to beat out to FX70. FX74 has the potential to defeat kentsfield.
We all know FX IS NOT A JOKE CPU.
Just like nVidia could release a two card single slot GPU and be faster AMD can add a second socket to FX.
What's the problem. Should AMD lay down and die because Intel finally stopped melting cases and is really competitive with K8?
I don't think so.
Uh? What have you been smoking?
Where is this report of an 80% performance increase per clock over Core 2 Duo?
The initial 4x4, is nothing more then Two K8 Dual Core CPU's on a single motherboard. In heavy Multithreaded applications it will perform admirably yet noticeably slower then Kentsfield.
Under games and other apps it will lag behind a Core 2 Duo E6700 and maybe even E6600. It's all logical. No games out there, save for Quake 4, take advantage of more then two cores. Most that take advantage of Multiple Cores don't even get that large of an increase in performance. The proof is achieved when looking at how Kentsfield performs under the current crop of games.
4x4 K8 will be slower then Kentfield under all applications save for a few memory benchmarks and any applications (can't think of one though) that may be limited by Kentsfield's memory bandwidth. 4x4 K8 will be slower then Core 2 Duo under all games and the vast majority of applications.
So seriously.. why waste cash on something inferior?
I'm not saying AMD should roll over and die. I'm saying consumers should use there head when purchasing.
And also.. I think all AMD fanboys and/or Intel fanboys should roll over and die.
My 2 Cents.
Hopefully not what you've been smoking. I only compare AMD products to other AMD products so 4x4 is 80% faster THAN FX62.
Since 4x4 is SAID to be 80% faster in multithreading and allows you to shunt OS threads to two extra cores the beauty of it will be that it's much faster than FX62 and will carry twice as many active threads.
I would say that the 3.0GHz versions will get close to or pass Q6700 in certain FP applications like ScienceMark which means that engr's will be LOVING IT.
I'm sorry you hate AMD but that doesn't change the fact that if an FX62 system is $1700 and a 4x4 system is $2200 you are paying $500 for up to 80% more perf. That's a good deal. High end GPUs use up to 250W so I guess no one should buy those either.
Power Consumption
Watts
X1950XTX X1900XT 256MB X1950 Pro
Idle
152
151
144
Load
285
279
225
Power consumption with the 80nm die shrink is quite impressive. Under load the Radeon X1950 Pro manages to consume a mere 225 watts—54 watts less than the Radeon X1900XT 256MB. While the Radeon X1900XT 256MB delivers more pixel shading power, the Radeon X1950 Pro offers slightly better performance-per-watt in gaming.
Linkage for the weak.
Xeon Dempsey uses enough power to run a refrigerator nearly.
Linkage for the weak
965 EE loses in lots of benchmarks to X2 3800+ and probably uses 3X the power.
Where is your indignation about that? Where was your outrage as more HeatBurst SKUs were added?
Get a life. AMD has one.
UH? You sound like an even bigger idiot then before (if that's possible).
Why should I be outraged now about Netburst? It served it's purpose, and as far as i'm concerned died after the Pentium 4C. (yes idiot, Pentium 4C was an overall superior and more useful processor then the AthlonXP, Prescott was crap, I didn't buy one and I sure as hell have NEVER recommended one).
So 4x4 is up to 80% faster then an Athlon64 FX-62 according to you. Now could you kindly show us where you obtained this number?
Also I'd like to know under which applications. Now if it's 80% faster then an Athlon64 FX-62 when running Multithreaded applications, well no offense.. but umm NO SH!T! It's using two AMD Athlon64 FX level CPU's.. the same as having two Athlon64 FX-62 pretty much in SMP. So yes, it can now run 4 threads simultaneously and in applications sensitive to Multiple threads (like encoding) it could very well reach an 80% improvement.
But now, let's get something straight here. Per clock Core 2 Duo is 20-40% faster then AMD Athlon64 FX. Therefore if the 4x4 design gives you an 80% increase over a single AMD Athlon64 FX-62 then that would mean that Kentsfield would be around 100-120% faster then an AMD Athlon64 FX-62, thus still 20-40% faster per clock then 4x4.
Get my drift?
As for engineers loving a CPU that is faster under a single benchmarking program, I doubt it. Engineers use as many tweaks and tricks to get the best performance out of there software. Right now, Integer performance is more important then FPU performance. Even then, you're left with Core 2 Duo being faster per clock then Athlon64 FX-62 in sciencemark, therefore Kentsfield... you got it.. will be faster then 4x4 still.
Sisoft Memory Benchmark should be about just the only place 4x4 wins.
I don't hate AMD, I'm a realist.. and I don't take too kindly to liars.