AMD 4x4 - 15% faster than core in gaming..

[StyM]

source here

The Perspective Battle
So how will Core 2 Extreme stacks up against AMD’s 4x4 Platform? The answer is tough. Why tough? First AM2 has a speculation of being capable to do Reverse HyperThreading just by adding the instruction into the drivers. Even though users’ reports of Core2 series have discovered the multiplexing technique which is logically the same as in Reverse HT, 4x4 Platform will likely to beat Core2 into the second podium by a little margin. Theoretically when we take the average benchmarks between both the Athlon FX-62 and Core 2 Extreme have, an estimated 37-43% performance difference can be seen ( FX-62@2.8Ghz / Core2Extreme@2.93Ghz ) which is won by Core2Extreme. Let’s take an assumption if both chips have the ability to do Reverse HT. AMD’s 4x4 Platform will take its advantage by probably 50%( minus 5% of performance impact of HT Bus) extra performance since more threads can be executed in this platform compared with a single FX-62. Thus we can say that theoretically, AMD’s 4x4 Platform can outperform Core 2 Extreme by 10-15 percent in a condition where Reverse Hyper Threading is implemented.

[Major_A]

It's all over the place that reverse hyper threading is not real.

[cMw]

It's all over the place that reverse hyper threading is not real.

yet AMD announces it and they are going to come out again and say its fake or its not going to happen?

yea...ok because thats a great PR move..

if that the source said is true then AMD may hold onto its "best for gaming" title
(Intel hasn't released conroe yet so kinda hasn't taken anything away from AMD)

[[cTx]Philosophy]

Lets see how this pans out...
Ive had my opty rig for sale for a week now and noone was biting..
Might just have to keep er now and see how this 4X4 is gonna handle the conroe and kentsfield battle..
Im thinking theres gona be a war against amd and intel in the next few months to regain the crown, and its gonna be upgrade extravaganza if you have the loot..

Ima wait it out and see what happens..

I mean if amd can do this why cant intel smash together 2 conroes or kents and destroy the world :P

Im an amd fanboi forever but I think this is just a desperate attempt to steal some of Intels steam ATM... We shall see..

[DilTech]

Lets see how this pans out...
Ive had my opty rig for sale for a week now and noone was biting..
Might just have to keep er now and see how this 4X4 is gonna handle the conroe and kentsfield battle..
Im thinking theres gona be a war against amd and intel in the next few months to regain the crown, and its gonna be upgrade extravaganza if you have the loot..

Ima wait it out and see what happens..

I mean if amd can do this why cant intel smash together 2 conroes or kents and destroy the world :P

Im an amd fanboi forever but I think this is just a desperate attempt to steal some of Intels steam ATM... We shall see..

3 little letters... F. S. B.

Intel's front side bus would choke the poor little chips to death. Kentsfield already has to talk between the 2 dies amongst the FSB, now add another 2 dies(making 4 dies total for kentsfield) and have them all communicate via that 1 fsb.

What do YOU think would happen?

[Eastcoasthandle]

Wow only 15% INCREASE between 4 vs. 2...shouldn't that be theortically 25%?

[n00b 0f l337]

Hmmm amazing, so all there saying is it takes TWO DUAL CORES to beat a single intel DUAL CORE nowadays.
More heat output, more heatsinks, more case room needed.
I think this is more an anti-AMD thing.

AND NOTICE.

Its saying 2x fx62, thats how much money?

[BlackX]

Hmmm amazing, so all there saying is it takes TWO DUAL CORES to beat a single intel DUAL CORE nowadays.
More heat output, more heatsinks, more case room needed.
I think this is more an anti-AMD thing.

AND NOTICE.

Its saying 2x fx62, thats how much money?

2x FX-62 = 1600$
1x X6800 = 1300$

[n00b 0f l337]

Even if this is real, why is everyone happy that they should buy a SECOND cpu, a SECOND hsf/cooling system, deal with a SECOND power consumer, and buy a new mobo?

x6800 is only $1300 right now becuase of scalping, when fx-62 came out it was $1200!
So why not $2400 vs $1300. I can play the unfair play game as well.

[n00b 0f l337]

Well lets say you get dual lower speed dual cores, will it beat a conroe @ 4ghz of the same price?

[NoSinder]

http://www.theinquirer.net/default.aspx?article=33081

Good numbers, odd benches

By Charlie Demerjian: Monday 17 July 2006, 11:33
WELL, THE FIRST set of AMD 4x4 numbers are starting to trickle out, and interestingly enough, they come our way running on Vista.

The three benchmarks are all amenable to heavy threading, which probably explains why they chose Vista to test them.

Our source claims they saw an 80% average increase going from two to four cores, but exact testing methodology was not disclosed. Whether or not this carries over to current games is the hot topic of the moment, none of these benches were gaming benches.

That said, the preliminary numbers are encouraging. µ

......

[n00b 0f l337]

If the AMD people really start sticking to that ^^^ then they have no right to say SuperPi is unfair because it doesn't show real world gaming benches.

[n00b 0f l337]

Yeah speculation gets us knowhere.
But I hope that $300 for CPU's also includes $100 in HSF's for BT's. :P
And e6600 may be doing around 4ghz on air, so we'll see... Oh we'll see.

[Celeron Gamer]

I wonder how does the K8L yields, I hope it is be 65nm overclock crazy. If it remains on the same "lucky 3GHz OC" then AMD is in deep trouble

[Celeron Gamer]

I seriously doubt that, my expectations from MATURE 65nm process are about 4 GHZ on air. If amd is lucky their process will be mature by the time they start making K8L

I'll look forward to that, insane scaling AMD's

But Intel took AMD's efficient clock cycle idea, made it more efficient, and overclocks like a prescott, and shoves it back up their ass, so AMD better poop out a superchip

[b0bd0le]

where is AMD saying they have reverse HT coming out?

where is the undeniable, unrefutable evidence that AMD claims they have this techonology?

[gallag]

I seriously doubt that, my expectations from MATURE 65nm process are about 4 GHZ on air. If amd is lucky their process will be mature by the time they start making K8L

lol 4ghz lol

look at the history, PLEASE would someone post what ghz increases and power reductions the last 2 die shrinks have gained?

and can we stop comparing conroe to 4x4(which does sound pretty good) and start comparing it to kentsfield (which sounds very good)

[DoubleZero]

Sadly, AMD has delayed their 65nm process twice now. And that does not bode well to 1) the clocks they are getting out of the chips, 2) the yields, or 3) both.

Sadly, you don't have a clue of what you're talking about, but that's nothing new...

[oohms]

kentsfield will choke on it's fsb

Kentsfield benches here at XS have shown it not to be very fsb limited at all... it actually scales better from 2 to 4 cores than opterons do

[nn_step]

yet AMD announces it

Where have they announced it publically?

[K404]

Errr...everyone seems to have missed the obvious from the 1st post.

Theres SPECULATION as to performance increase based on SPECULATIVE scaling, SPECULATIVE efficiency, SPECULATIVE CPU use (FX or not) and..oh yea!..SPECULATIVE coding (R-HT)

speculation x speculation x speculation x speculation x speculation= BS!

Any of the guys here could string together the same number of possibilities? and get told to shut the hell up. Bad link...bad thread, sorry to say.

[gallag]

kentsfield will choke on it's fsb

13nm K8 chips ware about 60-70W at 2.4 Ghz, E6 chips ware about 30-35W, I don't mean the numbers amd says, but actual measured consumption. The average oc on 130nm was about 2.4 Ghz - went to 3 Ghz after 90n process got mature - thats 25% clock and 50% power gain (I talk about single core chips)

but lets stay on 4x4 shall we?

Mate, are you sure the AVERAGE oc for 90n is 3 GHz?

I guess I am asking why do you think a 25nm shrink to 65nm is going to yield a full 1 GHz headroom while a 40nm shrink from 130 to 90 only gained about 200-400mhz average?

You see, what you are doing is finding the absolute best numbers to support your wild clames and the absolute worst to support it while ignoring the actual average.

Just for clarification I am not a "Intel fan boy" and I laughed Hartley when people tried to say the pd was superior to the x2 but now every thing has just done a 180.

My thoughts

If 4x4 can use ANY x2 then it will be a winner
If it cannot and is limited to fx then it will be crap (but still blinding fast)

And what about Kentfield?
You say it will be choked by its fsb but that is just a statement pulled from netburst days, lets look at the facts

memron is the same as Conroe

memron runs on a 667fsb and could still put a fx62 in its place

So why should kentsfield be choked by its fsb when it will be double that of the very fast memron?

I honestly think that k8l quad core will be the winner not 4x4

[Ailleur]

speculation x speculation x speculation x speculation x speculation= BS!

Awesome!

When will you people get it in your thick heads that reverse hyper threading cannot bloody exist.
Its like wishing AMD's K8L will be the first quantum chip, ridiculous. Hell actually its not even like that, cause quantum chips have a chance of existing some day.

[Ailleur]

@Ailleur...you`re agreeing with me, right?

I was agreeing with your basic equation, the rest of my comment was directed at everybody else, sorry for the confusion

[d@rkn1ko]

with 4x4 you need 4 memory stick same as server motherboard?

2cpu
motherboard with 2 socket (more than 300$)
4 stick memory
psu ?
new case?
cooling ?

LOL