Rumour: Bulldozer 50% Faster than Core i7 and Phenom II.

[jimbo75]

http://www.youtube.com/watch?v=yUWckPBv6Fs

There is also one of him running WEI, and he claims his 4 core is 23% faster than the 2600K, make of that what you will. He looks like a good AMD fanboy so might just be wishful thinking.

 

[Tuna-Fish]

http://www.youtube.com/watch?v=yUWckPBv6Fs

There is also one of him running WEI, and he claims his 4 core is 23% faster than the 2600K, make of that what you will. He looks like a good AMD fanboy so might just be wishful thinking.

FX4110 being 20% faster than 2600 in general tasks sounds far-fetched even to a hardened AMD lover like me. Nothing of the leaks shown so far even hints at that kind of performance.

(Note that based on what we know, it's entirely feasible that FX4110 will be 20%, or even much more, faster than 2600 when you pick the task carefully enough -- for example, a single-threaded program that is almost entirely SSE float adds. That should be 100% faster clock for clock, minus whatever you lose to the slower cache/memory subsystem.)

But if it is true, then it truly is a jackpot to AMD -- FX4110 should be much smaller in die size than 2600K.

 

[jimbo75]

Yeah, the WEI vid looks completely legit to me, although apparently you can hack the score so it doesn't change?

http://www.youtube.com/watch?v=6zK4WAPKNYo&feature=mfu_in_order&list=UL

 

[Zstream]

FX4110 being 20% faster than 2600 in general tasks sounds far-fetched even to a hardened AMD lover like me. Nothing of the leaks shown so far even hints at that kind of performance.

(Note that based on what we know, it's entirely feasible that FX4110 will be 20%, or even much more, faster than 2600 when you pick the task carefully enough -- for example, a single-threaded program that is almost entirely SSE float adds. That should be 100% faster clock for clock, minus whatever you lose to the slower cache/memory subsystem.)

But if it is true, then it truly is a jackpot to AMD -- FX4110 should be much smaller in die size than 2600K.

Nice find

 

[Voo]

Yeah, the WEI vid looks completely legit to me, although apparently you can hack the score so it doesn't change?

http://www.youtube.com/watch?v=6zK4WAPKNYo&feature=mfu_in_order&list=UL

Well obviously you could just reverse engineer what programs (that's published) with what command line arguments are called and how they return their values and then replace those with whatever you want, so sure it can be hacked like every other benchmark out there (only easier).

Make with that what you want - never really understood the appeal of cheats for that stuff.

 

[Zstream]

Well obviously you could just reverse engineer what programs (that's published) with what command line arguments are called and how they return their values and then replace those with whatever you want, so sure it can be hacked like every other benchmark out there (only easier).

Make with that what you want - never really understood the appeal of cheats for that stuff.

That is true but who knows... I think if it's real, it shows good progress and they must have some chips almost ready for release.

 

[Tuna-Fish]

That is true but who knows... I think if it's real, it shows good progress and they must have some chips almost ready for release.

By now, they should have warehouses full of these things if they want to have a hard launch next month.

 

[formulav8]

By now, they should have warehouses full of these things if they want to have a hard launch next month.

True. I also keep thinking that AMD really does have these things ready to go. They would have known months ago if there would be a problem releasing in June. Since its only a couple or so weeks away, it looks like there will be NO delay and thus an actual cpu release. :thumbsup: (I originally kept thinking it was going to be delayed).

 

[RyanGreener]

Yeah, I'm glad that info shows that there will be no delay. I'm dying to see this thing.

 

[Accord99]

I take it (the denial at the behest of the Intel fanboys) as proof positive that AMD probably does have a monster on their hands and they are about to conroe the market themselves.

Only another month or so to go!

The difference though is that a rough idea of Conroe's performance was known in March 2006, and numerous enough ES leaks provided a clear picture of Conroe's performance well before launch.

 

[mosox]

No way. AMD themselves placed the Bulldozer in between the i5 and the i7 in their roadmap pics. Intel should be better still and AMD should stick to their low cost decent CPUs policy (the Phenom II X4 955 is $120 right now - great bang for the buck) instead of trying to (re)gain the performance crown. Cheap quads, that's what they need to offer.

If there's only two car manufacturers in the world you can't have them Ferrari and Porsche. Somebody has to make cheap cars.

 

[podspi]

If there's only two car manufacturers in the world you can't have them Ferrari and Porsche. Somebody has to make cheap cars.

Or, both car companies could compete in the high and low-end market :sneaky:

 

[Voo]

If there's only two car manufacturers in the world you can't have them Ferrari and Porsche. Somebody has to make cheap cars.

Well I've seen worse car analogies but it suffers from the same fate as all those. But if we would want to use it, the ferrari under the CPUs is neither AMD nor Intel, but more IBM - low volume and quite expensive CPUs. And if we look at the other manufacterers they mostly have both higher and lower models for sale.

 

[Terzo]

Question:

If AMD hits a home run with Bulldozer how will pricing go? With the performance crown I'm sure AMD could charge a premium, but if Intel wanted couldn't the lower prices a lot and put AMD under a lot of pressure? That being under the assumption that Intel can afford lower profits to steal (or keep away) business from it's competitor, while AMD would have to balance making profit with not losing sales to Intel?

 

[Cerb]

Question:

If AMD hits a home run with Bulldozer how will pricing go? With the performance crown I'm sure AMD could charge a premium, but if Intel wanted couldn't the lower prices a lot and put AMD under a lot of pressure? That being under the assumption that Intel can afford lower profits to steal (or keep away) business from it's competitor, while AMD would have to balance making profit with not losing sales to Intel?

Yes, and it has happened, before (slot Athlon, Opteron, Athlon64 X2). AMD's advantage is that Intel loves their margins, and are always being looked at for antitrust violations, so Intel very rarely goes so far as a full on price war (they have been known to kick AMD where it hurts, in the low cost value arena).

 

[Terzo]

Yes, and it has happened, before (slot Athlon, Opteron, Athlon64 X2). AMD's advantage is that Intel loves their margins, and are always being looked at for antitrust violations, so Intel very rarely goes so far as a full on price war (they have been known to kick AMD where it hurts, in the low cost value arena).

:thumbsup: Thanks for the quick response.

 

[mosox]

And if we look at the other manufacterers they mostly have both higher and lower models for sale.

Yep but Intel doesn't have cheap quads and no triple cores while AMD doesn't have really high-end CPUs. I don't want two Intels (expensive quads, low socket life), somebody's got to offer $100 quads and long term socket compatibility.

 

[mosox]

This doesn't look real
http://shiftdelete.net/amd-bulldozer-benchmark-sonuclari-sizdirildi-29353.html

 

[itsmydamnation]

No way. AMD themselves placed the Bulldozer in between the i5 and the i7 in their roadmap pics. Intel should be better still and AMD should stick to their low cost decent CPUs policy (the Phenom II X4 955 is $120 right now - great bang for the buck) instead of trying to (re)gain the performance crown. Cheap quads, that's what they need to offer.

If there's only two car manufacturers in the world you can't have them Ferrari and Porsche. Somebody has to make cheap cars.

funny how people remember things the way they want to, not the way they actually are, ^_^.

in that picture the quad core is positioned against the lower end 2500 i assume that means the 2.7 model
the hex against the higher i5 2500
the oct against i7 2600


you have to remeber that this is Market positioning not product performance. Other things to consider:

when that graph was released did AMD have final silicon?
was AMD playing it safe in terms of expected performance?
was the graph even legit?

its pretty funny we still know almost nothing about performance.

 

[Abwx]

its pretty funny we still know almost nothing about performance.

Worse is that we dont even know about fake performances...

BD is already unique in respect of the timely overstreched secrecy
that make it some kind of ultra edition model..

True that it s the first genuine 21th century processor design..

 

[mosox]

I don't know if that's legit but if you look at the actual squares on the AMD side, the hexacore is still slower than the best i-5 and the 8-core slower than the i-7

 

[Cogman]

Worse is that we dont even know about fake performances...

BD is already unique in respect of the timely overstreched secrecy
that make it some kind of ultra edition model..

True that it s the first genuine 21th century processor design..

Or it may be the crappiest thing we've ever seen.

We have no idea how this processor performs, so why go and label it "The first genuine 21th century processor design."

 

[Abwx]

Or it may be the crappiest thing we've ever seen.

We have no idea how this processor performs, so why go and label it "The first genuine 21th century processor design."

You had surely noticed that there s no mention
of eventual performances in this sentence...:biggrin:

Only that it s the first processor design of this century..

 

[Tuna-Fish]

Or it may be the crappiest thing we've ever seen.

Absolutely. We have so little knowledge of it that pretty much anything is still possible.

We have no idea how this processor performs, so why go and label it "The first genuine 21th century processor design."

That has nothing to do with performance -- a genuinely new design can be utter crap, like P4 proved.

Most processors are very conservative evolutions of the ones that came before them. I think that BD is sufficiently different to be called a new design. Whether this is good or bad remains to be seen.

 

[dma0991]

I don't know if that's legit but if you look at the actual squares on the AMD side, the hexacore is still slower than the best i-5 and the 8-core slower than the i-7

I can't even see the word performance in that chart. The way that I interpret it, AMD stated that you can get a 8 core BD system for about the same price as a Core i7 2600K SB. There was no mention that the performance of a 8 core BD is equivalent to a Core i7 2600K.