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

[Castiel]

Notice that they say i7 950 on the right, yet on the left they say package 1156?

Bulldozer is just BULLSHIT till some real results are released.

 

[bryanW1995]

1. All still estimated.
2. Still 8 core vs 4 core.
3. Price will be based on performance, so it will cost more than a 2600K.
4. No word on speeds yet.

Don't get me wrong, it is a huge step for AMD no question. But I do not think it is the "second coming" as some people tend to think.

if BD 8 core is strong enough to trump 2600k then that will be a win imho. remember, it was originally supposed to go up against nehalem, if it's ~ SB then that's going to bode very well for amd for a few years.

 

[Emulex]

so bulldozer needs to be lga-1156 standard or its chipset is going to nail it's coffin.

amd needs to make pin compatible cpu's again.

 

[IntelUser2000]

The 3 tests are all very well threaded, so it shows the advantages of 8 cores in the best light. Still, the results are really good.

Overall: Probably there's few % to gain in multi-threading just because Thuban was a hackjob to get 6 cores without expanding at all on the cache subsystem. It'll also gain from having Turbo Core that can be active when all cores are active. Currently, Thubans don't have that ability.

Cinebench: From the score, the gains in pre-AVX FP performance is pretty good. I wouldn't be surprised if there's 20% or so gain from the FP on Cinebench.

 

[maddie]

i7-950 130W TDP
1100T 125W TDP

the comparison is with that 2 processors

i know SB is at 95W and the HD2000 waste ~15W at full load... SB 4x is also much smaller than the ~310 mm2 Bulldozer

Is this really true? Where have you seen this?

 

[busydude]

Is this really true? Where have you seen this?

http://www.semiaccurate.com/forums/showpost.php?p=95274&postcount=55

Cannot vouch for accuracy though.

 

[JFAMD]

They are not half cores according to JF. They are 8 true cores.

And there is no way that an 8 core BD is @ 95W (which SB is at). Most likely 130W. So it can not be comparing at the same TDP. Now an i7 950 is at 130W, so that is more likely.

Why couldn't it be 95W? There are 95W client parts today.

There will be a 35W TDP 8-core server part and there will be a <95W TDP 16-core server part as well.

 

[Martimus]

so bulldozer needs to be lga-1156 standard or its chipset is going to nail it's coffin.

amd needs to make pin compatible cpu's again.

AMD has been using the same pin configuration since May 23, 2006, when they released the socket AM2 configuration. They have yet to release a different pin configuration in the following 5 years for consumer desktop processors.

Why would they release a chipset that uses a dead-end intel pin configuration? I don't understand the logic in that.

 

[JFAMD]

Pin compatibility is meaningless because the chipset will determine what you can plug into that socket.

 

[Cogman]

so bulldozer needs to be lga-1156 standard or its chipset is going to nail it's coffin.

amd needs to make pin compatible cpu's again.

Why? The last pin compatible AMD processor was the K6 back in 1997.

The problem with basing your standards off of another company is that you have no control over what they do with those standards.

 

[Cogman]

Pin compatibility is meaningless because the chipset will determine what you can plug into that socket.

That too. Even if AMD could magically work with intel's chipset (highly unlikely) they would pretty much have to completely redesign their chips to be compatible.

 

[out.of.order]

The 3 tests are all very well threaded, so it shows the advantages of 8 cores in the best light. Still, the results are really good.

Overall: Probably there's few &#37; to gain in multi-threading just because Thuban was a hackjob to get 6 cores without expanding at all on the cache subsystem. It'll also gain from having Turbo Core that can be active when all cores are active. Currently, Thubans don't have that ability.

Cinebench: From the score, the gains in pre-AVX FP performance is pretty good. I wouldn't be surprised if there's 20% or so gain from the FP on Cinebench.

yeah that's a AMD slide, i think they show pro-AMD benchmarks... that's marketing

Is this really true? Where have you seen this?

just my estimate, 300-330 mm2...

Why couldn't it be 95W? There are 95W client parts today.

There will be a 35W TDP 8-core server part and there will be a <95W TDP 16-core server part as well.

well... why compare a 95W CPU with two 125-130W cpus? btw... if that performances come from a 95W bulldozer you shoud change the name in "GODZiller" ^_^

 

[out.of.order]

Notice that they say i7 950 on the right, yet on the left they say package 1156?

maybe a typo...

Exactly. That is why comparing the 2 is like comparing apples to oranges.

i'm agree, 1155 is a mainstream platform that will compete with Llano, bulldozer should compete with the enthusiast socket of intel, maybe 2011 or 1356...

btw all depend on the price/performance ratio and availability

 

[sawtx]

well... why compare a 95W CPU with two 125-130W cpus? btw... if that performances come from a 95W bulldozer you shoud change the name in "GODZiller" ^_^

What about price, never even thought that the comparison being based on TDP. Newegg has the 1100T at $270 and the 950 at $295, so maybe this Bulldozer comes in that price range.

 

[AtenRa]

4 core SBs are 220mm2 vs a 4 Module 8 Threads BD at 320mm2 ??

what about 4 core SBs vs 2 Module 4 Threads BD at 200-230mm2 ??

 

[Castiel]

SB is more a 80W TDP Processor then a 95W. IGP is factored in to that number and with it disabled in a P67 chipset the number comes down.

 

[out.of.order]

4 core SBs are 220mm2 vs a 4 Module 8 Threads BD at 320mm2 ??

what about 4 core SBs vs 2 Module 4 Threads BD at 200-230mm2 ??

4 core SB without GPU are around 180 mm2

 

[Riek]

Notice that they say i7 950 on the right, yet on the left they say package 1156?

Because they do not match?

column it is listing the desktop platform for the intel and AMD for 2010. Next to it they do a test between an intel cpu and amd cpu that are comparable with each other in performance. But it is nowhere stated which platform the tested cpu has. that is something you make of it

 

[SickBeast]

If that leaked slide is accurate, AMD has a beast of a CPU on their hands that can only be viewed as an enormous success for the company. Just looking at that graph, it looks like SB will absolutely crush Intel's CPUs in gaming and rendering. I really hope it's accurate!

 

[reb0rn]

I don`t see anything spectacular about BD slide, it even look like PR crap...

 

[Arkadrel]

8 Cores going to be faster then 4 cores.. WOW who would have ever thought that? Not sure where they pulled that graph for gaming from since maybe 2 games take advantage of more then 4 cores. I want to see some Cinebench single threaded benchmarks.

Its a fake, why?

1) would AMD really write "Cinibench" (comments on the slide's bottom)?
2) "Industry's first and only native 8-core processor" - did they forget Nehalem-EX existed?
3) would AMD use words as "epic" in marketing slides?
4) use horizontal bars(which they didn't use in the past) instead of vertical bars?
5) would PR people would use such insanely drab colours in a marketing slide?
6) the box around the results wouldn't extend off the slide like that. It looks sloppy and unprofessional.

copy-pasted from SA, a few of the things that make people think its a fake.



Also JF-AMD is quoted for saying he wont tell any performance #s until launch day. I doubt this came from him or his people, its simply a fake some guy made in photoshop to score hits on his tech page or something.




That said Im looking forwards to the bulldozer, because of things like:

flexible floating point Unit (supposed to make it a FP monster)

Moduals where for 12&#37; extra die space they get a extra core, saveing on diespace/power useage (allowing them to put more cores into the cpu, at the same thermals/powerdraws as exsisting cpus).
--Performance/watt should improve drastically from this.

HyperTransport technology 3.1.
AES encoding acceleration

4 threaded workloads: (this one has me excited, and it can still do well on 8 threads)
"A single thread gets all of the front end, all of the FPU and all of the L2 cache if there is not a second thread on the module." (JF-AMD)

Turbo Core, makes use of "headroom" in watts, so it ll probably give differnt increases depending on how many cores are in use. Again this should help with stuff that isnt highly threaded. Ive heard upwards of atleast +500mhz (if this is for all 8 cores, then I imagine when only 1-4 cores are in use, it ll be much more).


I think bulldozer is promising ^-^

 

[ShadowVVL]

So is BD going to be 4 cores and 4 half cores or 8 cores and 8 half cores or just 8 cores?

 

[Edrick]

So is BD going to be 4 cores and 4 half cores or 8 cores and 8 half cores or just 8 cores?

Yes.

 

[Arkadrel]

So is BD going to be 4 cores and 4 half cores or 8 cores and 8 half cores or just 8 cores?

There wont be a differnce between the 2 cores in a modual,... so its not like 1 core will be differnt from the one next to it. And their both "real" cores, they simply share some stuff (that doesnt impact performance when your only useing 1 thread in a modual, and when your useing all your threads... well... do intel make 16 cores / thread cpus? any program that can use so many threads these processors will be extremly fast in)

besically this means the bulldozer will be really fast both in lightly threaded aplications, and in ones that can make use of many threads.

 

[ShadowVVL]

well i was just wondering if the 8 core BD is just like a quad core with hyper threading only using half cores instead of virtual threads? Or will it be 8 cores and 8 half cores where 1 and a half show as 1 core in the task manager?