AMD Bulldozer Details Leaked From Gigabyte

Status
Not open for further replies.
How fast is that super pi time compared to intels offerings?

Edit: For what its worth the i5 in my laptop took about 17 seconds for 1m
 
However much I want that to be true, I am not holding my breath until bulldozer actually launches or we get benchmarks from [H], Anand or somewhere else fairly credible.
 
Last edited:
lol, fake

Someone doesn't understand that SuperPi is a single threaded application.
 
bastage said:
Right so it must be fake.. because the CPU cant be that fast @ 3.8 Ghz..
Click to expand...
Yes, I'm glad you also notice that it's faked. ;)

A clumsy fake deserves to be laughed at. There are a few other problems with the data in the screen shot. Try to find them.
 
How the heck do you get a higher rez version? This one is useless.
 
So even if some stuff is photoshopped that means its all fake?? Maybe they wanted to get other suff in the pic that otherwise wouldnt work???I dont know but some of you guys maybe to quick to gudge. Who knows this could all be true or not....I call plausible
 
MACK-ATTACK said:
So even if some stuff is photoshopped that means its all fake?? Maybe they wanted to get other suff in the pic that otherwise wouldnt work???I dont know but some of you guys maybe to quick to gudge. Who knows this could all be true or not....I call plausible
Click to expand...

It is called trolling. These Intel trolls can't stop the Bulldozer!
 
Bulldozer @ 3.8 faster than Sandy Bridge @ 5.0 in a single-threaded application? I'll believe it when the final product is delivered and reviewed by several sites. That would be a ridiculous improvement on AMD's part, and they have only been claiming 50% more performance with 33% more cores themselves. Pretty sure this is fake.
 
While I agree it's probably fake, I would not at all be surprised if BD fpu single thread performance bested intel offerings. It is a fat fpu shared between two cores after all.

It would be more surprising if sandy's single thread FPU was faster.
 
Last edited:
sirmo said:
While I agree it's probably fake, I would not at all be surprised if BD fpu single thread performance bested intel offerings. It is a fat fpu shared between two cores after all.

It would be more surprising if sandy's single thread FPU was faster.
Click to expand...

They're 256-bit wide, right? If so, they're no bigger than Sandy Bridge's. The bigger advantage Bulldozer has is its "extra" 4 threads will in theory be executed faster than Sandy Bridge's since they're actual cores where SB is just hyperthreaded. Multi-threaded performance on Bulldozer beyond 4 cores will probably be pretty awesome, but 4 and under I'm willing to put my money on Sandy Bridge.
 
LAiN_OfTheWired said:
They're 256-bit wide, right? If so, they're no bigger than Sandy Bridge's. The bigger advantage Bulldozer has is its "extra" 4 threads will in theory be executed faster than Sandy Bridge's since they're actual cores where SB is just hyperthreaded. Multi-threaded performance on Bulldozer beyond 4 cores will probably be pretty awesome, but 4 and under I'm willing to put my money on Sandy Bridge.
Click to expand...

You can't tell who's got bigger FPUs purely on the width of the instructions they can take. BD shares a single FPU per 2 cores so it's only natural they have the transistor/power budget to make them beefy. Hammer's (2003) FPU was quite a bit ahead of Intel at its release.

Checkout the wetstone benches last time AMD released a new architecture : http://www.tomshardware.com/reviews/duel-titans,620-20.html
 
Last edited:
I understand Superpi also benefits from lots of cache; if they compare single threads on BD chips with 16mb versus SB with (6? 8?), that could make a substantial difference.

And... are we sure that non-AVX instructions can't be distributed across both FP "halves"? Assuming that superpi isn't already utilizing 256-bit instructions.
 
sirmo said:
Checkout the wetstone benches last time AMD released a new architecture : http://www.tomshardware.com/reviews/duel-titans,620-20.html
Click to expand...
*Whetstone

http://www.youtube.com/watch?v=6vQpW9XRiyM :p That was a few competitive architectures ago, innit?

Anyways, the link brings up an interesting point. SuperPI was compiled a long time ago. It doesn't have any tweaks to benefit modern processors, which is why it's often criticized as a poor, inefficient benchmark. The gcc *compiler* tests done at THG have a purpose: to show improvements on code compiled for the new architecture (Whetstone is not a very good benchmark... Spec tests are better). That is not the same as showing how legacy code performs, and is not automatic as this fake SuperPI score tries to imply. It could have been more believable if the hoaxer wasn't so ignorant of the benchmark. ;)
 
7 years in making.. possible knowledge transfer from ATI and all pxc can believe is BD will be a die shrink of Phenom 2. Why so negative?

AMD hasn't had a significant new architecture since the days of Pentium 4, since the time they were on top.
 
pxc said:
Keep open flames away from that straw man. :D

"I want to believe (a ridiculous fake is real)"
Click to expand...

man your posts have already shown that your an obvious intel fanboy so why are you still posting here.. if you don't want to believe its real then don't.. But everything on the interwebz points to you being wrong as well as AMD's direct comments about bulldozer being new architecture.. So go troll elsewhere.
 
bastage said:
But everything on the interwebz points to you being wrong as well as AMD's direct comments about bulldozer being new architecture..
Click to expand...
Bizarre that you got that from my posts, but another excellent example of a straw man. <citizenkaneclapping.gif>

I'm glad that BD is rumored to be introduced in about a month. The cold splash of reality will be refreshing. :)
 
come on guys, he is a skeptic ;) just read his quotes

needless to say, AMD would be stupid to have just a die shrink of an architecture that is already behind....but then again as was pointed out, everything points towards a new architecture overall..
 
if it is true great for amd!!! that is at a 1ghz oc from the stock speed according to cpuz. but i thought i remember reading they were aiming for 3.6-8 for these chips?
 
I'd say scepticism is well advised in case of AMD technology ;)

Still I'd prefer if optimistic scenario happened:
1. Because I'd rather put new well performing Bulldozer in my unofficially hopefully compatible AM3 mobo
2. If mobo turns out not compatible I would be able to buy Sandy Bridge cheaper due to competition ;)
 
tricky0502 said:
if it is true great for amd!!! that is at a 1ghz oc from the stock speed according to cpuz. but i thought i remember reading they were aiming for 3.6-8 for these chips?
Click to expand...

That is supposed to be their new advanced turbo. Low base clocks (at least for the 8 core) with a high turbo. There has been no indication on how many cores the turbo will operate however.
 
I am really getting sick of all the this person is a troll because they are skeptical of a unofficial benchmark posted on some random site on the internet.. It's really annoying and the also main reason why I left AMDZone (after being a member for 5 to 8 years). Is discussion or opposing views on this forum also not permitted?
 
I don't really see name calling as a problem. It's silly and immature, yes. It actually makes me laugh a bit because I buy and recommend both major x86 processor manufacturers -- yet I'm called the "fanboy" by users who only buy from one (second place performance) manufacturer. Great stuff. :D

The problem I have is when people misrepresent the posts. Like the idea that I haven't recognized BD as a new architecture (I was given grief years ago for pointing out that AMD's new roadmap pushed the new architecture back, and how the old K10 was being stretched in the meantime), or that I'm somehow against BD (I've called the BD architecture interesting several times right here in this sub-forum).

However, I do feel that on the desktop, for typical users wanting a high performance processor for general gaming/internet/non-workstation use, translating 8 core server performance is not good in the face of its competition. Even JF's rosy (vague) marketing projections don't change that. Anyways, BD release isn't too far away anymore and we'll see how it does instead of drool over or laugh at these lame fake screenshots. :p
 
At this point, do you guys honestly think the Bulldozer will be faster then the CPU that was announced last week? IIRC, it'll be released this year.
 
MACK-ATTACK said:
So even if some stuff is photoshopped that means its all fake?? Maybe they wanted to get other suff in the pic that otherwise wouldnt work???I dont know but some of you guys maybe to quick to gudge. Who knows this could all be true or not....I call plausible
Click to expand...
EVERYTHING is plausible. Until proven guilty. :D
 
jonneymendoza said:
what cpu is that?
Click to expand...

I was off though, it's "early 2012".

"The first products to hit the market using this new manufacturing technique will be Intel's "Ivy Bridge" line, the successor line to the current "Sandy Bridge" line. This means the new technology will first appear in laptops, desktops, and servers that use Intel's chips. The faster switching speeds, lower voltage operation, and lower leakage should make Ivy Bridge processors considerably more energy-efficient than the Sandy Bridge CPUs in systems today. These products are expected to hit the market in early 2012."

http://www.pcworld.com/article/227260/intels_3d_transistor_why_it_matters.html
 
I have a real hard time believing that AMD BD Turbo mode will go from 2.8-3.8 Ghz or 3.0-4.0 Ghz on desktops. (It would most likely have to jump its CPU voltage from .150V-.200V simply for those speeds stable. :( Because that speed/voltage would generate alot of additional heat for a stock heat sink to deal with. I could believe it if their heat sink cooler was designed like a Noctua NH-D14 or is watercooled. :p

I doubt those Turbo speeds, and also if you look @ current CPU's from both Intel and AMD for example. 2600k Turbo goes from 3.4Ghz to 3.8 Ghz, simply because the voltage required on a CPU to vary 400 mhz stable is negligible with 45 or 32nm tech. The 1100T also has a Turbo boost from 3.3Ghz to 3.7 only 300Mhz.

Actually though with AMD Bulldozer Module chip design, what if it's so INTELLIGENTLY designed that on an 8 Core FX that it drops half the module to less than idle) which equals cooler temps) and CAN actually run that one part of the 4 modules @ such a HIGH voltage / Clock speed. Rendering it a Quad Core with insane stock OCing? Because it is only heating up half of the chip it normaly does? for this stable speed increase?! For single threaded and to up to 4 core scenarios / benchmarks I could see this competing with SB and winning.

Then when running a multi-threaded program it intelligently returns back to stock speeds with the traditional old-school version of a 300-400Mhz OC with all 8 cores running for encoding for example or programs that DO use all 8 cores?!!!? So they would have like 3.1Ghz or 3.3 Ghz 8 core multi threading with an insane 1 Ghz OC on the 4 core / half module Turbo boost. :eek: Will there be 2 sets of Turbo Boost for BD FX?

Because of its NEW AMD DESIGN!! I have a hard time not believing this now is completely possible! :D ;)

Cue the Intel Fan-boys :D
 
Last edited:
I'd love to see AMD come from behind with an awesome design, competition is great. Problem is, we've not heard a peep from them.
 
Status
Not open for further replies.
Back
Top