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

[Dribble]

It's not just the cpu - the increased power draw on the motherboard strains the power regulation circuitry and can cause that to die young, particularly if you are using a cheaper motherboard with lower quality capacitors on it.

 

[garagisti]

http://www.xtremesystems.org/forums...-info-fans-!&p=4898473&viewfull=1#post4898473

3.7Ghz stock for an 8-core chippery... me likey

 

[videoclone]

yeah, but if 1 Mhz on an i7-2600 = 1.2 Mhz on a FX8130 then it will need to be clocked allot more then 3.7Ghz

 

[Imouto]

Nop doubling the core count.

 

[Mopetar]

They've stated that it will be able to turbo boost 500 MHz with all cores, so many people are assuming that when it's only using one or two cores, that it will be able to boost much higher.

If the stock clock rate truly is 3.7 GHz, the single core boost could conceivably take it as high as 4.7 GHz. If the chip could sustain that amount boost almost indefinitely, the single-threaded performance should be in the same ballpark as what Intel can offer.

 

[bridito]

But again, if this is for business, I wouldn't do it. Just because chances are pretty high that you really don't need a faster CPU. So going through the whole rigmarole of overclocking and testing is quite worthless.

It's definitely for business but I do need a faster CPU because... I'm a speed freak! Actually all I need is more RAM, but I could easily make the trusty old i7 940 keep going for at least another year or two by changing the 2GB modules for 4GBs. At 24GB I'd likely be able to take on anything the next couple of years could throw at me!

I have sort of done that. The problem is that they become obsolete in 3-5 years, and the last CPU I had die was an AMD K7 XP2400.

If they last 5 years, thats all I care about.

I don't think I've ever owned a PC for five years if we don't count leaving them in the garage to be a nest for mice!

 

[Mem]

I saw this link in their CPU forums posted over@OCUk http://hwbot.org/forum/showthread.php?t=28172

 

[AtenRa]

http://translate.google.com/transla...ldozer-FX-islemcisi-icin-test-sonuclari_1.htm

 

[badb0y]

Quick someone give me a rundown, how does it look!?

 

[soccerballtux]

that doesn't look great? 4.2ghz? Or perhaps only 3.2ghz and ~20seconds 1M calc.
My 3.5ghz does it in 19.680s.

 

[soccerballtux]

Quick someone give me a rundown, how does it look!?

looks bad to me. It'll be great at multithreading though, for the die area. Single threading performance not much better...

 

[yottabit]

looks bad to me. It'll be great at multithreading though, for the die area. Single threading performance not much better...

Yeah, I'm not sure I would trust the results from that second linked translated page. The HW bot stuff looks legit to me, at least believable, but there's no performance numbers. What he says makes sense about the gimped L3 cache and I imagine the Superpi times (19 secs) would be with that model and not the fully functioning one. If not it's totally fubared, as I'm pretty sure my stock i5 750 will wreck that. That's about in line with a ~3 Ghz Core 2.

Makes no sense, because stars cores are ~1:1 with Core 2 and at 4.2+ Ghz Turbo the BD would have to have way worse IPC than stars. This makes me think it's a gimped ES, especially combined with what OBR said (as incredulous as he may be)

I think we are still waiting for real performance numbers

Is it really a 186W TDP or is that a glitch? That almost necessitates water cooling IMO if you want any real thermal headroom

I'm honestly much more excited and hopeful for BD than I was before. You all can think whatever you want about IPC, but I've always assumed it will have better IPC than stars or AMD wouldn't release it. That being said if you can clock it at 4.4 Ghz on stock volts and the IPC is somewhere in between Core 2 and SB (which I think is very likely) I think they'll have a real winner on their hands for content creation

IMO it's really going to be a content creation processor though. I don't see why people are getting so excited about it for gaming. SB is and will be a much more well rounded gaming platform simple because it's not so many core and more focused on single core performance, and that should've been clear from the beginning

It's funny, I almost feel like BD would be perfect for Macs. They have crappy gaming anyway, and with the Mac Pro line they've always been about multithreading. Hmm.. Mac Pro is due for refresh...

 

[Arzachel]

Those benchmarks are... less than convincing. Not really worth making a fuss over.

 

[GaiaHunter]

http://translate.google.com/transla...ldozer-FX-islemcisi-icin-test-sonuclari_1.htm

that doesn't look great? 4.2ghz? Or perhaps only 3.2ghz and ~20seconds 1M calc.
My 3.5ghz does it in 19.680s.

looks bad to me. It'll be great at multithreading though, for the die area. Single threading performance not much better...

That one has been marked as fake since Bulldozer is only supported by CPU-Z 1.58 and the one in atenra link is using 1.57.1.

The one linked by Mem shows 1.58.

EDIT: Scratch that - Apparently CPU-Z can show BD on both 1.57 and 1.57.1.

 

[soccerballtux]

Yeah, I'm not sure I would trust the results from that second linked translated page. The HW bot stuff looks legit to me, at least believable, but there's no performance numbers. What he says makes sense about the gimped L3 cache and I imagine the Superpi times (19 secs) would be with that model and not the fully functioning one. If not it's totally fubared, as I'm pretty sure my stock i5 750 will wreck that. That's about in line with a ~3 Ghz Core 2.

Makes no sense, because stars cores are ~1:1 with Core 2 and at 4.2+ Ghz Turbo the BD would have to have way worse IPC than stars. This makes me think it's a gimped ES, especially combined with what OBR said (as incredulous as he may be)

I think we are still waiting for real performance numbers

Is it really a 186W TDP or is that a glitch? That almost necessitates water cooling IMO if you want any real thermal headroom

I'm honestly much more excited and hopeful for BD than I was before. You all can think whatever you want about IPC, but I've always assumed it will have better IPC than stars or AMD wouldn't release it. That being said if you can clock it at 4.4 Ghz on stock volts and the IPC is somewhere in between Core 2 and SB (which I think is very likely) I think they'll have a real winner on their hands for content creation

IMO it's really going to be a content creation processor though. I don't see why people are getting so excited about it for gaming. SB is and will be a much more well rounded gaming platform simple because it's not so many core and more focused on single core performance, and that should've been clear from the beginning

It's funny, I almost feel like BD would be perfect for Macs. They have crappy gaming anyway, and with the Mac Pro line they've always been about multithreading. Hmm.. Mac Pro is due for refresh...

I mean I would think with this 256 wide FPU the SuperPi should BLOW everything else out of the water. But IDK.

 

[soccerballtux]

That one has been marked as fake since Bulldozer is only supported by CPU-Z 1.58 and the one in atenra link is using 1.57.1.

The one linked by Mem shows 1.58.

good

 

[Schmide]

http://translate.google.com/transla...ldozer-FX-islemcisi-icin-test-sonuclari_1.htm

Considering the L2 read speed is less than memory, something is gimped.

 

[bridito]

Considering the L2 read speed is less than memory, something is gimped.

There are more fake BD benchys out there than there are fake breasts at a Hollywood cocktail party!

If next week's Dallas BD showing turns out to be a big fat nuthin', I'm going to officially abandon hope!

 

[Schmide]

There are more fake BD benchys out there than there are fake breasts at a Hollywood cocktail party!

If next week's Dallas BD showing turns out to be a big fat nuthin', I'm going to officially abandon hope!

Huh? The fact that you can find a factor that can explain the performance gives it some credibility. Why would you fake crappy performance???

 

[skipsneeky2]

LOL this thread is so damn long and not a single piece of physical proof if amd has their head up their butt for the warmth or realty stating facts about true performance.

Casey anthony wants a baby lets give her a baby in prison and leave the bulldozer thread alone till she is knocked up then we could argue some more about technology that is not here yet or wont be till the second coming of jesus.

Which ever comes first ill welcome it

 

[bridito]

Huh? The fact that you can find a factor that can explain the performance gives it some credibility. Why would you fake crappy performance???

I was referring to the various opinions that it's fakes. And I assure you that there are more than enough fanbois on either side who are staying up nights trying to figure out ways to fake benchys on their rival's CPUs!

 

[bridito]

LOL this thread is so damn long and not a single piece of physical proof if amd has their head up their butt for the warmth or realty stating facts about true performance.

Casey anthony wants a baby lets give her a baby in prison and leave the bulldozer thread alone till she is knocked up then we could argue some more about technology that is not here yet or wont be till the second coming of jesus.

Which ever comes first ill welcome it

Can I volunteer to get my hands on Casey Anthony first? (Note to the FBI, this is idle forum chatter and I am not now nor ever have advocated violence against lying ALLEGED baby killing scumbags as the Constitutional power of a jury is never to be disputed even when they have a case of rectumhead which is unparalleled even by certain CPU manufacturer executives.)

Anyway, back to the topic... :thumbsup:

IMHO: There is no evidence to show that BD outperforms 2600K or underperforms a Pentium II. None. Everything that's come out so far is nothing more than fanboi fantasies. I just pray that when Jesus with the Mayans and the rest of the prophets return on Dec. 21 2012 at least we would have had a chance to see a real BD benchy!

 

[Schmide]

I was referring to the various opinions that it's fakes. And I assure you that there are more than enough fanbois on either side who are staying up nights trying to figure out ways to fake benchys on their rival's CPUs!

So stop crying fanboi and talk about the figures! We are here to discuss the details not declare a conspiracy.

 

[dma0991]

There is no evidence to show that BD outperforms 2600K or underperforms a Pentium II

You wouldn't want to release a product that can't even beat your competitor in at least a single test. If it doesn't outperform your competition in any test at all then it is back to the drawing board which is what the latest stepping is all about.

I don't have to be a sales person to tell you that. A product that doesn't compete in anything at all won't sell obviously. So when it comes to BD, I am looking forward that it kicks ass in multi threaded applications but not so in single threaded. If BD were to suck at both MT and ST, then there is no point of selling it.

 

[skipsneeky2]

It would only sell dma If it was priced at the same price points as their current line up while outperforming it.

But even then it would be a useless product at best it would currently just close the performance gap at the same price point which it seems amd is good at with their higher end products.