Bulldozers first screens

[BeepBeep2]

AMD's current 8-series boards support SATA 6GB/s and USB3 already, lol.

Only SATA 6. USB3 is provided on board usually by an NEC chip.

[mak1skav]

AMD's current 8-series boards support SATA 6GB/s and USB3 already, lol.

pfff i guess i was away from AMD platform for a very long time now and i have missed even the most obvious facts. I guess that's why i am very interesting about the new BD processors.

Is the provided SATA performance from 8-series boards in pair with ICH10 and P67 Express chipset and most important for me does anyone has any informations about the SATA performance of the new 9-series ?

I can say that the only thing i know is that nForce chipset wasn't very good at HDD benchmarks comparing to Intel chipset. Is the same thing happening with AMD chipset boards too?

I apologize if my questions look very stupid to you people but as i said i was away from AMD platform for a really long time and i am finding BD a very good reason to give them a try in the near future.

[[XC] Synthetickiller]

I had a shot to get a cheap SB chip in a trade, which fell through. That said, I'm looking towards AMD now.

I'll have to peruse this 36 page thread looking at actual benchmarks (I haven't seen any yet!). 16x/16x makes me happy. Doing SLI on dozer seems kick ass from every perspective.

Only question I have (which can't be answered) is what type of OCing are we looking at and is air enough or do we need water for max OC?

[Olivon]

AMD Bulldozer architecture by Franck Delattre - Hardware.fr

However, AMD's architecture has great assets in his pocket, and is able to make a nice career. The flexibility of the modules versions promises to more than 8 cores in the near future, and if the frequencies are at the rendezvous, the performance will be present, especially in multi-threaded processing.

Still, it will still be patient. If the first sample Zambezi begin to be distributed, the launch should not occur before July. It will take a little longer to judge the effectiveness of the new architecture into practice and see if AMD has won his bet, an appointment can not miss it!

Great article

[sniper_sung]

Am I late with this?



BTW since when was this thread moved to the AMD section? I thought it was in the news section

[Olivon]

Am I late with this?



BTW since when was this thread moved to the AMD section? I thought it was in the news section

Nice Vantage PhysX score

[Manicdan]

i was wondering how good 80k is, when phsyx is activated, lol

[sniper_sung]

i was wondering how good 80k is, when phsyx is activated, lol

For comparison, here is the Vantage CPU score of 2600K@5GHz with 560 Ti PhysX:



Now the only question is how trusty the leakage is

[Manicdan]

the 560 in the SB score is also overclocked,
so the BD one is somehow +30% more efficient when clock rates of the gpu are taken into account
(15% faster score, with a 15% slower gpu)

[EniGmA1987]

I'll have to peruse this 36 page thread looking at actual benchmarks (I haven't seen any yet!).

no legit benchmarks

[kaktus1907]

just compare cpu test 1.. physx only works at test 2.. my Q9650 @4.5ghz gets 2300 plans/s

and that image needs more pixels

[Olivon]

CPU-Z seems faky ...

Real legit is this one from the CPU-Z's author himself :



http://www.hardware.fr/articles/833-...bulldozer.html

[undone]

CPU-Z seems faky ...

Real legit is this one from the CPU-Z's author himself :



http://www.hardware.fr/articles/833-...bulldozer.html

Seems all of those CPU-z are not fake. Their instruction set are ordered in same arrangement.(sse4.1,4.2,4A,x86-64......) Maybe for different internal use.

Edit:

BTW socket name is AM3b, not AM3+



Edit 2:

Any necessity to reconsider this FAKE is fake or true?

http://www.rumorpedia.net/wp-content...dbulldozer.jpg

[RussC]

So which is it, fake or not? I guess if its legit, it beats a 5GHZ i7-2600k in those CPU tests?

RussC

[Manicdan]

nearly 6k planes is a little unreal. considering it would be at stock clocks and using up to 125W of power, while the SB is probably way past 125W to reach 5ghz, which then makes BD look insanely more efficient than SB

[sniper_sung]

So which is it, fake or not? I guess if its legit, it beats a 5GHZ i7-2600k in those CPU tests?

RussC

That's hard to tell. The file property of the original image has:



That needs some analysis if anyone is interested

[RussC]

Ya, those performance #s seem to good to be true! We all know what that means

RussC

[DarthShader]

That score seem to be inline with the with earlier leaks, isn't it? (Did we even have any vantage scores earlier?? I am getting confused in all this )

Interesting is, again we have 3.8GHz listed. Is this fixed turbo? Normal? Idle? Only 95W? the 8110 is indeed supposed to have 95W TDP, it's the 3rd model in line. So why 3.8Ghz??

That needs some analysis if anyone is interested

When you paste a screenshot into CS to save/resize it, it's going to have CS listed as orgin. End of analysis, or did I miss something?

[Mechanical Man]

So which is it, fake or not? I guess if its legit, it beats a 5GHZ i7-2600k in those CPU tests?

RussC

Why would it not beat it, it has double the cores. Amount of beat is something to consider. If it beats more than double... Well i can easily believe in under double performance at same clock, but over would be some serious .

[sniper_sung]

That score seem to be inline with the with earlier leaks, isn't it? (Did we even have any vantage scores earlier?? I am getting confused in all this )

Interesting is, again we have 3.8GHz listed. Is this fixed turbo? Normal? Idle? Only 95W? the 8110 is indeed supposed to have 95W TDP, it's the 3rd model in line. So why 3.8Ghz??


When you paste a screenshot into CS to save/resize it, it's going to have CS listed as orgin. End of analysis, or did I miss something?

I'm not familiar with AMD processors at all, so I'm not sure if that core voltage means overclocking...

I think there are algorithms to detect image processing by photoshop, but I'm not doing research in that area so I can't tell

[Manicdan]

Why would it not beat it, it has double the cores. Amount of beat is something to consider. If it beats more than double... Well i can easily believe in under double performance at same clock, but over would be some serious .

cause its 5ghz vs 3.8ghz

we dont know the actual speed anyway, turbo could be pushing it up much higher. but if a stock BD can beat an OCed 2600k, then i know i wont be able to afford it.

[Lightman]

cause its 5ghz vs 3.8ghz

we dont know the actual speed anyway, turbo could be pushing it up much higher. but if a stock BD can beat an OCed 2600k, then i know i wont be able to afford it.

You sir are right!

Maybe this is the reason why AMD can afford to launch BD on the desktop first. With high prices of CPU's they will be able to meet demand (low volume)!

Normally server always launched first for AMD because of higher profits and lower volume. This time it's the opposite! Who knows? Maybe FX is not coming back without a reason

[Particle]

BTW socket name is AM3b, not AM3+

The people who make the physical socket can mold anything into it they want. The official title for the socket is AM3+.

[Manicdan]

You sir are right!

Maybe this is the reason why AMD can afford to launch BD on the desktop first. With high prices of CPU's they will be able to meet demand (low volume)!

Normally server always launched first for AMD because of higher profits and lower volume. This time it's the opposite! Who knows? Maybe FX is not coming back without a reason

i dont think its physically possible for it to be that strong

5775 planes on 8 cores at 3.8 ghz, would be 3800 (almost exactly) with 4 cores and at 5ghz, which means that BD with 4 cores would be worth as much as the best sandy bridge, and the hexcore intel chips would also be raped, and this would make the first launch of BD be in the price range of 300 to 1500$

sounds very unlikely to me

[mattkosem]

Is the provided SATA performance from 8-series boards in pair with ICH10 and P67 Express chipset

I can't comment on P67 (which is a new and different beast) but, comparing the SATA performance of SB850 to ICH10, they're pretty level. Intel usually wins out on the randoms and AMD wins out on the sequentials. I believe ICH10 is limited on the top end to 600MB/sec on RAID, and I've seen over 1100 on SB850 with 4R0 SSDs. That 1100 seems to be a SB850 limitation, since I see over 1250 easily on an add-in card with the same drives.

--Matt