Bulldozers first screens

[dpa]

How does this compare to i7 2600k, for example? How high would it have to be clocked to get similar results?

Is the voltage info legit?

if this is true.. then Intel needs to get something fast..

my 4.6Ghz sandy 2600k does a wprime 32m at 6.25sek... 4/8 threads.

[Sparky]

While I can hope it is real, at the same time, I'm trying to be real myself and not gonna get too hyped over a single screenshot that may or may not be fake

I don't want to base anything I purchase on pre-release screens. I suggest others keep a cool head as well

[ichime]

So @ Wprime, it's faster than the 2600K @ 4.8ghz and a bit slower than the 990X at stock speeds, according to that screenshot. Not too shabby!

[Mats]

if this is true.. then Intel needs to get something fast..

That's called the 2800K, according to gigabyte_cpuz_readme.txt. Dunno if it will be enough tho.

I guess we'll see much more stuff like this in the next two weeks..

[w0mbat]

well, phenom II x6 did well on wprime too and superpi is easy to excellerate.

[Hornet331]

No spuprise there, wprime scales good with cores and mem speed/timings. But results look good.

[muziqaz]

I done wprime at 3.9ghz thuban - 7.251sec(32M)
wprime 1024M - 227secs
superpi 1M - 17secs

[BeepBeep2]

well, phenom II x6 did well on wprime too and superpi is easy to excellerate.

SuperPi is easy to accelerate?
Look at AMD's history with SuperPi from Athlon (XP) to Phenom II...


This is a 100% increase in efficiency compared to Phenom II in SuperPi

No spuprise there, wprime scales good with cores and mem speed/timings. But results look good.

RAM, uncore do nothing in wprime on AMD. Not sure about intel side but tuning ram for wprime doesnt gain much. ...not saying this arch cant be different

[muziqaz]

Yes AMD just upgraded x87 instructions. Got tired from being owned by Intel cpus

[stangracin3]

SuperPi is easy to accelerate?
Look at AMD's history with SuperPi from Athlon (XP) to Phenom II...


This is a 100% increase in efficiency compared to Phenom II in SuperPi


RAM, uncore do nothing in wprime on AMD. Not sure about intel side but tuning ram for wprime doesnt gain much. ...not saying this arch cant be different

i wouldn't mind knowing the NB, HT, Mem speeds, and timings though.

[muziqaz]

You can see HT 3.2hz

[cold2010]

SuperPi Mod v1.5



wPRIME 2.03

[generics_user]

this is completely insane if those results are true

1 ghz turbo is very impressive; additionally this is only a STOCK 95W chip beating sandybridge ocd to 4.5ghz?!

imagine what the 3.2ghz base frequency 125W chip does? i smell 4ghz+ turbo and performance higher than 990x on a significantly cheaper platform (cheaper socket + dual channel + weaker pwm + less layers)


first reviews will tell the truth about real world performance in application but this looks much better than the older leaks

[muziqaz]

this is completely insane if those results are true

1 ghz turbo is very impressive; additionally this is only a STOCK 95W chip beating sandybridge ocd to 4.5ghz?!

imagine what the 3.2ghz base frequency 125W chip does? i smell 4ghz+ turbo and performance higher than 990x on a significantly cheaper platform (cheaper socket + dual channel + weaker pwm + less layers)

and there goes not being too excited about this shot

[w0mbat]

i smell fake btw.

[muziqaz]

Also, it would be safe to assume that during those tests chip was running at 3.8ghz all the time or most of the time with all cores, so if it is the case, then those results are not that impressive for those who do not oveclock. But in that case who would stop you from setting the multi at 19 and keep it that way all the time

[muziqaz]

i smell fake btw.

yeah, when it is too good to be true, with amd it rarely is

[AKM]

I'm confident BD will turn out to be a good chip but I also don't trust anything coming out of Rumorpedia.

[generics_user]

and there goes not being too excited about this shot

you missed the most important part: "if true"

i don't really trust this screen, the cpu-z screen lacks cache info despite supporting BD according to the readme, Interlagos Support is very unlikely because it's supposed to come 1 quarter after zambezi?


maybe this screen paved the way for some other leaks

[jimbo75]

That's not even the top binned chip, according to the earlier slide.



No comment on the validity of the benchmarks. I can believe the WPrime ones but that SuperPi one looks too fast for me to believe.

[Olivon]

I call BS

[DarthShader]

Also, it would be safe to assume that during those tests chip was running at 3.8ghz all the time or most of the time with all cores, so if it is the case, then those results are not that impressive for those who do not oveclock.

It would be good to know how come there is 3.8ghz shown - is something still running?

If it was at 3.8 for the benchmarks, then the improvments in wprime could be only assosiated with the 33% cores increase. But if we remember that the turbo for all cores running supposed to be 500mhz and wprime most likely doesn' t stress the whole cpu enough to disable turbo, then it was running at 3.3ghz during those tests... and that would mean over 50% improvment.... for 33% more cores, just like promised! :O

[muziqaz]

Hm, I remember John(JF-AMD) mentioning those 500Mhz with all cores active. But he was talking about Interlagos 16cores beast. I would think desktop Zambezi chips might have more of that turbo

[w0mbat]

guys, if u look at this ss u can find a few obvious errors -> fake.

[[XC] hipno650]

I just want to point out that on the same page of this thread there are polar opposite screen shots shown about it's performance... while this single screen shot is exciting I hate to see people getting hyped up about it....

this happens almost every time a new product comes out. I remember the Cayman launch and everyone speculated and hyped it up a bunch and guess what... It's no faster than a GTX 570 card... it was right where you would logically expect it to be...

it is rare that we see a launch that goes against the grain of current performance upgrade trending. the last time it really happened was with the launch of Conroe back in the day... That was a MASSIVE shift... but most of the things before than and since then have been pretty boring overall.

I don't want to be a buzzkill or anything but one of my fondest product speculation memories was the screen shot that showed up showing an R600 GPU (2900xt) scoring 30k in 3DMark06.... that was pretty funny considering the chip was slower than a 8800GTS 640MB and overall total trash...

I'm not saying I expect the same-thing to happen with BD as that would be bad news for everybody but It would be nice if FOR ONCE people did not get all reved up over a product that is over a month away from being released.... and has yet to have an half legitimate information on performance released.