• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Bulldozer Finally!

:eek: :eek: :eek: :D

Source
Today is a good day as we have just received a great picture from Kirllos which was acquired by mega-d botnet.
Where is it from? According to email, it was borrowed from Gigabyte’s corporate network.
Machine is running AMD FX 8110 processor with the maximum TDP of 95W @ 2.8 Ghz and Socket AMD3b 942 package. If you are wondering why it says 3.8 GHz as well, it’s because o the AMD turbo boost.

According to the Gigabyte CPUZ History:
1.58.1gb – April 2011 update includes:
- AMD Bulldozer processor new instructions…
- AMD Zambezi, Valencia, Interlagos (32nm) support.
- AMD 990FX/RD990 + SB950 chipset.
- AMD AM3b/AMD+3 (black) socket detection.
- AMD Turbo Core TDP mode detection.
- NEW AMD logos / Tidy interface,
- DDR4 1.2v preliminary support.
- Add ID’s for future Intel 2650, 2700, 2800k editions.
- Several bugs fixed.
Click to expand...
Oh, it also has some benchmark numbers for you to check out.

amdbulldozer.jpg
 
Last edited:
HighlandeR said:
Fud are suggesting the 8 core fx desktop AMD chip will only just catch up with the 2600k
Click to expand...

If that turns out to be true, and if the FX 8100 is much more expensive than the i5, then I'm going to buy an i5 2500. Having seen benchmarks of the i5 and the i7 2600, there doesn't seem to be a massive difference in most of the tests and the i5 beats the i7 in some categories.
 
The fastest 95W Phenom II X6 is the Phenom II X6 1065T which runs at 2.9GHZ and can run half the cores at 3.4GHZ and downclocks the others to 800MHZ.

It seems that the 95W FX8110 runs at 2.8GHZ across all 8 cores. IIRC,second generation Turbo Core can run at upto 500MHZ across all cores. This would mean that the FX8110 is running 4 of the cores at 3.8GHZ in lightly threaded applications.

It will be interesting to see the clockspeeds for the 4 and 6 core versions.

Merlin5 said:
If that turns out to be true, and if the FX 8100 is much more expensive than the i5, then I'm going to buy an i5 2500. Having seen benchmarks of the i5 and the i7 2600, there doesn't seem to be a massive difference in most of the tests and the i5 beats the i7 in some categories.
Click to expand...

The i7 2600K is clocked higher than a i5 2500K and has an extra 2MB of L3 cache.

BTW,I would suggest you look at post 996.
 
Last edited:
zoomee said:
hmmm - 1.4v on a stock 32nm chip? - hope its true like but hey!
Click to expand...

Martini1991 said:
1.4v?!.. Jesus.. my 1055T at those clocks needs less.
Click to expand...


Voltage doesn't really matter, the 32nm process is pretty damn different to the 45nm one. THe key thing is power/temps/stability, at 1.4v a 1055t won't be using 95W, the main reason not to put 1.6v through a 45nm Phenom 2, is simply ability to cool the thing. with 95W at 1.4v, thats still a lot of headroom for power/temps/cooling capacity.

Also don't forget that the one thing CPU-Z will frequently get wrong is accurate voltage readings on new/unsupported mobo's. Things like spd settings on mem, cpu clock speed come from the same source on every mobo, things like chip voltage and the like are read from mobo monitoring chip which vary, hence the one thing that tends to be wrong(if anything) in CPu-Z.


Merlin5 said:
If that turns out to be true, and if the FX 8100 is much more expensive than the i5, then I'm going to buy an i5 2500. Having seen benchmarks of the i5 and the i7 2600, there doesn't seem to be a massive difference in most of the tests and the i5 beats the i7 in some categories.
Click to expand...

Well, the main thing to take from the screenshot, if its real, is sub 8 seconds on superpi at 3.8Ghz is completely immense. Superpi is one of the few things that can use the full width of a single i7 core, its kind of best case scenario. My system isn't set up for ultimate speed but right now my P2 quad at 3.84Ghz gives 17.5seconds on superpi, if a Zambezi is down to under 8 seconds, thats a ridiculous, massive increase in throughput and speed. matching Intel in Intel's best case scenario benchmark, you literally can't overstate how MASSIVE that is.

Also, the cheapest Bulldozer octo core will cost less than a 2600K, probably below a 2500k, while the highest bin might well be closer to a 2600k.

I've been quietly optimistic for Bulldozer but never in a million years would have expected even a sub 10 second superpi. Phenom is a 3 issue core(but a slow one) i7 is a 4 issue core, Bulldozer is a 2 issue core, but with MASSIVE improvements so it should be much closer to an i7 than a Phenom 2 is. Bulldozer isn't a well hidden secret that its designed to be a pretty awesome multicore chip.

People incorrectly took that to mean they didn't care about single thread performance, that was rubbish it just meant AMD wouldn't increase core size 20% to increase single thread performance 5%, its a waste. It was focusing on efficiency, but fast single thread performance IS efficient. But that is why I was expecting good but not epic single thread performance increases, that Super pi increase, if believed, is in the range of best ever increase in performance type EPIC.
 
I just noticed the scores!! :eek:

The wPrime scores are just 10% slower than a Core i7 980X:

http://www.bit-tech.net/hardware/cpus/2011/01/03/intel-sandy-bridge-review/7

A 95W 2.8GHZ 8 core FX8110 is just 10% slower than a 3.33GHZ 6 core 980X capable of running 12 threads!

The SuperPi scores look faster than what a Core i7 2600K running at 4GHZ can achieve!!

Of course the screenshot could be fake but if true then the FX8110 looks a very fast processor.
 
Last edited:
CAT-THE-FIFTH said:
I just noticed the scores!! :eek:

The wPrime scores are just 10% slower than a Core i7 980X:

http://www.bit-tech.net/hardware/cpus/2011/01/03/intel-sandy-bridge-review/7

A 95W 2.8GHZ 8 core FX8110 is just 10% slower than a 3.33GHZ 6 core 980X capable of running 12 threads!

The SuperPi scores look faster than what a Core i7 2600K running at 4GHZ can achieve!!

Of course the screenshot could be fake but if true then the FX8110 looks a very fast processor.
Click to expand...

oliver2601 said:
We haven't really seen pricing yet. Sure, it could be faster than a 2500k... but if it's £80 more would it really be worth it?
Click to expand...

Just to quantify both things, firstly its almost certainly running 3.8Ghz for turbo, this thing is going to be around 1Ghz Turbo and unlike Intel that will be with all the cores running, not just one core.

Second, it won't cost £80 more than a 2500k, it looks very likely the top bin will be over 3Ghz normal and 4Ghz or higher with Turbo. Even if not, get the lower bin and clock it up.

AMD need market share more than the best ever margin per chip, price competitiveness has been AMD's game for god knows how long.

Its very possible the top Bulldozer could and really should cost as much or even more than a 2600k but by earlier reports it would seem they'll be pricing it below a 2600K.

Just keep in mind, Superpi is where, theoretically you'd expect a 2 issue core architecture to be at its WEAKEST, not strongest ;)

Of course, maybe there is some very fancy mojo going on and it can run a single Superpi instance over a single module(but two cores inside that module).

remember the design of bulldozer, adding the second core to each module cost 5-10% die space. They didn't take a quad core and double its size, the point was to make a octo core, out of essentially what used to be a quad core, take efficiency( in size) through the roof vs their older Phenom 2 and be insanely competitive again. Pricing it out of affordability for most people would be completely and utterly counterproductive.
 
That SuperPi time seems too good to be true.
If real then great :)

Enthusiasts will buy then anyways myself included, even if the performance isn't as stellar as that leaked screenie indicates.

The Crosshair is so pretty i just have to buy the V.
NV support for SLI on the 9x chipsets is a nice bonus, no more SLI hack.

If somehow ASUS could persuade NV to let them include the SLI token in the CH 4 bios it would please a lot of folks
Somehow i doubt that though.

Sure i just heard my 2600K say "bring it on dude" :)
 
You must log in or register to reply here.
Back
Top Bottom