Monday, September 27th 2010
AMD Radeon HD 6700 Series ''Barts'' Specs Sheet Surfaces
Here is the slide we've been waiting for, the specs sheet of AMD's next-generation Radeon HD 6700 series GPUs, based on a new, radically redesigned core, codenamed "Barts". The XT variant denotes Radeon HD 6770, and Pro denotes HD 6750. AMD claims that the HD 6700 series will pack "Twice the Horsepower", over previous generation HD 5700 series. Compared to the "Juniper" die that went into making the Radeon HD 5700 series, Barts features twice the memory bandwidth thanks to its 256-bit wide high-speed memory interface, key components such as the SIMD arrays split into two blocks (like on Cypress), and we're now getting to learn that it uses a more efficient 4-D stream processor design. There are 1280 stream processors available to the HD 6770 (Barts XT), and 1120 stream processors to the HD 6750 (Barts Pro). Both SKUs use the full 256-bit memory bus width.
The most interesting specification here is the shader compute power. Barts XT churns out 2.3 TFLOP/s with 1280 stream processors, GPU clocked at 900 MHz, while the Radeon HD 5870 manages 2.72 TFLOP/s with 1600 stream processors, 850 MHz. So indeed the redesigned SIMD core is working its magic. Z/Stencil performance also shot up more than 100% over the Radeon HD 5700 series. Both the HD 6770 and HD 6750 will be equipped with 5 GT/s memory chips, at least on the reference-design cards, which are technically capable of running at 1250 MHz (5 GHz effective), though are clocked at 1050 MHz (4.20 GHz effective) on HD 6770, and 1000 MHz (4 GHz effective) on HD 6750. Although these design changes will inevitably result in a larger die compared to Juniper, it could still be smaller than Cypress, and hence, more energy-efficient.
Source:
PCinLife
The most interesting specification here is the shader compute power. Barts XT churns out 2.3 TFLOP/s with 1280 stream processors, GPU clocked at 900 MHz, while the Radeon HD 5870 manages 2.72 TFLOP/s with 1600 stream processors, 850 MHz. So indeed the redesigned SIMD core is working its magic. Z/Stencil performance also shot up more than 100% over the Radeon HD 5700 series. Both the HD 6770 and HD 6750 will be equipped with 5 GT/s memory chips, at least on the reference-design cards, which are technically capable of running at 1250 MHz (5 GHz effective), though are clocked at 1050 MHz (4.20 GHz effective) on HD 6770, and 1000 MHz (4 GHz effective) on HD 6750. Although these design changes will inevitably result in a larger die compared to Juniper, it could still be smaller than Cypress, and hence, more energy-efficient.
245 Comments on AMD Radeon HD 6700 Series ''Barts'' Specs Sheet Surfaces
So, 1920 4-D stream processors (I'm beginning to doubt AMD will continue to call these "stream processors"), and 2 GB of memory over a 256-bit wide GDDR5 memory interface clocked at 6.40 GHz (1600 MHz), might just do the trick. No doubt 28 nm process will be ready by late Q1, early Q2 at TSMC, but NVIDIA surely won't make a GPU with higher transistor count than GF100 on it right away. So Cayman is going to have a very long stint.
So, 3840 cores on the Antilles. If AMD does decide to double Barts in the SIMD department for Cayman, you're looking at 5120 cores.Or, I'll decide your avatar for a month.
shader die space in cypress is 60% and 4D shader is 80% of 5D shader in size and SIMD controller and TMU took about 15% then here will be 2(334x 0.6 x0.8)+2(334x0.15)+334x0.25 = 320.64 + 100.2 +83.5 = 504.34mm^2 + hard wiring = 510mm^2
that is huge die and such 510mm^2 only has 32 rops????and i don't see any reason why we'd need 640ALU for? folding@home?
and you expect a 510mm^2 chip using a narrow 256bit bus on it?
if the shader turn out to be 5120(1280ALU) then the die size will be:
4(334x 0.6 x0.8)+4(334x0.15)+334x0.25 = 641.28 + 200.4 + 83.5 = 925.18mm^2 + hard wiring = 940mm^2......
shader like this are pointless if you don't have more rops to push it. like g92 was bottleneck by its 16 rop while it had 128 ALU. and now cayman that has 1280 ALU but 32 rops....that is a big joke...
if the specification turn out to be 1920:96:64 512bit story will be vastly different from above
1.5(334x0.6x0.8)+1.5(334x0.15)+2(334x0.25) = 240.48 + 75.15 + 167 = 482.64mm^2 + hard wiring = 484mm^2
480ALU is what we need in existed 40nm..no go further....
GTX 475 is aimed at Barts XT.
I enjoy reading it and all, and if you seculate enough, some part of what you say is bound to be right, it's like flipping a coin.
code name of amd's hd 6000 line
Antilles (6970~6990???)
Cayman (6850/6870~6950/6970???)
Whistler (???~6850/6870???)
Barts (6750~6770??)
blackcombs
Turks
Caicos
*edit*I'm glad it is not just me is who is confused by that post :laugh:
wiki may seem to be unreliable in some case (like comparison of graphic processing unit.) but this part has been protected under the community so i'm sure that article is real and it hasn't change since june/july.
wccftech.com/2010/08/27/upcoming-ati-hd-6000-series-codenames-revealed-catalyst-108/
www.nordichardware.com/news/69-cpu--chipset/40932-radeon-hd-6000-product-names-revealed-in-catalyst-108.html
this also prove its existence
*edit*Then i may have seen the picture of that assuming it was the small passive card, i can't remember the name as its too late/early here, i should be asleep.
gees...amd is acting like JJ Abrams as time goes.....actually everything in wiki are reliable except some fanboism controversy place like product comparison that's relate to spec of purely speculation. but when source come out it will be protect by community and these fanboi will go nowhere.
technology is progress as ground breaking/brutal force(new technology/architecture)=> tweak,efficient redesign(reconfigure/die shrink)=>upward and push further performance with brutal force again. that is what moore's law about and these average consumer are about to destroy it. however hd 6000 will be just as what moore's law predict and all IC industry will follow until the end of humanity! that won't change forever.
moore's law and technology only serve elite, not average joe.