qcmadness

http://news.softpedia.com/news/A ... overed-311697.shtml

引用:

Coming to face the “Enhanced Kepler” family of GPUs from NVIDIA, like the GK110, will be AMD's large and transistor-packed Sea Islands.

Chips will vary in might, but they will have up to 5.1 billion transistors, roughly 20% higher than the best Radeon HD 7000 GPU. That means 2560 stream processors. Other than that, the HD 8000 will have 42 ROPs (Tahiti has 32).

These are the specs of the top-tier chip, set to be used in the Radeon HD 8970 adapter.

The HD 8950 will be a bit weaker, obviously, with 2304 stream processors and, probably, lower clock speeds (they aren't known yet).

Lower in the hierarchy comes the Radeon HD 8870, with 1792 SPs, most likely. Then there is the HD 8850, with 1536.

And so we've arrived to the mid-range segment, where the “Oland” will show its worth. It should have around 896 processors and a bandwidth of 192 bits (GDDR5 VRAM). The Radeon HD 8770 will possess these specs.

Finally, the Radeon HD 8750 will offer customers 768 stream processors and a memory interface of 128 bits.

http://videocardz.com/35830/new- ... rce-gtx-780-details

引用:

NVIDIA GeForce GTX 780 Specification

    GPU: GK2xx
    Base Clock: 1100 MHz
    Boost Clock: 1150 MHz
    Memory Clock: 1625 MHz (6500 MHz effective)
    Memory type: 3GB GDDR5
    Memory Interface: 384-bit

According to Chip.de, the GeForce GTX 780, which would supposedly use this processor, may feature 2304 CUDA cores. The site also reveals that the card would use 3GB GDDR5 memory across a 384-bit interface. Nevertheless, our sources are less enthusiastic about this card. From what we know, it would have 1920 CUDA cores (2110 tops), 160 TMUs and 48 ROPs. Needless to say, probably the most certain information right now is the 384-bit interface. The site also mentions the exact clock speeds, but it seems that these are pure speculation, since NVIDIA does not usually launch their cards with such round clock numbers. Anyway, from what has been said so far, the GPU would be clocked at 1100/1150 MHz (for base and boost respectively), and the effective memory clock is 6500 MHz.

These specs suggest that the memory bandwidth would increase from 192 GB/s to 312 GB/s (that’s 62% increase). The texture fill-rate would be around 170.6 GT/s (32% increase) and pixel fill-rate 52.8 GP/s (64% increase)

又鬥大

Aware

又鏡面
我睇下TSMC今次又會整死邊間

GeN~XP

did it mention an anticipated release date ?

Puff

The rumored spec for Radeon seems consistent.

dom

Until AMD don't screwed up their driver
No More ATI until next gen

Puff

Charlie says that the initial Maxwell, the family after Kepler, is... is on 28nm. To be released in spring 2014. Consider that... err... AMD could have also del... moved their 28nm products a bit later than expected (fall 2012 to early 13?), may I draw a conclusion that the roadmaps of the entire foundry world slips one to two quarters?  lol... The delay in 28nm before gave no advantage to them. Perhaps AMD's Graphics Division can have a chance to WIN this time, with the ability to rapidly move to new process (if they can).

While the reality is that AMD will have no 20nm but 28nm products in 2014, probably except GPU in the first half and some big consumer stuff in the second half, it would be nice to see 20nm Cats and Excavator launch in 15H1.



[ 本帖最後由 Puff 於 2012-12-6 20:40 編輯 ]

符碌

gtx6xx系o岩o岩先出,咁快又話出7xx系,食懵左?

Puff

引用:

原帖由 符碌 於 2012-12-7 02:22 發表
gtx6xx系o岩o岩先出,咁快又話出7xx系,食懵左?

差唔多啦，一年一更新嘛。