GTX 460 rumored on the way...

http://www.nordichardware.com/en/ne...28www.nordichardware.com+News+%26+Articles%29

Latest saying more memory instead (1280 vs 1024) and no cutting of CUDA cores (384 vs 352 above). 384 was from the original rumours.
no way it will have 1280 if its a 256bit bus.

anyway here is an early review of the gtx465 and it is 352sp and 1024. it looks to be slower than a 5830 which makes sense considering how bad they butchered the sp. http://translate.google.com/transla...010/0512/A20100512652722.shtml&sl=zh-CN&tl=en
 
Link no-worky
yeah they probably pulled it.

anyway according to Fudzilla there will be a gtx465 based on the gf100 and then also a gtx460 based on gf104. http://www.fudzilla.com/content/view/18822/1/

"It appears that Nvidia's GTX 460 and GTX 465 branding has caused a bit of confusion among retailers and hacks. Some believe just one card is coming, while others are getting the specs a bit wrong.

The GTX 465 is a GF100-based card with 1024MB of memory and a 200W TDP. It should launch at Computex and it will pack a 256-bit memory interface and, according to unconfirmed reports, 352 shaders.

The GTX 460 on the other hand is a GF104-based card, so it will lack some features found on the GTX 465. However, just like the GTX 465 it will feature a 256-bit memory bus and 1024MB of memory. It has a 150W TDP, so it's clear that the GF104 is not as complex as the GF100 core. Listings claiming that the GTX 460 will pack 1280MB of memory seem to be wrong. As the GTX 465 is scheduled to launch at Computex, the GTX 460 should probably appear at a later date.

The big question with the GF104 core are the clocks. The core seems to be much more power efficient than the GF100 and Nvidia's decision to use it for its upcoming dual-GPU card is clear proof of this, therefore it would not surprise us if the GF104 ended up with somewhat higher clocks than GF100 cards. Be as it may, we still don't understand why Nvidia chose to use similar branding for two cards that appear to be based on entirely different architectures. "
 
Back
Top