Nvidia hogs the AD103 GPU at 42.5 percent for its midrange graphics card

from Sven Baudouin
In addition to the AD104 chip with 5,888 shader units, Nvidia must also use the larger AD103 GPU, previously reserved exclusively for the Geforce RTX 4080, for the smaller Geforce RTX 4070 and trim accordingly.

Like a Twitter user known for such “revelations”. @employee It was once again the first to report that Nvidia is reportedly planning some sort of “leftovers” from partially crippled AD103 graphics processors and use on the already released Geforce RTX 4070.

Use leftovers with the main data remaining the same

Little or nothing will change for gamers or Geforce RTX 4070 buyers. The AD103 GPU, which provides 10,240 shaders, 320 TMUs, 112 ROPs and 80 CUs when fully expanded and already used in the Geforce RTX 4080, should be increased in an expansion phase by 95 percent, to 5,888 ALUs plus 184 TMUs and 64 ROPs are cut in this project thus matching the Geforce RTX 4070’s smaller AD104-250-A1 graphics processor.

GeForce RTX 4070 Geforce RTX 4070 (“New”)*
building Ada Lovelace Ada Lovelace
GPU m 104 m 103
shading units 5,888 5,888
texture units 320 320
Royal Oman Police 112 112
CUs 64 64
graphics memory 12 GB GDDR6X 12 GB GDDR6X
memory interface 192 bits 192 bits
bandwidth 504 GB / s 504 GB / s
TDP 200 watts 200 watts

*) Not officially confirmed yet!

Switching to the AD103 GPU won’t be an obstacle for Nvidia’s integrated partners because both GPUs use the same pin layout. In terms of specs, the Geforce RTX 4070 will always be the Geforce RTX 4070, whether it’s running an AD103 GPU or an AD104 GPU.

See also  Tom Clancy's Ghost Recon Frontline: Trial postponed at the last minute

Leave a Reply

Your email address will not be published.