Specifications of the NVIDIA GeForce GTX 780 Ti and GeForce GTX 780 Ti Black Edition
This Thursday November 7 NVIDIA launched two editions of its GeForce GTX 780 Ti GPU ” GK110 “.
In just a few days before the launch of video cards based on the new GPU Nvidia GeForce GTX 780 Ti ” GK110 “, which will arrive on Thursday November 7.
So far the data we had on this new GPU was based on unsubstantiated rumors, but apparently, thanks to a leak inadvertently revealed by Galaxy, one of the major partners of Nvidia, we finally have almost complete specifications this new chip :
- Based on the GK110 Kepler graphics core.
- 2880 shader processors running at 876MHz ( to 928MHz Boost Mode ).
- 240 texture units ( TMUs ).
- 48 rendering units ( ROPs ).
- 120 shaders FP64.
- 3GB of GDDR5 memory with 384-bit bus running at 1752MHz ( 7008MHz effective).
- Between 250 and 300W TDP.
- Powered by 2 8 +6 pin connectors.
- MSRP from USD $ 699.
Additionally there will be a special edition called GeForce GTX 780 Ti Black Edition, which like the recently revealed GeForce GTX 780 GHz Edition GPU, takes also one of the nomenclatures used by rival AMD. This variant will be available with a cooling system in black and comes with 6 and 12GB memory, besides having a slightly higher operating frequency.
Apparently the AMD Radeon GPU R9 290x “Hawaii XT “, forced Nvidia to launch a GK110 Kepler GPU with almost all units active: GeForce GTX 780 Ti will have just 120 of its 960 active FP64 shaders, for what is expected that the rumored GeForce GTX Ultra Titan 2880 possesses 960 standard shaders shaders FP64.
We have no doubt that resume GeForce GTX 780 Ti graphics card will retain the performance crown for Nvidia, beating the AMD Radeon GPU beating R9 290x.
It is rumored that GeForce GTX Titan will maintain its high price ( $ 999 ) due to its high computational FP64 power compared to other GPUs from GeForce, meeting the needs of many users that require application performance graphics and an intensive accelerated computing GPU, that require a 64-bit floating point ( FP64 ).