The high-end killer for US $1200
Nvidia has officially announced its high-end killer in the form of Pascal-based Titan X, packing some serious specifications with its GP102 GPU and hopefully some impressive performance for its US $1200 price.
The new Nvidia Titan X is based on the 16nm GP102 Pascal GPU and features 56 streaming multiprocessors with a total of 3584 CUDA cores, 224 TMUs and 96 ROPs as well as provides an impressive 11 TFLOPs of 32-bit floating point compute performance. It also comes with 12GB of 10Gbps GDDR5X memory paired up with 384-bit memory interface, pushing 480GB/s of memory bandwidth.
The GP102 GPU behind the new Titan X is clocked at 1417MHz GPU base and 1531MHz GPU Boost clock and with a TDP of 250W, the Titan X draws power from a combination of 8-pin and 6-pin PCI-Express power connectors.
It features DisplayPort 1.4, HDMI 2.0b and DVI as display outputs but there are no details on the precise configuration.
The cooler on the new Nvidia Pascal-based Titan X is similar to what we have seen earlier but painted in black, trademark color for Nvidia's Titan lineup.
The new Titan X was unveiled by Jen-Hsun Huang at an event at Stanford University and while there are still a lot of unknowns, it will be coming on August 2nd and pack some impressive performance punch.
During its announcement of the new Nvidia Titan X, company introduced the new INT8 compute performance number in the specification list, which is used in neural network inference and Titan X pushes 44 TOPs of INT8 compute performance. Unfortunately, we don't have anything to compare it to, so we will take Nvidai's word that this one is fast.
Nvidia claims that the new Pascal-based Titan X is up to 60 percent faster than the Maxwell-based GTX Titan X and up to three times faster than the original Kepler-based Titan, but with 12 billion transistors, which is almost double the amount compared to previous Titan graphics cards and with a price of US $1200, it better be faster than anything on the market.
We will probably hear more about the new Titan X as we draw closer to the August 2nd launch date and we are looking forward to some performance numbers.
Source: Nvidia.com.
The new Nvidia Titan X is based on the 16nm GP102 Pascal GPU and features 56 streaming multiprocessors with a total of 3584 CUDA cores, 224 TMUs and 96 ROPs as well as provides an impressive 11 TFLOPs of 32-bit floating point compute performance. It also comes with 12GB of 10Gbps GDDR5X memory paired up with 384-bit memory interface, pushing 480GB/s of memory bandwidth.
The GP102 GPU behind the new Titan X is clocked at 1417MHz GPU base and 1531MHz GPU Boost clock and with a TDP of 250W, the Titan X draws power from a combination of 8-pin and 6-pin PCI-Express power connectors.
It features DisplayPort 1.4, HDMI 2.0b and DVI as display outputs but there are no details on the precise configuration.
The cooler on the new Nvidia Pascal-based Titan X is similar to what we have seen earlier but painted in black, trademark color for Nvidia's Titan lineup.
The new Titan X was unveiled by Jen-Hsun Huang at an event at Stanford University and while there are still a lot of unknowns, it will be coming on August 2nd and pack some impressive performance punch.
During its announcement of the new Nvidia Titan X, company introduced the new INT8 compute performance number in the specification list, which is used in neural network inference and Titan X pushes 44 TOPs of INT8 compute performance. Unfortunately, we don't have anything to compare it to, so we will take Nvidai's word that this one is fast.
Nvidia claims that the new Pascal-based Titan X is up to 60 percent faster than the Maxwell-based GTX Titan X and up to three times faster than the original Kepler-based Titan, but with 12 billion transistors, which is almost double the amount compared to previous Titan graphics cards and with a price of US $1200, it better be faster than anything on the market.
We will probably hear more about the new Titan X as we draw closer to the August 2nd launch date and we are looking forward to some performance numbers.
Source: Nvidia.com.
News by Luca Rocchi and Marc Büchel - German Translation by Paul Görnhardt - Italian Translation by Francesco Daghini
Previous article - Next article
comments powered by Disqus
comments powered by Disqus