NVIDIA TITAN Xp is the world’s most powerful graphics card. Incredible computing horsepower and groundbreaking NVIDIA Pascal™ architecture give you the power to accomplish things you never thought possible.
The new card uses 3,840 unlocked CUDA cores, besting the 3,584 in the 1080 Ti. It still packs 12GB of GDDR5X RAM, the same as the last model, but does so at 11.4GHz, up from 10GHz before. That helps it push pixels and voxels at 547.7 GB/s, up from 480GB/s, enough speed to let it handle Crysisas if it were Candy Crush.
The TITAN Xp is a $1200 card, so you will be paying quite a heavy premium to have the best performing card on the market – a premium of $500 to be precise. According to the benchmark given below, the score scores approximately 32,000 points on FireStrike 1.1 (look at the graphics score shown below the overall score) which is very impressive performance. In fact, this is roughly equal to a GTX 1070 SLI without any of the issues associated with SLI. It really is a good deal if you are a high end PC enthusiast and aren’t really worried about the value proposition.
TITAN Xp is crafted to offer superior heat dissipation using vapor chamber cooling technology in a die-cast aluminum body. It’s a powerful combination of brilliant efficiency, stunning design, and industry-leading performance.
Given below are the complete specifications for the new TITAN Xp as well as the comparison to its predecessor as well as the GTX 1080 Ti. As we can immediately see, the major improvement is in the form of the core count increase and the bandwidth increase. This is actually more bandwidth than the HBM1 specification we have been used to (which tops out at 512 GB/s). The major limiting factor here, for the 12.15 TFLOPs at stock clocks number (since performance is a function of 2 * core count * clock speed) is the clock speed. And the clock speed is of course bottleneck by the blower style cooler.
NVIDIA TITAN Xp (2017) Specifications Comparison
|NVIDIA Graphics Card||TITAN X (Pascal)||GTX 1080 Ti||TITAN Xp|
|GPU Die Size||471mm²||471mm²||471mm² (TBC)|
|CUDA Cores Per SM||64||64||64|
|FP32 CUDA Cores (Total)||3584||3584||3840|
|Compute Power||10.97 TFLOPs||11.34 TFLOPs||12.15 TFLOPs|
|Memory Size||12 GB GDDR5X||11 GB GDDR5X||12 GB GDDR5X|
|Memory Clock||10 Gbps||11 Gbps||11.4 Gbps|
|Bandwidth||480 GB/s||484 GB/s||547.7 GB/s|
Get it from here.