The Nvidia Titan RTX just went live on the NVIDIA website, the professional targeted graphics card was announced two weeks ago. The card holds 576 tensor cores and 72 raytracing cores, slightly more than the RTX 2080 Ti with it’s 544 and 68 values respectively.
With 24 GB of GDDR6 memory, the TU102 GPU should have plenty of it. The flagship card has 4608 shader processors and is aimed at AI researchers and data scientists, rather than gamers; Nvidia claims the card “transforms the PC into a supercomputer” allowing faster training and inference of neural networks and enables researchers to experiment with larger neural networks and data sets.
 |
|||
| Titan RTX | Titan V | RTX 2080 Ti Founders Edition |
|
| Shader Cores | 4608 | 5120 | 4352 |
| Tensor Cores | 576 | 640 | 544 |
| Core Clock | 1350MHz | 1200MHz | 1350MHz |
| Boost Clock | 1770MHz | 1455MHz | 1635MHz |
| Memory Clock | 14 Gbps GDDR6 | 1.7 Gbps HBM2 | 14 Gbps GDDR6 |
| Memory Bus | 384-bit | 3072-bit | 352-bit |
| Memory Bandwidth | 672GB/sec | 653GB/sec | 616GB/sec |
| Memory | 24GB | 12GB | 11GB |
| Single Precision | 16.3 TFLOPS | 13.8 TFLOPS | 14.2 TFLOPS |
| Double Precision | 0.51 TFLOPS | 6.9 TFLOPS | 0.44 TFLOPS |
| Tensor Performance (FP16 w/FP32 Acc) |
130 TFLOPS | 110 TFLOPS | 57 TFLOPS |
| GPU | TU102 (754mm2) |
GV100 (815mm2) |
TU102 (754mm2) |
| Transistor Count | 18.6B | 21.1B | 18.6B |
| TDP | 280W | 250W | 260W |
| Cooling | Active | Active | Active |
| Manufacturing Process | TSMC 12nm FFN | TSMC 12nm FFN | TSMC 12nm FFN |
| Architecture | Turing | Volta | Turing |
| Price | $2499 | $2999 | $1199 |
The card launches at in the U.S. and Europe for $2,499. You do get free shipping for that money though 😉
