The network has leaked unofficial NVIDIA Quadro M6000 graphics card specification, which is equipped with a chart-topping generation graphics processor of NVIDIA Maxwell.
NVIDIA Quadro M6000 Specifications- Maxwell GM200 professional equivalent to GeForce GTX 980 Ti
If you believe on the disclosed specifications, the new card is based on the layout of the Maxwell GM200 with 3072 CUDA cores and 96 ROPs – so we can say around 50% more than the Maxwell GM204 GPU used in the GeForce GTX 980. In addition, the VRAM on board is up to 12 GB of GDDR5 memory with 384-bit interface. Running at 988 MHz core clock and 6600MHz memory frequency.
There is no doubt it is true Quadro accelerators which has been designed for the professional use, but the graphics processor is usually also used in the construction of the other series. In this case, the Maxwell GM200 core will probably also used in the model of GeForce GTX 980 Ti, and the number of units may indicate a significant increase in performance with respect to the current leader of the GeForce GTX 980 (though obviously not at the level of 50%). Some alleged specifications of NVIDIA Quadro M6000 in compression of NVIDIA Quadro K6000.
Model | Quadro K6000 | GeForce GTX 980 | Quadro M6000 |
The layout | Kepler GK110 | Maxwell GM204 | Maxwell GM200 |
CUDA Cores | 2880 | 2048 | 3072 |
TMUs | 240 | 128 | 192 (?) |
ROPs | 48 | 64 | 96 |
GPU clock speed | 980 MHz | 1126/1216 MHz | 988 MHz |
Video memory | 12 GB of GDDR5 384-bit | 4 GB GDDR5 256-bit | 12 GB of GDDR5 384-bit |
Memory clock | 6000 MHz | 7010 MHz | 6600 MHz |
Memory Bandwidth | 288 GB / s | 224 GB / s | 317.4 GB / s |
It is worth mentioning that NVIDIA has prepared a special conference at the fair CES 2015 – perhaps we will be able to see the launch of new NVIDIA Quadro M6000.
Source: techpowerup