SIE launches its new GPU platforms for scientific computing based on NVDIA Titan X for scientific computing.
The new HPC systems allow incorporating these new cards recently launched by NVDIA. Here you can see a comparison in teraflops of the benefits of the TITAN X card, integrated on 8.000 transistors with 12 GB of RAM.
In the HPC Titan X benchmark, we see a difference of 1,9Tflops between the Geforce GTX Titan X and the GeForce GTX980 card, all at lower power consumption.
Dr. Faraudo has carried out NAMD tests with our equipment and you can see the result at the following link: NAMD TESTS WITH NVDIA GTX 980 AND CUDA 6.5
El Dr. Jordi Faraudo from the ICMAB (CSIC) he tells us:
The same test as the other time, now I get the following estimate:
-4 CPUs, 2 GeForce GTX 980 Benchmark time: 0.0313843 days/ns
The comparison with the other two machines (mine and the one I tested in October):
– speed 2 Tesla M2075 GPUs: 0.0874105 days/ns
– speed 2 GPU Titan black: 0.0281269 days/ns, that is, the GTX980 is practically the same speed as the TitanBlack
You may also be interested: How can I program on GPU?