Nvidia's next-generation H100 Tensor Core GPUs and Quantum-2 InfiniBand are now widely available, in Microsoft Azure and more than 50 partner systems from the company's partners including Asus, Atos, Dell Technologies, Gigabyte, HPE, Lenovo, and Supermicro.
As usual, Nvidia has used its annual GTC conference to announce multiple new products. On the enterprise side, they include Nvidia Spectrum-4, said to be world's first 400Gbps end-to-end networking platform for data centres.
GPU vendor Nvidia has announced that its AI and HPC software will be available for Arm-based computers by the end of 2019.
Updated: Hot on the heels of yesterday's Top500 supercomputer announcement comes a reminder that Australia is a significant player in the field with the only NVidia CUDA research centre in the southern hemisphere.
Most cybersecurity is making up for weak platforms. We need to address the fundamentals, design platforms that prevent out-of-bounds access[…]
For most developers the security/performance trade off is still the hardest one to tackle, even as the cost of processing[…]
RISC has been overhyped. While it is an interesting low-level processor architecture, what the world needs is high-level system architectures,[…]
There are two flaws that are widespread in the industry here. The first is that any platform or language should[…]
Ajai Chowdhry, one of the founders and CEO of HCL is married to a cousin of a cousin of mine.[…]