NVIDIA Tesla cGPUs future will be based in Denver
Yesterday announced its new microprocessor ARM Cortex-A57 (Atlas) and Cortex-A52 (Apollo) , the new duo big.LITTLE to 64 bits, but even before they are officially announced, many companies like Nvidia and AMD are developing future products based in them, among which we can mention the Nvidia CPUs Boulder and AMD Opteron ARM64 CPUs and hybrid (CPU + GPU) Nvidia Denver and the recently announced AMD Opteron APU ARM64 .
Nvidia reveals future Tesla cGPUs for super-computers and servers will be made up of hybrid designs that combine multiple CPUs armv8 (Cortex-A57) and multi-core gpu (shader processors) based on its proprietary graphics technologies, ie: your project Denver.
The hybrid chips allow applicants performing tasks as traffic data intensive cloud computing, financial transactions, data mining (among other possibilities) offering high performance and low power consumption, but for certain types of applications, such as targeted to complex scientific calculations, it will continue to require the x86 microprocessors from Intel and AMD.
Nvidia declined to provide an exact release date for these future products merely stating it will be released sometime this decade.
Link: Nvidia to pair CPUs, graphics chips in next Tesla products (ComputerWorld)Tags: ARM64, armv8, cGPU, Cortex-A57, CPU, denver, GPGPU, gpu, hybrid, Nvidia, server, supercomputer, Tesla