Intel Confirms Ponte Vecchio As 7nm General Purpose GPU
In addition to announcing the much anticipated oneAPI beta, Raja Koduri spent his time at Intel's event today also talking about "Ponte Vecchio" as their forthcoming general purpose GPU.
Ponte Vecchio will be their first general-purpose GPU optimized for HPC/AI workloads and based on the Xe architecture. The 7nm Ponte Vecchio will be focused on delivering "highly flexible discrete general-purpose GPU is architected for HPC modeling and simulation workloads and AI training."
This will be the first Intel Xe graphics card to market next year and make use of their Foveros 3D and EMID packaging, support Compute Express Link (CXL), and other yet to be announced features.
Intel confirmed the ANL Aurora supercomputer due to be ready in 2021 will feature two Xeon Scalable "Sapphire Rapids" CPUs and six Ponte Vecchio GPUs per node.
More details at Intel.com on their SC19 announcements.
Ponte Vecchio will be their first general-purpose GPU optimized for HPC/AI workloads and based on the Xe architecture. The 7nm Ponte Vecchio will be focused on delivering "highly flexible discrete general-purpose GPU is architected for HPC modeling and simulation workloads and AI training."
This will be the first Intel Xe graphics card to market next year and make use of their Foveros 3D and EMID packaging, support Compute Express Link (CXL), and other yet to be announced features.
Intel confirmed the ANL Aurora supercomputer due to be ready in 2021 will feature two Xeon Scalable "Sapphire Rapids" CPUs and six Ponte Vecchio GPUs per node.
More details at Intel.com on their SC19 announcements.
19 Comments