On March 21, AMD announced the availability of AMD Instinct ™ ecosystem with extended system support from partners, including ASUS, Dell Technologies, Gigabyte, HPE, Lenovo and Supermicro. The new AMD Instinct ™ MI210 accelerator and the powerful features of ROCM ™ software 5. In general, the AMD Instinct and ROCM ecosystem offers exphops-class technologies to a wide range of customers using high-performance computing and artificial intelligence, satisfying the growing demand for data centers workloads with acceleration of calculations. and reducing the time required for analysis and detection.

Ensuring future high-performance computing and artificial intelligence:

AMD Instinct AMD Accelerators of the MI200 series are designed to hold openings in exhaustopus systems, allowing researchers, scientists and engineers to solve the most pressing tasks, from climate change to vaccine studies. AMD Instinct Mi210 accelerators specially support Exphopus class technologies for customers who need a fantastic performance of high-performance computing and artificial intelligence in PCIE® format. AMD Instinct Mi210 Accelerators, based on AMD CDNA ™ 2 architecture, expand AMD leadership in double accuracy computing performance (FP64) on maps with PCIE form factor. They also represent a reliable solution for accelerated learning to deeply learning, offering a wide range of mixed accuracy capabilities based on AMD Matrix Core.

Stimulating implementation:

ROCM AMD ROCM platform with open software that allows researchers, scientists and engineers to use AMD Instinct accelerator capabilities for scientific discoveries based on numerous applications and libraries underlying the best applications for high-performance computing and artificial intelligence.

With ROCM 5 AMD expands its software platform by adding new hardware support for AMD Instinct MI200 series accelerators and the AMD Radeon ™ Pro W6800 professional video card, as well as support for Red Hat® Enterprise Linux® 8.5, increasing the availability of ROCM for developers and allowing outstanding performance with key Workloads.

In addition, via AMD Infinity Hub, a central place for open source applications, portable and optimized AMD graphics processors, end users can easily find, download and install HPC container applications and machine learning platforms. AMD Infinity Hub applications containers are designed to reduce the traditionally complex problem of receiving and installing software issues, allowing users to study on the basis of the overall experience and opportunities to solve problems.

Expansion of the affiliate and client ecosystem:

As more and more specialized applications are optimized for working with ROCM and AMD Instinct accelerators, AMD continues to expand its software ecosystem by adding commercial independent software providers, including ANSYS®, Cascade Technologies and Tempoquest. These ISV provide applications for accelerated workloads, including computing hydrodynamics (CFD), weather forecasting, automated engineering (CAE) and much more. These updates complement the existing application support in ROCM, including HPC, AI and machine learning applications, Amber, Chroma, CP2K, Grid, Gromacs, Laamps, Milc, Mini-Haac, Namd, Namd 3.0, Onnx-RT, Openmm, Pytorch, Relion , Specfem3D Cartesian, SpecFem3D Globe and TensorFlow.

AMD also allows its partners such as Asus, Dell Technologies, Gigabyte, HPE, Lenovo, Supermicro and system integrators, including Colfax, Exxact, Koi Computers, Nor-Tech, Penguin and Symmetric, offering differentiated solutions to solve the next generation computing tasks. Supercomputer customers already enjoy the advantages offered by these new client victories, including the FRONTIER installation in the National Ok-Ridge laboratory, KTH / DARDEL, CSC / LUMI and CINES / ADASTRA.

Ensuring access for customers and partners:

The AMD Accelerator Cloud cloud offers clients Wednesday to remote access and estimating AMD Instinct accelerators and AMD ROCM software. Whether it is transferring an outdated code, a benchmarking application or a multi-GPU scaling testing or multiple nodes, the AMD Accelerator Cloud cloud provides potential customers and partners to quick and easy access to the latest graphics processors and software. The AMD Accelerator Cloud cloud is also used for various events, such as Hakaton and ROCM training activities, offered to both existing and potential customers, which allows developers to drag their skills and find out how to get the maximum return from AMD Instinct accelerators.