Tachyum Presents Software Stack for HPC/AI and Supercomputer Management

Trending...
LAS VEGAS ~ Tachyum, a Las Vegas based company, released a white paper today detailing the software packages offered by the company and supported by its software team that enable software developers to build HPC applications deployed as part of Prodigy®-empowered data centers.

The white paper, titled "Tachyum HPC/AI Software Stack Overview", outlines the components of Tachyum's software stack which are designed to meet the high demands of cloud and HPC/AI workloads with Prodigy. The components include firmware such as BIOS (UEFI) and BMC, operating system level packages like LINUX Kernel and SELinux, storage and file-systems options like Ceph, OS level virtualization tools like Docker and containerd, Podman based on Kubernetes pods, communication stack and runtimes such as OpenMPI, MPICH, gPRC and RabbitMQ.

The user space utilities and scripting languages supported by Tachyum include Bash, SED, Gawk, Grep, Wget, Rsync, Strace, Git, Subversion as well as PHP Perl Ruby Tcl Top Htop Iotop Sar Nc. Parallelism standards such as OpenMP and SHMEM are also supported. Programming languages and compilers ported by Tachyum include GNU gcc compiler and go compiler. Debugging tools like GDB KGDB OpenOCD are also available along with emulators developed by Tachyum for internal use or for customers to translate X-86 ARM or RISC-V software to Tachyum's Prodigy architecture.

More on nvtip.com
Clustering management and job scheduling is handled by SLURM for jobs management Zabbix Prometheus with Grafana for monitoring Kubernetes for automating software deployment scaling and management of containerized applications Ansible for IT automation while networking capabilities are provided by SONIC network OS developed by Tachyum itself.

AI/HPC frameworks libraries supported include BLAS GEMM NumPy Eigen PyTorchTensorFlow while AI applications such as Resnet ConvMixer SSD Yolo Fast-SCNN UNET BERT are also available along with HPC applications Quantum espresso DeepMD LAMMPS.

Dr Radoslav Danilak founder and CEO of Tachyum said "Multiple different processing technologies are combined in today's HPC systems Achieving the required performance and functionality at an affordable power envelope places a lot of challenges to HPC software environments – especially on scheduling and resource management Tachyum's Prodigy the world's first Universal Processor integrates the functionality of CPU GPGPU and TPU into a single homogeneous architecture so that there is no additional complexity needing to be solved on the software level between hardware components"

More on nvtip.com
Prodigy delivers unprecedented data center performance power and economics reducing CAPEX OPEX significantly It is useful for both high-performance line-of-business applications allowing servers to seamlessly switch between workloads eliminating expensive dedicated AI hardware while increasing server utilization Performance up to 4x that of highest performing x86 processors (for cloud workloads) 3x that of highest performing GPU for HPC 6x for AI applications can be achieved with Prodigy powered data centers.

Those interested in reading "Tachyum HPC/AI Software Stack Overview" white paper can download it from https://www.tachyum.com/resources/whitepapers/2.... For more information about Tachyum follow them on Twitter LinkedIn or Facebook at https://twitter.com/tachyum https://www.linkedin.com/company/tachyum https://www.facebook.com/Tachyum/.
Filed Under: Business
Show All News | Report Violation

0 Comments

Latest on nvtip.com