ARC is proud to announce the release of a new cluster called Infer, which provides 18 Intel Skylake nodes each equipped with an Nvidia T4 GPU. The cluster’s name "Infer" alludes to the AI/ML inference capabilities of the T4 GPUs derived from the "tensor cores" on these devices. We think they will also be a great all-purpose resource for researchers who are making their first forays into GPU-enabled computations of any type. For more information about the T4 architecture on T4 architecture and performance relative to, e.g., V100 GPUs, see the following pages:
Cluster details and examples are provided on our Infer page. Users may login at infer1.arc.vt.edu and try it out at their earliest convenience. For now, software installs are mostly limited to CUDA and associated toolchains; if users have additional requests please, they may submit them via Help ticket.
ARC is happy to announce the addition of 39 new GPU nodes to the NewRiver cluster. Each of these nodes is equipped with two Intel Xeon E5-2680v4 (Broadwell) 2.4GHz GPU (28 cores/node in all), 512 GB memory, and two NVIDIA P100 GPUs. Each GPU is capable of up to 4.7 TeraFLOPS of double-precision performance, so including CPU and GPU these nodes add over 400 TFLOPS of peak double-precision throughput to ARC's resources.
Continue reading P100 GPU Nodes added to NewRiver
ARC is happy to announce the release of a new cluster, named Cascades, available at cascades1.arc.vt.edu and cascades2.arc.vt.edu. Cascades is a 196-node system capable of tackling the full spectrum of computational workloads, from problems requiring hundreds of compute cores to data-intensive problems requiring large amount of memory and storage resources. Cascade contains three compute engines designed for distinct workloads:
- General – Distributed, scalable workloads. With Intel’s latest-generation Broadwell processors, 2 16-core processors and 128 GB of memory on each node, this 190-node compute engine is suitable for traditional HPC jobs and large codes using MPI.
- GPU – Data visualization and code acceleration! There are four nodes in this compute engine which have - two Nvidia K80 GPUs, 512 GB of memory, and one 2 TB NVMe PCIe flash card.
- Very Large Memory – Graph analytics and very large datasets. With 3TB (3072 gigabytes) of memory, four 18-core processors and 6 1.8TB direct attached SAS hard drives, 400 GB SAS SSD drive, and one 2 TB NVMe PCIe flash card , each of these two servers will enable analysis of large highly-connected datasets, in-memory database applications, and speedier solution of other large problems.
Continue reading New ARC Cluster: Cascades
ARC is happy to announce the release of a new cluster, named DragonsTooth, available at
dragonstooth1.arc.vt.edu. DragonsTooth is made up of 48 nodes, each equipped with:
- 2 x Intel Xeon E5-2680v3 (Haswell) 2.5 GHz 12-core CPU (same CPU as NewRiver)
- 256 GB 2133 MHz DDR4 memory for large-memory problems
- 4 x 480 GB SSD Hard Drives for fast local I/O ($TMPDIR)
- 806 GFlops/s theoretical double-precision peak
Continue reading New ARC Cluster: DragonsTooth
ARC HPC Systems will undergo maintenance beginning at midnight on the morning of Tuesday, March 29, 2016. The purpose of this maintenance will be to migrate to a new shared Home directory on the file system that currently provides Home to NewRiver. This will provide two key benefits to users:
- All files in your Home directory will be visible from all clusters. For example, you will see the same files in $HOME from both NewRiver and BlueRidge. This will make it easier to migrate work between clusters based on which hardware is best suited to the task or which resource is less busy.
- The maximum Home directory size will be increased from 100 GB to 500 GB per user.
Continue reading ARC Migrating to Shared Home Directories, 29 Mar 2016