Tyan's Intel-powered server hardware balances the demands of compute, storage, memory and IO for specific application areas, rather than trying to be a one-size-fits-all system.

The challenge for modern high-performance computing (HPC) systems is to read and process increasingly vast amounts of data. But it's simply not enough to throw more processing power at the problem. Different application areas require different hardware design approaches in order to maximize workload performance. Customization, therefore, is key.

For example, IO-heavy HPC computing tasks (e.g. genome sequencing and weather forecasting) typically require large amounts of data to be read/written to disk. So, the speed of storage IO on the servers used for the file system has a big influence on how efficiently The file system performs. To avoid becoming IO-bound, servers can be optimized with fast NVMe storage, which is directly accessible to the processor.


This approach might look like the Tyan Thunder CX GC68A-B7126, a barebones cloud server that's ideally suited to data center use. Built around a 3rd Gen Intel® Xeon® Scalable processor and fully customizable, this 1U cloud server can take advantage of integrated workload acceleration features such as Intel® Deep Learning Boost. It not only incorporates 12x 2.5" SATA drive bays (with optional 2x NVMe U.2 hot-swap bays), but also supports a large memory capacity, 2x FH/HL PCIe Gen.4x16 slots for high performance NICs, IPMI 2.0, and (1+1) 850W redundant power supplies.

Big Data computing needs bulk storage

Of course, one server design doesn't fit every use-case. Or every computing vertical. For Big Data computing workloads, the hardware requirements revolve around distributed bulk storage rather than single node performance. Applications like Hadoop-Hive, Spark, and the Hadoop Distributed File Systems (HDFS) are typically used to manage the huge amounts of data used for analytics operations. This, in turn, necessitates computing nodes with storage that focuses squarely on large capacity at the lowest cost per byte.



A 2U dual-socket server like the Tyan Thunder SX TS65-B7126 is a smart solution in this regard, capable of scaling up or down as needed. With 12 drive bays, you can stack it with the full complement of 18TB Enterprise Class 7200RPM SATA drives to deliver 216TB of raw storage. And, by providing options for large amounts of memory and the latest Intel Xeon　Scalable processors (with options available with up to 40 cores), the SX TS65-B7126 can be configured to tackle AI-inference applications and other complex Big Data workloads.

Multi-GPU systems power Deep Learning

The last design consideration is for Accelerated HPC Computation, which includes both traditional HPC and Deep Learning systems. Applications in this area can include materials and molecular science, computational fluid dynamics, financial engineering, pharmacology and oil/gas exploration.

Again, Intel-powered solutions like the Tyan Thunder HX FT83A-B7129 can shine here, leveraging deep learning acceleration built directly into Intel Xeon processors. These 3rd Gen CPUs power the base computing platform (large memory, computational cores, operating system, IO, networking, etc.), while an array of GPUs act as accelerators for parallel processing. The HX FT83A-B7129, for example, can support up to 10 DW GPU alongside one spare PCIe Gen.4 x16 slot.

Built for the most demanding HPC workloads, the Tyan Thunder series is a line of world class barebones server systems based on Intel's latest technology, including 3rd Gen Intel Xeon　Scalable Processors. For more information about our newest products, view the Tyan/Intel Product Catalog 2022.