On many enterprise platforms, builders have totally different objectives for his or her software program, corresponding to: Excessive efficiency, throughput, reminiscence or addressing the primary downside of the day. It's one factor to write down software program on a pc sort after which deploy it to the info middle, nevertheless it's one other factor to write down code on the same pc that it’ll land on. Most massive server platforms, in a single kind or one other, have an OEM that gives the equal of a "growth tower" that enables customers to write down code with the at their desks. This time, GIGABYTE has developed an EPYC tower that’s used for each code growth and small / medium companies.
The W291-Z00, a reputation that may roll from the tongue, is an EPYC platform with a socket constructed round one of many GIGABYTE motherboards and as much as 32 cores, 64 threads, 1TB DRAM ( eight slots) and may accommodate 4 highly effective GPUs. The thought right here is that this may be constructed as a beast when wanted, and the chassis makes use of baffles to help even passive GPUs or accelerators. There's the equal of 9 PCIe slots within the chassis, so accelerators bigger than dual-slot could be supported on the bottom PCIe slot. These are powered by two redundant 1600 watt 80 plus platinum energy provides.
There are 4 hot-swappable entrance bays that help three.5 "and a couple of.5" SATA or SAS drives, and there may be additionally an M.2 slot on the motherboard. Extra storage is supported through riser playing cards within the PCIe slots or through the 4 mini-SAS ports for a complete of 16 drives. Administration is through the ASpeed AST2500 BMC. Every of the 4 followers could be set by means of the GIGABYTE console administration interface. The networking is finished with two Intel I210 AT Gigabit community ports and one other for the BMC.
GIGABYTE's goal markets for the W291-Z00 embrace small workplaces, small / medium companies and universities, in addition to these concerned in simulation, machine studying or AI / HPC workloads. With GPUs, the corporate additionally targets accelerated AI workloads in addition to graphical CAD and Rending workloads.
Because the product is offered right this moment as this product comes from the GIGABYTE server group, events should contact their native distributor for pricing.