The new Ampere technology,
both at the level of the Tesla A100 PCIe GPUs, as well as the RTX, it has been a real revolution in the sector.
The new Ampere technology, both at the level of the Tesla A100 GPUs and the RTX, has been a true revolution in the sector.
SIE launches its entire range of products, which will be available from the end of September to December, in its different configurations, based on second and third generation AMD processors Zen2 and Zen3, respectively. All solutions with NVIDIA Ampere technology and
PCI-e 4.0 connectivity, which allows twice the bandwidth of the current bus.
Sistemas Informáticos Europeos, as a Gigabyte wholesaler and direct distributor of PNY for southern Europe and other countries in the world such as Mexico, Chile or Nepal, offers its solutions based on these solutions, with a 3-year total warranty on labor and parts.
From lowest to highest computing power we present:
1st Workstation that support up to 64 cores and allow the incorporation of two NVIDIA RTX 30 cards that provide more than 20,000 CUDA cores.
2nd 4U rack servers that equip up to 10 Tesla A100 cards with PCI-e 4.0 bus and with configurations of AMD Rome (EPYC 7002 second generation) or the new AMD Milan, when available, without the need for hardware modification. Based on the Gigabyte G492-Z50 and Gigabyte G492-Z51 .
PLAT G492-Z50
These platforms also allow a large storage capacity for Big Data solutions and intensive analysis, where the amount of data is massive.
3rd Servers with Tesla A100 SXM4 technology, which provide an aggregate power of 40,960 CUDA cores, the 5120 Tensor Core and 320 GB of RAM, through the new NVLINK bus. This represents a power for Deep Learning and Artificial Intelligence of the order of 5 Exaflops.
PLAT G492-ZD0
In addition, this new technology represents improvements of the order of twice the performance, also in classic HPC solutions, such as programs for molecular dynamics
All our solutions, which we will launch between now and the end of the year, will allow us to take a giant step forward in terms of Big Data, Deep Learning, Machine Learning, HPC and Artificial Intelligence solutions.
They also incorporate all the CUDA 11 technology, which are available in more than 700 calculation applications, in sectors such as computational chemistry, aerodynamics, meteorological models, engineering, etc. Specific libraries such as Caffe2, Pytorch 1.6, Apache Spark, CuDNN 8, Tensorflow 2.3, Magma 2.5, Keras 2.2, allow us to offer our clients turnkey solutions, ready to work from day one.
All these solutions are available on the latest generation operating systems such as Ubuntu 20.04, Centos 8.2 and Windows 10 Workstation, which allow to get the most out of all SIE Ladon platforms.
Given the complexity of so many new releases, SIE offers its 20-year know-how within the HPC sector and solutions for scientific computing, which are completed with
high-performance storage solutions such as BeeGFS, Mellanox Ethernet connectivity in 100 Gb / s and Mellanox Infiniband in 200 Gb / s, in low latency, as well as the interconnection switches of the same manufacturer.
In addition, we offer virtualized solutions on an open source platform such as PROXMOX and hyperconvergence such as Open Stack and Open Nebula.
And finally, to announce that SIE is a Distributor for Europe and Latin American countries of Gigabyte Server, so that in the coming weeks it will have a team with these characteristics so that whoever is interested can test the cards, the equipment and their power, all remotely.
In order to test it, please fill out the following form by clicking on ‘MORE INFORMATION’ below.
