GeForce GPU servers / FPGA Server / Iris Pro GPU server

OVH тоже представила конфиги, которые потом будут продаваться на основном сайте.
Потестить можно уже сейчас.
Кстати, потом цены обычно меняются. Например бета-тестеры забирали i7-6700k по 50 евро :) А сейчас они стоят 90 евро.

NVIDIA Grid GPU servers

(300e/w / 600e/m)

• Xeon E5-1650v3 / 128GB DDR3 / 2x 240GB SSD SATA3 / GPU 2x Grid K2
• Xeon E5-1650v3 / 128GB DDR3 / 2x 240GB SSD SATA3 / GPU 2x Grid M60
• www.runabove.com/nvidia-grid-gpu-servers.xml

Specs K2

• Graphic units: 2GPU Kepler
• 3072 CUDA Cores
• 8GB RAM GDDR5

Specs M60

• Graphic units: 2GPU Maxwell
• 4096 CUDA Cores
• 16GB RAM GDDR5

Iris Pro GPU server

• Xeon E3-1285v4 / 32GB DDR3 / 2x 2TB SATA3 / Intel Iris Pro Graphics P6300 (30e/w 80e/m)
• www.runabove.com/iris-pro-gpu-server.xml

• GPU Cores: 48
• RAM: 128MB eDRAM
• Performance: 883 GFLOPS

GTX1080 GPU Server

• Xeon E5-1650v3 / 128GB DDR3 / 2x 240GB SSD SATA3 / 2x TITAN X — 60e/200e
• 2x Xeon E5-2630v3 / 128GB DDR3 / 1x240GB SSD / 2x TITAN X — 170e/500e
• 2x Xeon E5-2630v3 / 128GB DDR3 / 1x240GB SSD / 1xGeforce GTX 1080 — 100e/300e
• www.runabove.com/titan-x-gpu-servers.xml

NVIDIA GeForce TITAN X

• NVIDIA CUDA Cores: 3072
• Frequency: 1000 MHz
• Boosted Frequency: 1075 MHz
• Memory: 12GB GDDR5
• Memory: bandwidth 336,5 GB/s

NVIDIA GeForce GTX 1080

• NVIDIA CUDA Cores: 2560
• Frequency: 1607 MHz
• Boosted Frequency: 1733 MHz
• Memory: 8GB GDDR5X
• Memory: bandwidth 320 GB/s

FPGA Server

(200e/400e)

• FPGA Card: Bittware A10PL4 Card ALTERA Arria 10 GX 1150 FPGA Chip 2x4GB Memory
• CPU Intel E5-1650v3 / 128GB RAM / 2x240GB SSD
• www.runabove.com/FPGAaaS.xml

What are the typical application fields of FPGAs?
Because FPGA are hardware programmed to run one application. The performance can be much higher than on a classical CPU architecture (with software compilation). The main fields are:

• video & image processing,
• communications,
• digital signal processing,
• high performance computing,
• data mining.