€ 8.567,68 (€ 10.195,54 inkl. MwSt.)
tesla-v100-32GB
+

NVIDIA Tesla V100 32GB CoWoS HBM2 ECC Memory - 900GBps HPC/DL

TCSV100MPCIE-PB; Memory Bus Width: 4029bit; Memory Bandwidth: 732GB/s

NVIDIA® TESLA® V100 - THE MOST ADVANCED GRAPHICS PROCESSOR FOR RECHENZENTREN, available with GPU server only

Artificial intelligence for self-propelled cars. Forecasts for our climate. A new drug for cancer treatment. For some of the most important challenges in the world, we urgently need to find a solution that requires enormous computing effort. Today's data centers rely on many interconnected standard computing nodes. However, this slows down the performance required for important calculations in the areas of High Performance Computing (HPC) and Hyperscale. The NVIDIA® Tesla® V100 is the most advanced graphics processor ever designed for the data center. They use the new NVIDIA Volta™ GPU architecture to deliver the fastest computing node in the world, with performance equivalent to hundreds of slower standard computing nodes. Higher performance with fewer, faster-paced nodes enables data centers to significantly increase throughput while saving money. More than 400 HPC applications, 9 of them from the 10 most important, and all deep learning frameworks have already been accelerated. Every HPC customer can therefore use accelerating GPUs in their data center.

Order Information

Ordering Part #:
TCSV100MPCIE-PB

GPU Features

Architectur:
  • Volta
Dual-Link DVI:
DisplayPort:
HDMI:
VGA:
Slots:
  • 2
Form Factor:
  • 111,15 mm (H) x 267,7 mm (L) Dual Slot, Full Height
  • Full Profile
Interface:
  • PCI-E 3.0 x16
Max Power:
250W
Cooling Solution:
  • passive
Special Features:
  • CUDA
  • Deep Learning
  • DirectCompute
  • OpenACC
  • OpenCL
  • Page Migration Engine
Package Content:
  • 1 x Power adapter (2x PCIe 8-pit auf single CPU 8-pin)
Compute APIs:
CUDA, DirectCompute, OpenCL ™ , OpenACC

GPU Memory Feature

GPU Memory:
32GB
Memory Bus Width:
  • 4029bit
Memory Bandwidth:
732GB/s
Memory Clock:
1.4MHz
Error Correcting Code Memory ECC:

GPU Performance

Tensor Performance:
112 Tflops (GPU Boost Clocks)TFLOPS
Tensor Cores:
640Tensor Cores
CUDA Processing Cores:
5120Cores
Boost Clock:
1300MHz
Single Precision FP32:
14TFLOPS
Double Precision FP64:
7TFLOPS