About: NVIDIA Tesla K80 GPU Passive Accelerator

The Tesla K80 GPU Passive Accelerator from NVIDIA Boosts technology to dynamically maximize performance of a Tesla K80 GPU based on workloads. Peak Double Precision Floating Point Performance 2.91 Tflops Peak Singe Precision Floating Point Performance 8.74 Tflops Memory Bandwidth (ECC Off) 480 GB/sec (240 GB/sec per GPU) Memory Size and Type 24GB GDDR5 ECC (12 GB per GPU) CUDA Cores 4992 (2496 per GPU) Thermal Solution Passive heatsink.

Key Features

Peak Double Precision Floating Point Performance 2.91 Tflops
Peak Singe Precision Floating Point Performance 8.74 Tflops
Memory Bandwidth (ECC Off) 480 GB/sec (240 GB/sec per GPU)
Memory Size and Type 24GB GDDR5 ECC (12 GB per GU)
ECC Default Setting Enabled (configurable by user)
ECC Usage Overhead 6.25% (~22.5 GB available)
Memory Clock 2.5 GHz
Number of GPUs 2x GK210
CUDA Cores 4992 (2496 per GPU)
Base Core Clock 560 MHz Boost Clocks 562 MHz to 875 MHz
Cache L1 and L2
GPU Interconnect On-board PLX switch System Interface PCI Express 3.0 x16