Tesla V100S PCI-E Nvidia Graphic Card is a dual-slot 10.5 inch PCI Express Gen3 card with a single NVIDIA Volta GV100 graphics processing unit (GPU). It uses a passive heat sink for cooling, which requires system air flow to properly operate the card within its thermal limits. The Tesla V100 PCIe supports double precision (FP64), single precision (FP32) and half precision (FP16) compute tasks, unified virtual memory and page migration engine. For performance optimization, NVIDIA GPU Boost™ feature is supported. By automatically adjusting the GPU clock dynamically, maximum performance is achieved within the power cap limit. Tesla V100 PCIe boards are shipped with ECC enabled by default to protect the GPU’s memory interface and the on-board memories. ECC protects the memory interface by detecting any single, double, and all odd-bit errors. The GPU will retry any memory transaction that has an ECC error until the data transfer is error-free. ECC protects the DRAM content by fixing any single-bit errors and detecting double-bit errors. The Tesla V100 PCIe with 16GB of HBM2 memory has native support for ECC and has no ECC overhead, both in memory capacity and bandwidth. Tesla V100 PCIe supports Maximum Performance (Max-P) and Maximum Efficiency (Max-Q) modes. In Max-P mode, the Tesla V100 PCIe Accelerator will operate unconstrained up to its thermal design power (TDP) level of 250 W to accelerate applications that require the fastest computational speed and highest data throughput. Max-Q mode allows data center managers to tune power usage of their Tesla V100 PCIe Accelerators to operate with optimal performance per watt. A power limit can be set via software across all GPUs in a rack, reducing power consumption dramatically, while still obtaining excellent rack performance for target applications. Max-Q is not tied to a specific power number. The data center manager can set the Tesla V100 to a power budget (as long as it is below 250 W) that delivers the best perf/watt for the target Overview Tesla V100 PCIe GPU Accelerator PB-08744-001_v03 | 2 workload. Max-Q gives the data center manager the flexibility to optimize the throughput at a node-level, rack-level or data center-level based on the power budget.

Product Specifications

Product SKUs                                NVPN: 699-2G500-0200-XXX

Total board power                        250 W

GPU SKUs                                      GV100-893-A1

Maximum memory clock        877 MHz

Memory size                                16 GB HBM2

Memory bus width                    4096-bit

Peak memory bandwidth        Up to 900 GB/

Tesla V100S PCI-E Nvidia Graphic Card