(Bundle Sale) NVIDIA® H100 PCIe cards is compute-optimized GPU built on the NVIDIA Hopper architecture with dual-slot 10.5-inch PCI Express Gen5 interface in a passive heatsink cooling design suitable for data centers (Part Number: SKY-TESL-H100-80P).

(Bundle Sale) NVIDIA® A100 80GB PCIe cards are compute-optimized GPUs built on the NVIDIA Ampere architecture with dual-slot 10.5-inch PCI Express Gen4 interface in a passive heatsink cooling design suitable for data centers (Part Number: SKY-TESL-A100-80P).

(Bundle Sale) NVIDIA® A30 24GB PCIe cards are compute-optimized GPUs built on the NVIDIA Ampere architecture with dual-slot 10.5-inch PCI Express Gen4 interface in a passive heatsink cooling design suitable for data centers (Part Number: SKY-TESL-A30-24P).

(Bundle Sale) NVIDIA® A40 PCIe cards are computing and graphics GPUs built on the NVIDIA Ampere architecture with PCI Express Gen4 interface in a passive heatsink cooling design suitable for mainstream enterprise and visual computing servers (Part Number: SKY-TESL-A40-48P).

(Bundle Sale) NVIDIA® A10 PCIe cards are computing and graphics GPUs built on the NVIDIA Ampere architecture with PCI Express Gen4 interface in a passive heatsink cooling design suitable for mainstream enterprise and visual computing servers (Part Number: SKY-TESL-A10-24P).

(Bundle Sale) NVIDIA® A2 PCIe cards are computing and graphics GPUs built on the NVIDIA Ampere architecture with PCI Express Gen4 interface in a passive heatsink cooling design suitable for mainstream enterprise and visual computing servers (Part Number: SKY-TESL-A2-16P).

Графический процессор NVIDIA® T4 ускоряет облачные рабочие нагрузки, основан на новой архитектуре NVIDIA Turing™ и упакован в энергоэффективный 70-ваттный компактный форм-фактор PCIe, оптимизирован для основных вычислительных сред и поддерживает многоточность Turing Turing Tensor и новые ядра RT.

NVIDIA® Tesla® V100 Tensor Core — передовой графический процессор для ЦОД, созданный для ускорения ИИ, высокопроизводительных вычислений (HPC), обработки данных и графики. Основан на архитектуре NVIDIA Volta и обеспечивает производительность до 100 процессоров в одном графическом процессоре.