Product | NVIDIA® A2 GPU Computing Accelerator - 16GB GDDR6 - PCIe 4.0 x8 - Passive Cooler | NVIDIA® A10 GPU Computing Accelerator - 24GB GDDR6 - PCIe 4.0 x16 - Passive Cooler (w/o CEC) | NVIDIA® A16 GPU Computing Accelerator - 64GB (4x 16GB) GDDR6 - PCIe 4.0 x16 - Passive Cooler | NVIDIA® A40 GPU Computing Accelerator - 48GB GDDR6 - PCIe 4.0 x16 - Passive Cooling |
Action | Select | Select | Select | Select |
Main Specifications | ||||
Product Series | Nvidia A2 | Nvidia A10 | Nvidia A16 | Nvidia A40 |
Core Type | NVIDIA TENSOR | NVIDIA TENSOR | NVIDIA TENSOR | NVIDIA TENSOR |
Core Clock Speed | 1440 MHz (1770 MHz Boost Clock) | 885 MHz (1695 MHz Boost Clock) | ||
Host Interface | PCI Express 4.0 x8 | PCI Express 4.0 x16 64GB/s | PCI Express 4.0 x16 | PCI Express 4.0 x16 |
GPU Architecture | Ampere | Ampere | Ampere | Ampere |
Detailed Specifications | ||||
Streaming Processor Cores | 1280 CUDA Cores | 10752 CUDA Cores | ||
NVIDIA Tensor Cores | 40 | Gen 3 | 336 Tensor Cores | ||
NVIDIA RT Cores | 10 | Gen 2 | 72 RT Cores | 84 RT Cores | |
Memory Clock Speed | 6251 MHz | 1563 MHz | ||
Memory Interface | 128-bit | 384-bit | ||
Memory Speeds (GT/s) | 14.5Gbps GDDR6 | |||
Max Memory Size | 16 GB GDDR6 ECC | 24 GB GDDR6 | 4x 16GB GDDR6 with error-correcting code (ECC) | 48 GB GDDR6 with error-correcting code (ECC) |
Max Memory Bandwidth | 200 GB/s | 600 GB/s | 4x 232GB/s | 696 GB/s |
INT8 Tensor Core | 250 TOPS | 500 TOPS | |||
TF32 Tensor Core | 9 TFLOPS | 18 TFLOPS Sparsity | 62.5 teraFLOPS | 125 teraFLOPS | ||
FP32 | 4.5 TFLOPS | 31.2 teraFLOPS | ||
Peak BFLOAT16 Tensor Core | 125 teraFLOPS | 250 teraFLOPS | |||
Peak FP16 Tensor Core | 18 TFLOPS | 36 TFLOPS Sparsity | 125 teraFLOPS | 250 teraFLOPS | ||
Peak INT4 Tensor Core | 500 TOPS | 1,000 TOPS | |||
Total NVLink Bandwidth | NVIDIA NVLink 112.5 GB/s (bidirectional) PCIe Gen4 16 GB/s | |||
NVIDIA CUDA™ Technology | 11.1 or later | |||
Peak INT4 Performance | 72 TOPS | 144 TOPS Sparsity | |||
Peak INT8 Performance | 36 TOPS | 72 TOPS Sparsity | |||
ECC Protection | On by Default | |||
DisplayPort Connectors | 3x DisplayPort 1.4 A40 is configured for virtualization by default with physical display connectors disabled. The display outputs can be enabled via management software tools. | |||
Cooling | Passive | Passive | Passive | Passive |
Dual Slot | Single-slot | Single-slot | Dual-slot | 2-slot Low-profile |
Dimensions | 6.61” L x 2.71” H | FHFL | 4.4" (H) x 10.5" (L) | |
Form Factor | Low-Profile PCIe | |||
Lithography | 8 nm | Samsung 8nm | ||
Supplementary Power Connectors | None | 8-pin CPU | 1x 8-pin CPU (EPS12V) | |
Max Graphics Card Power (W) | 40-60 W | Configurable | 150W | 250W | 300W |
Action | Select | Select | Select | Select |