| Revew | Lenovo NVIDIA RTX A1000 8GB GDDR6 Graphics CardThe Lenovo NVIDIA RTX A1000 (part number 4X61Q73041) is a professional workstation graphics card based on the NVIDIA Ampere architecture. It combines 2,304 CUDA cores, 72 third-generation Tensor cores and 18 second-generation RT cores with 8GB of GDDR6 memory with ECC support. The card has a single-slot low-profile design, draws up to 50W and takes all of its power from the PCIe slot, so it also fits compact workstations. Four Mini DisplayPort 1.4a outputs drive up to four displays, which makes it a good fit for CAD, design, content creation and multi-display work. 
Key features- NVIDIA Ampere architecture - 2,304 CUDA cores, 72 third-generation Tensor cores and 18 second-generation RT cores deliver 6.7 TFLOPS of single-precision performance, hardware ray tracing and acceleration for AI workloads.
- 8GB GDDR6 with ECC - 8GB of memory on a 128-bit interface with 192GB/s of bandwidth, plus ECC error correction that protects data integrity in professional work.
- Compact and efficient - A single-slot low-profile card measuring 2.7 x 6.4 inches, with active cooling and a maximum power draw of 50W. No auxiliary power cable is needed.
- Up to four displays - Four Mini DisplayPort 1.4a outputs with audio, HDCP 2.2 and NVIDIA Mosaic support. Drive four 4K displays at 120Hz, four 5K displays at 60Hz or two 8K displays at 30Hz.
- Video encode and decode - One encode engine and two decode engines, including hardware AV1 decode, for efficient video work and streaming.
- Ready for professional software - Supports DirectX 12, OpenGL 4.6, Vulkan 1.3, CUDA and OpenCL 3.0. NVIDIA RTX professional graphics cards are certified for a broad range of professional applications.
Specifications| Manufacturer | Lenovo | | Part number | 4X61Q73041 | | GPU | NVIDIA RTX A1000 | | Architecture | NVIDIA Ampere | | Cores | 2,304 CUDA, 72 Tensor (3rd gen), 18 RT (2nd gen) | | Performance | 6.7 TFLOPS FP32, 13.2 TFLOPS RT, 53.8 TFLOPS Tensor (FP16 with sparsity) | | Memory | 8GB GDDR6 with ECC | | Memory interface | 128-bit, 192GB/s bandwidth | | Interface | PCI Express 4.0 x8 | | Outputs | 4 x Mini DisplayPort 1.4a | | Display support | 4 x 4096x2160 at 120Hz, 4 x 5120x2880 at 60Hz or 2 x 7680x4320 at 30Hz | | Encode / decode | 1 encode engine, 2 decode engines, AV1 decode | | APIs | DirectX 12, Shader Model 6.6, OpenGL 4.6, Vulkan 1.3, CUDA, DirectCompute, OpenCL 3.0 | | Power consumption | Up to 50W, no auxiliary power connector | | Cooling | Active | | Form factor | Low profile, single slot, 2.7 x 6.4 in | | NVLink | Not supported |
|
