PNY NVIDIA L4
SKU: NVBGPUNTL4
NVIDIA’s AI platform and full-stack approach make L4 the best GPU for video and inference at scale. It can handle a wide range of AI applications, such as recommendations, voice assistants, generative AI, visual search, and contact center automation. L4 is the most efficient NVIDIA accelerator for mainstream servers. It can boost AI Video performance by up to 120 times compared to CPUs. It can also improve generative AI performance by 2.7 times and graphics performance by over 4 times compared to the previous GPU generation. L4 is versatile and energy-efficient. It has a single-slot, low-profile design that fits in any server and any location, including the edge.
AI and video are growing fast, and so is the need for efficient, affordable computing. NVIDIA L4 GPUs can save up to 99% of energy and cost compared to CPU-based infrastructure. This means less rack space, less carbon footprint, and more scalability for enterprises. Switching from CPUs to NVIDIA L4s in a 2 MW data centre can power over 2,000 homes for a year or offset the carbon from 172,000 trees grown over 10 years*.
Available on back-order
Quantity:
Quantity:
System Configuration
Available on back-order
Details & Specifications
Details & Specifications
Manufacturer | |
---|---|
Manufacturer Part Number |
NVL4TCGPU-KIT |
Additional:
Ambient Operating Temperature |
0 to 50 Degrees Centigrade |
---|---|
Display Connectors |
None | vGPU Only |
NVENC | NVDEC |
2 | 4 | 4| JPEG Decoders | AV1 Encode and Decode |
Operating Humidity |
5% to 85% Relative Humidity |
Cooling:
Hardware Root of Trust |
Yes |
---|---|
Server Options |
Partner and NVIDIA-Certified Systems with 1-8 GPUs |
Dimensions:
Form Factor |
6.61” L x 2.71” H | Single Slot | Low-Profile |
---|---|
NEBS ready |
Yes | Level 3 |
GPU:
Game Clock |
795 MHz Base | 2040 MHz Boost |
---|---|
GPU Architecture | |
GPU Slot Size | |
Lithography |
4nm |
Peak BFLOAT16 Tensor Core Performance (TFLOPS) |
242 | Sparsity |
Peak FP16 Tensor Core Performance (TFLOPS) |
242 | Sparsity |
Peak FP32 Performance (TFLOPS) |
30.2 |
Peak FP32 Tensor Core Performance (TFLOPS) |
120 | Sparsity |
Peak FP8 Tensor Core Performance (TFLOPS) |
485 | Sparsity |
Peak INT8 Tensor Core Performance (TOPS) |
485 | Sparsity |
Memory:
Memory Bandwidth (GB/s) |
300 |
---|---|
Memory Capacity (GB) |
24 |
Memory Frequency (MHz) |
6251 |
Memory Interface (bit) |
192 |
Memory Type |
ECC |
PCIe:
PCIe Version |
PCIe 4.0 |
---|
Software:
Supported OS |
Linux, Windows |
---|
Warranty:
Mfr Warranty Length (M) |
36 |
---|---|
Mfr Warranty Type |
Direct To Manufacturer |
* Specifications for older discontinued variants can be view by clicking here