* Images are for reference only. See product specifications (datasheet).
Technical Documentation:
insert_drive_file SKY-TESL-L40-48P_L4-24P_DS(053023)20230530201321.pdf
Terms and Conditions
| NVIDIA GPU Cores | GPU Architecture | Ada Lovelace |
| Memory Size | 24 GB DDR6 (with ECC) | |
| CUDA Cores | 7680 | |
| Tensor Cores | 240 | |
| RT Cores | 60 | |
| Power Consumption | 72 W | |
| GPU Computing | Max FP32 (TFLOPS) | 31.3 |
| Max FP64 (TFLOPS) | – | |
| Multi-Instance GPU | – | |
| Design | Compute + Graphics | |
| Display & Graphics | NVENC | 2 |
| NVDEC | 4 | |
| Environment | Operating Temperature | 0~50 °C |
| Mechanical | Form Factor | Single slot, half height |
| Dimension (L x H) | 167.6 x 68.6 mm |
Data sheet
After placing your order, you will receive an email with a proforma invoice that details the most advantageous shipping cost. With us, you won't pay any hidden fees; we do not incorporate shipping costs into the product price. We will also provide you with an estimated delivery date. By confirming the proforma invoice, you give us the green light to dispatch your order and agree to our delivery terms.
We are proud certified partners of brands like Advantech, Westermo, Elproma, and Korenix. This ensures that you are purchasing quality and verified products directly from a trusted supplier. Visit the link to verify our collaboration directly on the official website.
Our team has extensive experience in designing and implementing large industrial projects. We have already collaborated with significant industrial enterprises, metro systems, and railway transport. If you are planning a larger project or need consultation, we are here for you! Our specialists are happy to assist you and prepare an offer tailored precisely to your needs.
Do you have questions or need quick assistance? Our customer service is ready to respond as quickly as possible and provide you with all the information you need.
We believe in the quality of our products and services. If for any reason you are not satisfied, we are happy to help you find a solution.
Tesla L40 48GB PCIe DSFH w/ATX BKT NVIDIA Ada-Lovelace GPU architecture. Universal Compute & Graphics GPU. 18176 NVIDIA® CUDA® Cores. 568 NVIDIA® Tensor Cores; 142 NVIDIA® RT Cores. 48GB with ECC memory. Up to 864 GB/s memory bandwidth. Max. power consumption: 300W. Graphics bus: PCIe 4.0 x16. Thermal solution: Passive.
Tesla L40S 48GB PCIe DSFH w/ATX BKT NVIDIA Ada-Lovelace GPU architecture. Universal Compute & Graphics GPU. 18176 NVIDIA® CUDA® Cores. 568 NVIDIA® Tensor Cores; 142 NVIDIA® RT Cores. 48GB with ECC memory. Up to 864 GB/s memory bandwidth. Max. power consumption: 350W. Graphics bus: PCIe 4.0 x16. SKY-TESL-L40S-48P.
Tesla H100 80GB PCIe DSFH w/ATX BKT NVIDIA Hopper GPU architecture. Compute-optimized GPU. 14592 NVIDIA® CUDA® Cores. 456 NVIDIA® Tensor Cores. 80GB HBM2e memory with ECC. Up to 2TB/s memory bandwidth. Max. power consumption: 350W. Graphics bus: PCI-E 5.0 x16. Thermal solution: Passive.
Tesla H100 NVL 94GB PCIe DSFH w/ATX BKT NVIDIA Hopper GPU architecture. Compute-optimized GPU. 14592 NVIDIA® CUDA® Cores. 456 NVIDIA® Tensor Cores. 94GB HBM2e memory with ECC. Up to 3.9TB/s memory bandwidth. Max. power consumption: 400W. Graphics bus: PCI-E 5.0 x16. Thermal solution: Passive.
Tesla L40 48GB PCIe DSFH w/ATX BKT NVIDIA Ada-Lovelace GPU architecture. Universal Compute & Graphics GPU. 18176 NVIDIA® CUDA® Cores. 568 NVIDIA® Tensor Cores; 142 NVIDIA® RT Cores. 48GB with ECC memory. Up to 864 GB/s memory bandwidth. Max. power consumption: 300W. Graphics bus: PCIe 4.0 x16. Thermal solution: Passive.
Tesla L40S 48GB PCIe DSFH w/ATX BKT NVIDIA Ada-Lovelace GPU architecture. Universal Compute & Graphics GPU. 18176 NVIDIA® CUDA® Cores. 568 NVIDIA® Tensor Cores; 142 NVIDIA® RT Cores. 48GB with ECC memory. Up to 864 GB/s memory bandwidth. Max. power consumption: 350W. Graphics bus: PCIe 4.0 x16. SKY-TESL-L40S-48P.
Tesla H100 80GB PCIe DSFH w/ATX BKT NVIDIA Hopper GPU architecture. Compute-optimized GPU. 14592 NVIDIA® CUDA® Cores. 456 NVIDIA® Tensor Cores. 80GB HBM2e memory with ECC. Up to 2TB/s memory bandwidth. Max. power consumption: 350W. Graphics bus: PCI-E 5.0 x16. Thermal solution: Passive.
Tesla H100 NVL 94GB PCIe DSFH w/ATX BKT NVIDIA Hopper GPU architecture. Compute-optimized GPU. 14592 NVIDIA® CUDA® Cores. 456 NVIDIA® Tensor Cores. 94GB HBM2e memory with ECC. Up to 3.9TB/s memory bandwidth. Max. power consumption: 400W. Graphics bus: PCI-E 5.0 x16. Thermal solution: Passive.
