Database Mart NVIDIA GPU List for Server Workloads
Our GPU servers power AI and machine learning, high-performance computing, 3D rendering, and cloud gaming. This page is Database Mart’s reference NVIDIA GPU list: compare CUDA cores, VRAM, and architecture before you deploy on bare metal in US data centers. Use the NVIDIA graphics cards list table below for workstation and datacenter rows, then check the GPU price list on our server pages when you are ready to order.
A GPU List Built for Server Decisions
The table below is our reference NVIDIA graphics cards list for hardware offered on Database Mart GPU servers. It is not a shopping cart—use this GPU list with engineering before you pick an instance. Datacenter and workstation rows highlight NVIDIA server GPU list options (Tesla, H-series, L4, RTX A/Pro); GeForce rows cover cost-sensitive inference and graphics. For GPU price list and NVIDIA graphics card price list amounts, use live listings on GPU server hosting or contact us—SKUs and monthly rates change with supply.
| Model | Type | Launch Time | Architecture | CUDA Cores | Tensor Cores | GPU VRAM | vGPU Support | NVLink Support | FP32 Performance | Power Consumption |
|---|---|---|---|---|---|---|---|---|---|---|
| GT 710 | Desktop | Mar. 2014 | Kepler | 192 | No | 2GB DDR3 | No | No | 0.3 TFLOPS | 19W |
| GT 730 | Desktop | Jun. 2014 | Kepler | 384 | No | 2GB DDR3 | No | No | 0.7 TFLOPS | 25W |
| GTX 1060 | Desktop | Dec. 2016 | Pascal | 1152 | No | 6GB GDDR5 | No | No | 3.9 TFLOPS | 120W |
| GTX 1080 | Desktop | May. 2016 | Pascal | 2560 | No | 8GB GDDR5X | No | No | 8.9 TFLOPS | 180W |
| GTX 1650 | Desktop | Apr. 2019 | Turing | 896 | No | 4GB GDDR6 | No | No | 3.0 TFLOPS | 75W |
| GTX 1660 | Desktop | Mar. 2019 | Turing | 1408 | No | 6GB GDDR6 | No | No | 5.0 TFLOPS | 120W |
| RTX 2060 | Desktop | Jan. 2019 | Turing | 1920 | 240 | 6GB GDDR6 | No | No | 6.5 TFLOPS | 160W |
| RTX 2080 | Desktop | Sep. 2018 | Turing | 2944 | 368 | 8GB GDDR6 | No | Yes | 10.1 TFLOPS | 215W |
| RTX 3050 | Desktop | Feb. 2024 | Ampere | 2304 | 72 | 8GB GDDR6 | No | No | 6.8 TFLOPS | 70W |
| RTX 3060 | Desktop | Feb. 2021 | Ampere | 3584 | 112 | 12GB GDDR6 | No | No | 12.7 TFLOPS | 170W |
| RTX 3060 TI | Desktop | Dec. 2020 | Ampere | 4864 | 152 | 8GB GDDR6 | No | No | 16.2 TFLOPS | 200W |
| RTX 3090 | Desktop | Sep. 2020 | Ampere | 10496 | 328 | 24GB GDDR6X | No | Yes | 35.6 TFLOPS | 350W |
| RTX 4060 | Desktop | Jun 2023 | Ada Lovelace | 3072 | 96 | 8GB GDDR6 | No | No | 15.1 TFLOPS | 115W |
| RTX 4060 TI | Desktop | May 2023 | Ada Lovelace | 4352 | 136 | 8GB GDDR6 | No | No | 20.1 TFLOPS | 160W |
| RTX 4080 | Desktop | Nov. 2022 | Ada Lovelace | 9728 | 304 | 16GB GDDR6X | No | No | 48.8 TFLOPS | 320W |
| RTX 4090 | Desktop | Oct. 2022 | Ada Lovelace | 16384 | 512 | 24GB GDDR6X | No | Yes | 82.6 TFLOPS | 450W |
| RTX 5060 | Desktop | May 2025 | Blackwell | 3840 | 120 | 8GB GDDR7 | No | No | 19.2 TFLOPS | 145W |
| RTX 5090 | Desktop | May 2025 | Blackwell | 21760 | 680 | 32GB GDDR7 | No | No | 104.8 TFLOPS | 575W |
| Quadro K620 | Workstation | Jul. 2014 | Maxwell (1st Gen) | 384 | No | 2GB GDDR5 | No | No | 0.77 TFLOPS | 45W |
| Quadro K1200 | Workstation | Jan. 2015 | Maxwell (1st Gen) | 512 | No | 4GB GDDR5 | No | No | 1.08 TFLOPS | 45W |
| Quadro K2200 | Workstation | Jul. 2014 | Maxwell (1st Gen) | 640 | No | 4GB GDDR5 | No | No | 1.34 TFLOPS | 68W |
| Quadro K4000 | Workstation | Mar. 2013 | Kepler | 768 | No | 3GB GDDR5 | No | No | 2.17 TFLOPS | 80W |
| Quadro M2000 | Workstation | Apr. 2016 | Maxwell (2nd Gen) | 768 | No | 4GB GDDR5 | No | No | 1.8 TFLOPS | 75W |
| Quadro M4000 | Workstation | Jun. 2015 | Maxwell (2nd Gen) | 2048 | No | 8GB GDDR5 | No | No | 4.3 TFLOPS | 150W |
| Quadro M6000 | Workstation | Mar. 2015 | Maxwell (2nd Gen) | 3072 | No | 24GB GDDR5 | No | No | 6.1 TFLOPS | 250W |
| Quadro P600 | Workstation | Feb. 2017 | Pascal | 384 | No | 2GB GDDR5 | No | No | 1.2 TFLOPS | 40W |
| Quadro P620 | Workstation | Feb. 2018 | Pascal | 512 | No | 2GB GDDR5 | No | No | 1.5 TFLOPS | 40W |
| Quadro P1000 | Workstation | Feb. 2017 | Pascal | 640 | No | 4GB GDDR5 | No | No | 1.9 TFLOPS | 47W |
| Quadro T1000 | Workstation | May 2021 | Turing | 896 | No | 8GB GDDR6 | No | No | 2.5 TFLOPS | 50W |
| Quadro P4000 | Workstation | Feb. 2017 | Pascal | 1792 | No | 8GB GDDR5 | No | No | 5.3 TFLOPS | 105W |
| Quadro P5000 | Workstation | Oct. 2016 | Pascal | 2560 | No | 16GB GDDR5X | No | No | 8.9 TFLOPS | 180W |
| RTX A400 | Workstation | Apr. 2024 | Ampere | 768 | 24 | 4GB GDDR6 | No | No | 2.7 TFLOPS | 50W |
| RTX A1000 | Workstation | Apr. 2024 | Ampere | 2304 | 72 | 6GB GDDR6 | No | No | 6.8 TFLOPS | 50W |
| RTX A4000 | Workstation | Apr. 2021 | Ampere | 6144 | 192 | 16GB GDDR6 | Yes | No | 19.2 TFLOPS | 140W |
| RTX A5000 | Workstation | Apr. 2021 | Ampere | 8192 | 256 | 24GB GDDR6 | Yes | Yes | 27.8 TFLOPS | 230W |
| RTX A5500 | Workstation | Mar. 2022 | Ampere | 10240 | 320 | 24GB GDDR6 | Yes | No | 34.1 TFLOPS | 230W |
| RTX A6000 | Workstation | Apr. 2021 | Ampere | 10752 | 336 | 48GB GDDR6 | Yes | Yes | 38.7 TFLOPS | 300W |
| RTX Pro 2000 | Workstation | Aug. 2018 | Turing | 1024 | 128 | 4 GB GDDR5 | Yes | No | 3.0 TFLOPS | 75W |
| RTX Pro 4000 | Workstation | Mar. 2025 | Blackwell | 8960 | 280 | 24GB GDDR7 | Yes | No | 46.9 TFLOPS | 140W |
| RTX Pro 4500 | Workstation | Mar. 2025 | Blackwell | 10496 | 328 | 32GB GDDR7 | Yes | No | 54.9 TFLOPS | 200W |
| RTX Pro 5000 | Workstation | Mar. 2025 | Blackwell | 14080 | 440 | 48GB GDDR7 | Yes | No | 73.7 TFLOPS | 300W |
| RTX Pro 6000 | Workstation | Mar. 2025 | Blackwell | 24064 | 752 | 96GB GDDR7 | Yes | Yes | 110 TFLOPS | 600W |
| Tesla K40 | Datacenter | Oct. 2013 | Kepler | 2880 | No | 12GB GDDR5 | No | No | 5.0 TFLOPS | 235W |
| Tesla M40 | Datacenter | Nov. 2015 | Maxwell | 3072 | No | 24GB GDDR5 | No | No | 8.7 TFLOPS | 235W |
| Tesla K80 | Datacenter | Nov. 2014 | Kepler | 4992 | No | 24GB GDDR5 | No | Yes | 6.8 TFLOPS | 300W |
| Tesla P40 | Datacenter | Sep. 2016 | Pascal | 3840 | No | 24GB GDDR5 | No | No | 11.8 TFLOPS | 250W |
| Tesla P100 | Datacenter | Jun. 2016 | Pascal | 3584 | No | 16GB HBM2 | Yes | Yes | 9.3 TFLOPS | 250W |
| Tesla V100 | Datacenter | Jun. 2017 | Volta | 5120 | 640 | 16GB HBM2 | Yes | Yes | 13.2 TFLOPS | 250W |
| Tesla T4 | Datacenter | Sep. 2018 | Turing | 2560 | 320 | 16GB GDDR6 | Yes | No | 8.1 TFLOPS | 70W |
| Tesla A30 | Datacenter | Apr. 2021 | Ampere | 3584 | 224 | 24GB HBM2 | Yes | Yes | 10.3 TFLOPS | 250W |
| Tesla A40 | Datacenter | Oct. 2020 | Ampere | 10752 | 336 | 48GB GDDR6 | Yes | Yes | 37.4 TFLOPS | 300W |
| Tesla A100 | Datacenter | May. 2020 | Ampere | 6912 | 432 | 40GB HBM2e | Yes | Yes | 19.5 TFLOPS | 250W |
| Tesla H100 | Datacenter | Mar. 2022 | Hopper | 14592 | 456 | 80GB HBM3 | Yes | Yes | 51.2 TFLOPS | 350W |
| Tesla H200 | Datacenter | Nov. 2024 | Hopper | 16896 | 528 | 141GB HBM3e | Yes | Yes | 60.3 TFLOPS | 600W |
| NVIDIA L4 | Datacenter | Mar. 2023 | Ada Lovelace | 7424 | 232 | 24GB GDDR6 | Yes | No | 30.3 TFLOPS | 72W |
Note for production: verify Tesla H200 VRAM column against NVIDIA official specs before publish; live site previously showed 41GB.
Explanation of Key GPU Parameters
Each row in our NVIDIA GPU list uses the fields below. Reading them together turns the NVIDIA graphics cards list from a flat table into a shortlist you can defend in architecture reviews.
Type: Refers to main use and target users. GeForce is for gaming and everyday graphics; Quadro and RTX A are for professional 3D design and rendering; Tesla and H-series entries belong on the NVIDIA server GPU list for AI, HPC, and data centers.
Architectuer: Represents the underlying chip design generation (Kepler through Blackwell). Newer architectures improve efficiency and AI features when you map an NVIDIA GPU list row to production.
CUDA Cores: Determines parallel processing power, essential for AI training and complex computations across the NVIDIA graphics cards list.
Tensor Cores: Boosts AI computations and deep learning efficiency; prioritize these when you filter the NVIDIA GPU list for training or inference.
VRAM (Video Memory): Affects handling of large datasets; critical for deep learning and video rendering when you scan the GPU list.
Memory Bandwidth: Impacts data transfer speed between GPU and memory, important for high-performance computing on NVIDIA server GPU list hardware.
FP32 Performance(TFLOPS): Measures floating-point computing power, crucial for machine learning and simulations when ranking the NVIDIA GPU list.
Power Consumption (Wattage): Important for energy efficiency and infrastructure planning on dedicated GPU servers tied to this NVIDIA graphics cards list.
FAQs About Our NVIDIA GPU List and GPU List
Common questions about how to read this NVIDIA graphics cards list, where the GPU price list appears, and how the NVIDIA server GPU list differs from retail specs—before you contact sales.
Is this the full NVIDIA GPU list Database Mart offers?
The table reflects GPUs we deploy or offer on GPU servers; availability varies. Confirm stock before you treat any row as a guaranteed NVIDIA GPU list entry.
How is a server GPU list different from a retail card list?
Our NVIDIA server GPU list covers cards on dedicated servers with remote access and US data centers, not retail boxes alone.
How often does Database Mart update the NVIDIA GPU list?
We add or retire SKUs as inventory changes; refresh the NVIDIA GPU list table when you plan a new deployment.
Where do I see the GPU price list?
On GPU server hosting and related pages—the NVIDIA graphics card price list there shows current monthly or hourly rates, not this reference NVIDIA graphics cards list.
Can I filter the GPU list for AI only?
Focus on Tensor Cores, VRAM, and datacenter Type rows in the GPU list (Tesla A/H, L4). For sizing help, see AI hosting or contact us.
Who should use this NVIDIA graphics cards list?
Engineers comparing the NVIDIA GPU list before sales quotes, buyers cross-checking the NVIDIA server GPU list, and partners who need a canonical NVIDIA graphics cards list link instead of scattered PDFs.
Questions About Our NVIDIA GPU List?
Share workload and budget—we will map you to the right NVIDIA graphics cards list entry and point you to the current GPU price list without upselling SKUs you do not need.