NVIDIA GeForce RTX 5090
Tensor Cores : 680
Ray Tracing Cores : 170
Streaming Multiprocessors : 170
Texture Mapping Unit : 680
Render Output Units : 176
Boost Clock : 2407 MHz
Memory Clock : 3500 MHz
Effective Memory Speed : 28 Gbps
Memory Bus Width : 512-bit
Memory Bandwidth : 1,792 GB/s
The TSMC N4P manufacturing process allows for 92.2 billion transistors on a 750 mm² die. To draw everything on your screen, the GPU uses 21,760 CUDA Cores, 680 Texture Mapping Units, and 176 Render Output Units. It also has 170 Ray Tracing Cores for realistic lighting and reflections, plus 680 Tensor Cores to accelerate AI features.
To handle high-resolution textures, the NVIDIA GeForce RTX 5090 comes with 32 GB of GDDR7. Its 512-bit bus delivers 1,792 GB/s of memory bandwidth, ensuring the GPU stays fed at high settings. The Founders Edition boost clock is up to 2,407 MHz.
Blackwell upgrades the graphics pipeline with fourth-gen RT Cores (Mega Geometry), fifth-gen Tensor Cores with DLSS 4 Multi-Frame Generation, Raytracing (RT) Cores, and an AI (Tensor) Processor to coordinate on-GPU AI workloads. For creators, ninth-gen NVENC adds AV1 quality and speed gains and supports up to three encoders and two decoders.
Before you buy, check your case specs. The Founders Edition is 304 mm long and uses a 2-slot design. Display outputs are 3× DisplayPort 2.1b and 1× HDMI 2.1b, with support up to 4K 480 Hz or 8K 165 Hz via DSC (monitor-dependent). Power is via one 16-pin PCIe Gen5 cable (adapter included). The card’s total graphics power is 575 W and a 1000 W PSU is recommended.
| SKU ID |
SKU-166
|
|---|---|
| Product Type | Graphic Cards |
| Price (excl. tax) | NPR0.00 |
| Availability | Out of stock |
| Launch | 2025 |
| Segment | Dedicated Graphics Card |
| Brand | NVIDIA |
| Series | NVIDIA GeForce 50 |
| Model | RTX 5090 |
| GPU Model | GB202-300 (GB202) |
| Architecture | Blackwell |
| Fabrication Process | 4 nm (TSMC N4P) |
| Die Size | 750 mm² |
| Transistor Count | 92.2 Billion |
| Transistor Density | 122.9M TRAN/mm² |
| GPU |
CUDA Cores : 21760 Tensor Cores : 680 Ray Tracing Cores : 170 Streaming Multiprocessors : 170 Texture Mapping Unit : 680 Render Output Units : 176 |
| Clock Speeds |
Base Clock : 2017 MHz Boost Clock : 2407 MHz Memory Clock : 3500 MHz Effective Memory Speed : 28 Gbps |
| Memory Configuration |
Graphics RAM : 32768 MB / 32 GB GDDR7 Memory Bus Width : 512-bit Memory Bandwidth : 1,792 GB/s |
| Display Outputs |
3 × DisplayPort 2.1b 1 × HDMI 2.1 |
| API Support |
DirectX 12.2 Vulkan 1.3 OpenGL 4.6 OpenCL 3.0 |
| Performance |
Pixel Fillrate : 423.6 GPixels/s Texture Fillrate : 1.6 TTexel/s Peak FP32 : 104.8 TFLOPS FP32 Perf. per Watt : 182.2 GFLOPS/W FP32 Perf. per mm² : 139.7 GFLOPS/mm² |
| Power Connectors | 1× 16-pin |
| TDP/TBP | 575 W |
| Warranty | 1 Year |
Customer Reviews
No reviews yet.
Be the first to write one