NVIDIA RTX 5090
The NVIDIA RTX 5090 is a consumer graphics card based on the Blackwell architecture. Released on 2025-01-30 with a launch price of $1,999.00, it features 21760 CUDA cores and 32 GB of GDDR7 memory.
General info about RTX 5090
| Architecture | Blackwell (GB102) |
|---|---|
| Market Segment | Consumer |
| Release Date | 2025-01-30 |
| Launch Price | $1,999.00 |
| Manufacturing Process | 4nm |
Technical specs of RTX 5090
| CUDA Cores | 21760 |
|---|---|
| Base Clock Speed | 1300 MHz |
| Boost Clock Speed | 2600 MHz |
| Transistor Count | 100B |
| VRAM Capacity | 32 GB GDDR7 |
| Memory Bus Width | 384 bits |
| Memory Bandwidth | 1152.0 GB/s |
| TDP | 575 W |
Key technical parameters of the NVIDIA RTX 5090 include its 21760 CUDA cores, 1300 MHz base clock, and 2600 MHz boost clock, delivering high performance for consumer applications.
Benchmark performance of RTX 5090
| FP32 (float) | 55 TFLOPS |
|---|---|
| FP16 (half) | 110 TFLOPS |
The NVIDIA RTX 5090 achieves 55 TFLOPS in FP32 performance, making it a top contender in consumer workloads.
Gaming performance of RTX 5090
| 1080p | 300 FPS |
|---|---|
| 4K | 140 FPS |
| 8K | 60 FPS |
In gaming, the NVIDIA RTX 5090 delivers 300 FPS at 1080p and 140 FPS at 4K, showcasing its capability for high-resolution gaming.
Features and connectivity of RTX 5090
| Tensor_cores | Gen 5 |
|---|---|
| Dlss | 4.0 |
| Ports | HDMI 2.1, DisplayPort 2.0 x3 |
Physical dimensions of RTX 5090
| Length | 11in |
|---|---|
| Width | 4.5in |
| Height | 3-slot |
Rankings and efficiency of RTX 5090
| Ranking Position | #2 |
|---|---|
| Popularity Ranking | #10 |
| Cost Effectiveness | 1.6/5 |
| Power Efficiency | 1.7/5 |
Top GPUs for Training and Inference
| Category | Rank 1 | Rank 2 | Rank 3 |
|---|---|---|---|
| Best for Training | NVIDIA H200 | NVIDIA H100 | NVIDIA B200 |
| Best for Inference | NVIDIA A40 | NVIDIA A100 | NVIDIA A10 |