NVIDIA GeForce RTX 4090 Max-Q Full Specs
9,728 Shaders 1.46GHz |
16GB GDDR6576GB/s |
·· 28.31 TFLOPS |
Form Factor Soldered |
TDP 80W |
Clock Speed ··· |
Peak OPS 905.9 TOPSINT4 Tensor Sparse |
Tensor FP8-16 226.5 TFLOPSFP8-16 Tensor Sparse 452.9 TFLOPSTensor FP8-32 113.2 TFLOPSFP8-32 Tensor Sparse 226.5 TFLOPS |
Tensor FP16-16 113.2 TFLOPSFP16-16 Tensor Sparse 226.5 TFLOPSTensor FP16-32 56.62 TFLOPSFP16-32 Tensor Sparse 113.2 TFLOPS |
BF16 28.31 TFLOPSTensor BF16 56.62 TFLOPSBF16 Tensor Sparse 113.2 TFLOPS |
Tensor TF32 28.31 TFLOPS |
FP32 28.31 TFLOPS |
FP64 442.3 GFLOPS |
Tensor INT4 452.9 TOPS |
Tensor INT8 226.5 TOPS |
Ray 65.43 TOPS |
Pixel Rate 163 GPixel/s |
Texture Rate 442.3 GTexel/s |
Shaders 9,728 Shaders |
TMUs 304 TMUs |
ROPs 112 ROPs |
Tensor Cores 304 T-Cores |
RT Cores 76 RT-Cores |
SMs 76 SMs |
Base Clock 930MHz |
Boost Clock 1.46GHz |
L2 Cache 65.5MB shared |
16GB GDDR6 |
Memory Bus 256-bit |
Memory Speed 18GT/s |
Memory Bandwidth 576GB/s |
ECC No |
TDP 80W |
Multi-Monitor 4 |
HDCP HDCP 2.3 |
Encoder Model 2x NVENC 8 |
Decoder Model NVDEC 5 |
Form Factor Soldered |
Manufacturer |
Chip Designer |
Architecture |
Family |
Branding  |
Codename NV183 |
Chip Variant AD103-275-A1 |
Market Segment Laptop |
Release Date Feb 8, 2023 |
Foundry TSMC |
Fabrication Node 4N |
Die Size 379mm² |
Transistor Count 45.9 Billion |
Transistor Density 121.2 MTr/mm² |
No images available
