EVGA GeForce RTX 3090 Ti FTW3 ULTRA vs NVIDIA GeForce RTX 5090 D v2 Full Specs
10,752 Shaders 1.92GHz | 21,760 Shaders 2.41GHz |
24GB GDDR6X1.01TB/s | 24GB GDDR71.34TB/s |
·· 41.29 TFLOPS | ·· 104.8 TFLOPS |
Form Factor PCIe Card | Form Factor PCIe Card |
TDP 450W | TDP 575W |
Power Connectors 1x 16-Pin 12VHPWR | Power Connectors 1x 16-Pin 12VHPWR |
GeForce RTX 3090 Ti FTW3 ULTRA1.32 POPSINT4 Tensor Sparse
GeForce RTX 5090 D v22.38 PFLOPSFP4 Tensor SparseClock Speed ··· | Clock Speed ··· |
Peak OPS 1.32 POPSINT4 Tensor Sparse | Peak OPS 2.38 PFLOPSFP4 Tensor Sparse |
- | Tensor FP4 1.19 PFLOPSFP4 Tensor Sparse 2.38 PFLOPS |
- | Tensor FP8-16 593.8 TFLOPSFP8-16 Tensor Sparse 1.19 PFLOPSTensor FP8-32 296.9 TFLOPSFP8-32 Tensor Sparse 593.8 TFLOPS |
Tensor FP16-16 165.2 TFLOPSFP16-16 Tensor Sparse 330.3 TFLOPSTensor FP16-32 82.58 TFLOPSFP16-32 Tensor Sparse 165.2 TFLOPS | Tensor FP16-16 296.9 TFLOPSFP16-16 Tensor Sparse 593.8 TFLOPSTensor FP16-32 148.4 TFLOPSFP16-32 Tensor Sparse 296.9 TFLOPS |
BF16 41.29 TFLOPSTensor BF16 82.58 TFLOPSBF16 Tensor Sparse 165.2 TFLOPS | BF16 104.8 TFLOPSTensor BF16 148.4 TFLOPSBF16 Tensor Sparse 296.9 TFLOPS |
Tensor TF32 41.29 TFLOPS | Tensor TF32 74.22 TFLOPS |
FP32 41.29 TFLOPS | FP32 104.8 TFLOPS |
FP64 645.1 GFLOPS | FP64 1.64 TFLOPS |
Tensor INT4 660.6 TOPS | Tensor INT4 - |
Tensor INT8 330.3 TOPS | Tensor INT8 593.8 TOPS |
Ray 62.27 TOPS | Ray 317.7 TOPS |
Pixel Rate 215 GPixel/s | Pixel Rate 423.6 GPixel/s |
Texture Rate 645.1 GTexel/s | Texture Rate 1.64 TTexel/s |
Shaders 10,752 Shaders | Shaders 21,760 Shaders |
TMUs 336 TMUs | TMUs 680 TMUs |
ROPs 112 ROPs | ROPs 176 ROPs |
Tensor Cores 336 T-Cores | Tensor Cores 510 T-Cores |
RT Cores 84 RT-Cores | RT Cores 170 RT-Cores |
SMs 84 SMs | SMs 170 SMs |
Base Clock 1.56GHz | Base Clock 2.01GHz |
Boost Clock 1.92GHz | Boost Clock 2.41GHz |
Tensor Clock - | Tensor Clock 2.27GHz |
L2 Cache 6.1MB shared | L2 Cache 98.3MB shared |
24GB GDDR6X | 24GB GDDR7 |
Memory Bus 384-bit | Memory Bus 384-bit |
Memory Speed 21GT/s | Memory Speed 28GT/s |
Memory Bandwidth 1.01TB/s | Memory Bandwidth 1.34TB/s |
TDP 450W | TDP 575W |
Max Temp 92°C Max | Max Temp 90°C Max |
Multi-Monitor 4 | Multi-Monitor 4 |
HDCP HDCP 2.3 | HDCP HDCP 2.3 |
- | 3x DisplayPort 2.1 |
3x DisplayPort 1.4 | - |
1x HDMI 2.1 | 1x HDMI 2.1 |
Encoder Model NVENC 7 | Encoder Model 3x NVENC 9 |
Decoder Model NVDEC 5 | Decoder Model 2x NVDEC 6 |
Form Factor PCIe Card | Form Factor PCIe Card |
PCIe 3.4-Slots | PCIe 2.1-Slots |
Height 136.8 mm (5.39")Width 300 mm (11.81")Depth 70 mm (2.76") | Height 137 mm (5.39")Width 304 mm (11.97")Depth 42 mm (1.65") |
Cooling Open-Air 3x Fans | Cooling Open-Air 2x Fans |
Power Connectors 1x 16-Pin 12VHPWR | Power Connectors 1x 16-Pin 12VHPWR |
Manufacturer | Manufacturer |
Chip Designer | Chip Designer |
Architecture | Architecture |
Family | Family |
Branding  | Branding  |
Codename NV172 | Codename NV192 |
Chip Variant GA102-200-KD-A1 | Chip Variant GB202-250-A1 |
Market Segment Desktop | Market Segment Desktop |
Release Date Mar 30, 2022 | Release Date Jan 6, 2025 |
Foundry Samsung | Foundry TSMC |
Fabrication Node 8N | Fabrication Node 4NP |
Die Size 628mm² | Die Size 762mm² |
Transistor Count 28.3 Billion | Transistor Count 92.2 Billion |
Transistor Density 45.04 MTr/mm² | Transistor Density 121.1 MTr/mm² |
No images available
No images available