NVIDIA GeForce RTX 5070 vs ZOTAC GeForce RTX 3080 Ti AMP Holo Full Specs
6,144 Shaders 2.51GHz | 10,240 Shaders 1.71GHz |
12GB GDDR7672GB/s | 12GB GDDR6X936.2GB/s |
·· 30.87 TFLOPS | ·· 35.02 TFLOPS |
Form Factor PCIe Card | Form Factor PCIe Card |
TDP 250W | TDP 350W |
Power Connectors 1x 16-Pin 12VHPWR - | Power Connectors - 2x 8-Pin |
GeForce RTX 5070987.8 TFLOPSFP4 Tensor Sparse
GeForce RTX 3080 Ti AMP Holo1.12 POPSINT4 Tensor SparseClock Speed ··· | Clock Speed ··· |
Peak OPS 987.8 TFLOPSFP4 Tensor Sparse | Peak OPS 1.12 POPSINT4 Tensor Sparse |
Tensor FP4 493.9 TFLOPSFP4 Tensor Sparse 987.8 TFLOPS | - |
Tensor FP8-16 246.9 TFLOPSFP8-16 Tensor Sparse 493.9 TFLOPSTensor FP8-32 123.5 TFLOPSFP8-32 Tensor Sparse 246.9 TFLOPS | - |
Tensor FP16-16 123.5 TFLOPSFP16-16 Tensor Sparse 246.9 TFLOPSTensor FP16-32 61.73 TFLOPSFP16-32 Tensor Sparse 123.5 TFLOPS | Tensor FP16-16 140.1 TFLOPSFP16-16 Tensor Sparse 280.2 TFLOPSTensor FP16-32 70.04 TFLOPSFP16-32 Tensor Sparse 140.1 TFLOPS |
BF16 30.87 TFLOPSTensor BF16 61.73 TFLOPSBF16 Tensor Sparse 123.5 TFLOPS | BF16 35.02 TFLOPSTensor BF16 70.04 TFLOPSBF16 Tensor Sparse 140.1 TFLOPS |
Tensor TF32 30.87 TFLOPS | Tensor TF32 35.02 TFLOPS |
FP32 30.87 TFLOPS | FP32 35.02 TFLOPS |
FP64 482.3 GFLOPS | FP64 547.2 GFLOPS |
Tensor INT4 - | Tensor INT4 560.3 TOPS |
Tensor INT8 246.9 TOPS | Tensor INT8 280.2 TOPS |
Ray 93.62 TOPS | Ray 52.82 TOPS |
Pixel Rate 201 GPixel/s | Pixel Rate 191.5 GPixel/s |
Texture Rate 482.3 GTexel/s | Texture Rate 547.2 GTexel/s |
Shaders 6,144 Shaders | Shaders 10,240 Shaders |
TMUs 192 TMUs | TMUs 320 TMUs |
ROPs 80 ROPs | ROPs 112 ROPs |
Tensor Cores 192 T-Cores | Tensor Cores 320 T-Cores |
RT Cores 48 RT-Cores | RT Cores 80 RT-Cores |
SMs 48 SMs | SMs 80 SMs |
Base Clock 2.16GHz | Base Clock 1.4GHz |
Boost Clock 2.51GHz | Boost Clock 1.71GHz |
L2 Cache 41MB shared | L2 Cache 6.1MB shared |
12GB GDDR7 | 12GB GDDR6X |
Memory Bus 192-bit | Memory Bus 384-bit |
Memory Speed 28GT/s | Memory Speed 19.5GT/s |
Memory Bandwidth 672GB/s | Memory Bandwidth 936.2GB/s |
TDP 250W | TDP 350W |
Max Temp 85°C Max | Max Temp 93°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 9 | Encoder Model NVENC 7 |
Decoder Model NVDEC 6 | Decoder Model NVDEC 5 |
Form Factor PCIe Card | Form Factor PCIe Card |
PCIe 2.1-Slots | PCIe 3.2-Slots |
Height 112 mm (4.41")Width 242 mm (9.53")Depth 42 mm (1.65") | Height 131.8 mm (5.19")Width 317.8 mm (12.51")Depth 64.6 mm (2.54") |
Cooling Open-Air 2x Fans | Cooling Open-Air 3x Fans |
Power Connectors 1x 16-Pin 12VHPWR | Power Connectors 2x 8-Pin |
Manufacturer | Manufacturer |
Chip Designer | Chip Designer |
Architecture | Architecture |
Family | Family |
Branding  | Branding  |
Codename NV195 | Codename NV172 |
Chip Variant GB205-300-A1 | Chip Variant GA102-200-KD-A1 |
Market Segment Desktop | Market Segment Desktop |
Release Date Jan 6, 2025 | Release Date Jun 3, 2021 |
Foundry TSMC | Foundry Samsung |
Fabrication Node 4NP | Fabrication Node 8N |
Die Size 263mm² | Die Size 628mm² |
Transistor Count 31.1 Billion | Transistor Count 28.3 Billion |
Transistor Density 118.3 MTr/mm² | Transistor Density 45.04 MTr/mm² |
No images available
No images available