NVIDIA GeForce RTX 5070 vs NVIDIA 3x GeForce RTX 5090
NVIDIA GeForce RTX 5070
$549
NVIDIA 3x GeForce RTX 5090
$5,997
|
6144 Shaders
12GB GDDR7
2512MHz
|
3x 21760 Shaders
96GB (3x 32GB) GDDR7
2407MHz
|
|
Peak AI Performance
987.76 TFLOPS
FP4 Tensor Sparse
|
Peak AI Performance
10.06 PFLOPS
FP4 Tensor Sparse
|
|
FP32
30.87 TFLOPS
|
FP32
314.26 TFLOPS
|
|
FP16
30.87 TFLOPS
|
FP16
314.26 TFLOPS
|
|
Form Factor
PCIe Card
2.1-Slots
|
Form Factor
PCIe Card
3x 2.1-Slots
|
|
TDP
250W
|
TDP
1725W
|
|
Power Connectors
1x 16-Pin 12VHPWR
|
Power Connectors
1x 16-Pin 12VHPWR
|
Peak AI Performance
|
|
GeForce RTX 5070 - 987.76 TFLOPS FP4 Tensor Sparse
x1
3x GeForce RTX 5090 - 10.06 PFLOPS FP4 Tensor Sparse
x10.18
FP32
|
|
GeForce RTX 5070 - 30.87 TFLOPS FP32
x1
3x GeForce RTX 5090 - 314.26 TFLOPS FP32
x10.18
FP16
|
|
GeForce RTX 5070 - 30.87 TFLOPS FP16
x1
3x GeForce RTX 5090 - 314.26 TFLOPS FP16
x10.18
|
|
GeForce RTX 5070 - 672GB/s
x1
3x GeForce RTX 5090 - 1792GB/s
x2.67
|
|
GeForce RTX 5070 - 250W
x1
3x GeForce RTX 5090 - 1725W
x6.9
|
0%
|
0%
|
|
0%
|
0%
|
|
0%
|
0%
|
|
0%
|
0%
|
|
0%
|
0%
|
|
0%
|
0%
|
|
0%
|
0%
|
|
0%
|
0%
|
|
0%
|
0%
|
|
Manufacturer
NVIDIA
|
Manufacturer
NVIDIA
|
|
Chip Designer
NVIDIA
|
Chip Designer
NVIDIA
|
|
Architecture
Blackwell
|
Architecture
Blackwell
|
|
Family
GeForce 50
|
Family
GeForce 50
|
|
Codename
NV195
GB205
Variant
GB205-300-A1
|
Codename
NV192
GB202
Variant
GB202-300-A1
|
|
Market Segment
Desktop
|
Market Segment
Desktop
|
|
Release Date
1/6/2025
|
Release Date
1/6/2025
|
|
Foundry
TSMC
|
Foundry
TSMC
|
|
Fabrication Node
4NP
|
Fabrication Node
4NP
|
|
Die Size
263 mm²
|
Die Size
3x 762 mm²
|
|
Transistor Count
31.1 Billion
|
Transistor Count
3x 92.2 Billion
|
|
Transistor Density
118.25M/mm²
|
Transistor Density
121.07M/mm²
|
|
Form
PCIe Card
|
Form
PCIe Card
|
|
Shading Units
6144 Shaders
|
Shading Units
3x 21760 Shaders
|
|
Texture Mapping Units
192 TMUs
|
Texture Mapping Units
3x 680 TMUs
|
|
Render Output Units
80 ROPs
|
Render Output Units
3x 176 ROPs
|
|
Tensor Cores
192 T-Cores
|
Tensor Cores
3x 680 T-Cores
|
|
Ray-Tracing Cores
48 RT-Cores
|
Ray-Tracing Cores
3x 170 RT-Cores
|
|
Streaming Multiprocessors
48 SMs
|
Streaming Multiprocessors
3x 170 SMs
|
|
Graphics Processing Clusters
4 GPCs
|
Graphics Processing Clusters
3x 11 GPCs
|
|
2160MHz Base
2512MHz
|
2010MHz Base
2407MHz
|
|
Peak AI Performance
987.76 TFLOPS
FP4 Tensor Sparse
|
Peak AI Performance
10.06 PFLOPS
FP4 Tensor Sparse
|
|
FP4
493.88 TFLOPS Tensor
987.76 TFLOPS Tensor Sparse
|
FP4
5.03 PFLOPS Tensor
10.06 PFLOPS Tensor Sparse
|
|
FP8
246.94 TFLOPS Tensor (FP16 Accumulate)
493.88 TFLOPS Tensor (FP16 Accumulate) Sparse
123.47 TFLOPS Tensor (FP32 Accumulate)
246.94 TFLOPS Tensor (FP32 Accumulate) Sparse
|
FP8
2.51 PFLOPS Tensor (FP16 Accumulate)
5.03 PFLOPS Tensor (FP16 Accumulate) Sparse
1.26 PFLOPS Tensor (FP32 Accumulate)
2.51 PFLOPS Tensor (FP32 Accumulate) Sparse
|
|
FP16
30.87 TFLOPS
123.47 TFLOPS Tensor (FP16 Accumulate)
246.94 TFLOPS Tensor (FP16 Accumulate) Sparse
61.74 TFLOPS Tensor (FP32 Accumulate)
123.47 TFLOPS Tensor (FP32 Accumulate) Sparse
|
FP16
314.26 TFLOPS
1.26 PFLOPS Tensor (FP16 Accumulate)
2.51 PFLOPS Tensor (FP16 Accumulate) Sparse
628.52 TFLOPS Tensor (FP32 Accumulate)
1.26 PFLOPS Tensor (FP32 Accumulate) Sparse
|
|
FP32
30.87 TFLOPS
|
FP32
314.26 TFLOPS
|
|
FP64
480 GFLOPS
|
FP64
4.91 TFLOPS
|
|
BF16
30.87 TFLOPS
61.74 TFLOPS Tensor
123.47 TFLOPS Tensor Sparse
|
BF16
314.26 TFLOPS
628.52 TFLOPS Tensor
1.26 PFLOPS Tensor Sparse
|
|
TF32
30.87 TFLOPS Tensor
61.73 TFLOPS Tensor Sparse
|
TF32
314.26 TFLOPS Tensor
628.52 TFLOPS Tensor Sparse
|
|
INT8
246.94 TOPS Tensor
493.88 TOPS Tensor Sparse
|
INT8
2.51 POPS Tensor
5.03 POPS Tensor Sparse
|
|
INT32
30.87 TOPS
|
INT32
314.26 TOPS
|
|
Ray Tracing
93.6 TOPS
|
Ray Tracing
317.7 TOPS
|
|
Pixel Fillrate
200.96 GPixel/s
|
Pixel Fillrate
423.632 GPixel/s
|
|
Texture Fillrate
482.304 GTexel/s
|
Texture Fillrate
1636.76 GTexel/s
|
|
L1
64KB/SM Tex
128KB/SM
|
L1
64KB/SM Tex
128KB/SM
|
|
L2
40MB Shared
|
L2
96MB Shared
|
|
12GB
GDDR7
|
96GB (3x 32GB)
GDDR7
|
|
Bus Width
192Bit
|
Bus Width
512Bit
|
|
Clock
2210MHz
Transfer Rate
28GT/s
Bandwidth
672GB/s
|
Clock
2210MHz
Transfer Rate
28GT/s
Bandwidth
1792GB/s
|
|
TDP
250W
|
TDP
1725W
|
|
Temp
85°C Max
|
Temp
90°C Max
|
|
3x DisplayPort 2.1
1x HDMI 2.1
|
9x DisplayPort 2.1
3x HDMI 2.1
|
|
Max Resolution
7680x4320
|
Max Resolution
7680x4320
|
|
Max Resolution Refresh Rate
165Hz
|
Max Resolution Refresh Rate
165Hz
|
|
Variable Refresh Rate
G-Sync
FreeSync
|
Variable Refresh Rate
G-Sync
FreeSync
|
|
Display Stream Compression (DSC)
Supported
|
Display Stream Compression (DSC)
Supported
|
|
Multi Monitor Support
4
|
Multi Monitor Support
4
|
|
Content Protection
HDCP 2.3
|
Content Protection
HDCP 2.3
|
|
Model
NVENC 9
|
Model
3x NVENC 9
|
|
Codec
AVC (H.264)
HEVC (H.265)
AV1
|
Codec
AVC (H.264)
HEVC (H.265)
AV1
|
|
Model
NVDEC 6
|
Model
2x NVDEC 6
|
|
Codec
MPEG-1
MPEG-2
MPEG-4
VC-1
VP8
VP9
AVC (H.264)
HEVC (H.265)
AV1
|
Codec
MPEG-1
MPEG-2
MPEG-4
VC-1
VP8
VP9
AVC (H.264)
HEVC (H.265)
AV1
|
|
Direct X
12
Direct 3D
12_3
|
Direct X
12
Direct 3D
12_3
|
|
OpenGL
4.6
OpenCL
3.0
Vulkan
1.3
|
OpenGL
4.6
OpenCL
3.0
Vulkan
1.3
|
|
Shader Model
6.8
CUDA
12.8
PureVideo HD
VP13
VDPAU
Feature Set M
|
Shader Model
6.8
CUDA
12.8
PureVideo HD
VP13
VDPAU
Feature Set M
|
|
2x Fans
|
2x Fans
|
|
Power Connectors
1x 16-Pin 12VHPWR
|
Power Connectors
1x 16-Pin 12VHPWR
|
|
Slots Required
2.1
PCIe Version
5.0
PCIe Lanes
16
|
Slots Required
2.1
PCIe Version
5.0
PCIe Lanes
16
|
|
Height
112 mm (4.41 in)
Width
242 mm (9.53 in)
Depth
42 mm (1.65 in)
|
Height
137 mm (5.39 in)
Width
304 mm (11.97 in)
Depth
42 mm (1.65 in)
|
Copy Link