NVIDIA RTX 4000 Ada Generation vs NVIDIA GeForce RTX 4070 Ti SUPER
We compared two Desktop platform GPUs: 20GB VRAM RTX 4000 Ada Generation and 16GB VRAM GeForce RTX 4070 Ti SUPER to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA RTX 4000 Ada Generation 's Advantages
More VRAM (20GB vs 16GB)
Lower TDP (130W vs 285W)
NVIDIA GeForce RTX 4070 Ti SUPER 's Advantages
Boost Clock has increased by 20% (2610MHz vs 2175MHz)
Larger VRAM bandwidth (672.3GB/s vs 360.0GB/s)
2304 additional rendering cores
Score
Benchmark
FP32 (float)
RTX 4000 Ada Generation
26.73 TFLOPS
GeForce RTX 4070 Ti SUPER
+64%
44.1 TFLOPS
3DMark Time Spy
RTX 4000 Ada Generation
14340
GeForce RTX 4070 Ti SUPER
+69%
24314
3DMark Time Spy Extreme
RTX 4000 Ada Generation
6708
GeForce RTX 4070 Ti SUPER
+77%
11874
3DMark Speed Way
RTX 4000 Ada Generation
3666
GeForce RTX 4070 Ti SUPER
+74%
6396
Blender
RTX 4000 Ada Generation
4737
GeForce RTX 4070 Ti SUPER
+67%
7921
OctaneBench
RTX 4000 Ada Generation
492
GeForce RTX 4070 Ti SUPER
+73%
856
Graphics Card
Aug 2023
Release Date
Jan 2024
Workstation Ada
Generation
GeForce 40
Desktop
Type
Desktop
PCIe 4.0 x16
Bus Interface
PCIe 4.0 x16
Clock Speeds
1500 MHz
Base Clock
2340 MHz
2175 MHz
Boost Clock
2610 MHz
2250 MHz
Memory Clock
1313 MHz
Memory
20GB
Memory Size
16GB
GDDR6
Memory Type
GDDR6X
160bit
Memory Bus
256bit
360.0GB/s
Bandwidth
672.3GB/s
Render Config
-
-
-
48
SM Count
66
6144
Shading Units
8448
192
TMUs
264
64
ROPs
96
192
Tensor Cores
264
48
RT Cores
66
128 KB (per SM)
L1 Cache
128 KB (per SM)
48 MB
L2 Cache
48 MB
-
-
-
Theoretical Performance
139.2 GPixel/s
Pixel Rate
250.6 GPixel/s
417.6 GTexel/s
Texture Rate
689.0 GTexel/s
26.73 TFLOPS
FP16 (half)
44.10 TFLOPS
26.73 TFLOPS
FP32 (float)
44.10 TFLOPS
417.6 GFLOPS
FP64 (double)
689.0 GFLOPS
Board Design
130W
TDP
285W
300 W
Suggested PSU
600 W
4x DisplayPort 1.4a
Outputs
1x HDMI 2.1
3x DisplayPort 1.4a
1x 16-pin
Power Connectors
1x 16-pin
Graphics Processor
AD104
GPU Name
AD103
-
GPU Variant
AD103-275-A1
Ada Lovelace
Architecture
Ada Lovelace
TSMC
Foundry
TSMC
5 nm
Process Size
5 nm
35.8 billion
Transistors
45.9 billion
294 mm²
Die Size
379 mm²
Graphics Features
12 Ultimate (12_2)
DirectX
12 Ultimate (12_2)
4.6
OpenGL
4.6
3.0
OpenCL
3.0
1.3
Vulkan
1.3
8.9
CUDA
8.9
6.8
Shader Model
6.7
