GPU Comparison
GEFORCE
NVIDIA GeForce RTX 3090 Ti
CORE STATE
GA102
VRAM
24 GB
CLOCK SPEED
1860 MHz
TDP
450 W
BUS WIDTH
384 bit
ARCHITECTURE
Ampere
PROCESS
8 nm
LAUNCH DATE
2022
VS
GEFORCE
RTX 4000 SFF Ada Generation
CORE STATE
AD104
VRAM
20 GB
CLOCK SPEED
1560 MHz
TDP
70 W
BUS WIDTH
160 bit
ARCHITECTURE
Ada Lovelace
PROCESS
5 nm
LAUNCH DATE
2023
PERFORMANCE BENCHMARKS
3dmark_3dmark_steel_nomad_dx12
geekbench_opencl
geekbench_vulkan
DETAILED SPECIFICATIONS
SPECIFICATION
RTX 3090 Ti
RTX 4000 SFF Ada Generation
Core Specs
Shading Units
10,752
6,144
-42.9%
Shaders
10,752
6,144
-42.9%
TMUs
336
192
-42.9%
ROPs
112
64
-42.9%
SM Count
84
48
-42.9%
Clocks
Base Clock
1560 MHz
720 MHz
Boost Clock
1860 MHz
1560 MHz
Memory Clock
1313 MHz
21 Gbps effective
1750 MHz
14 Gbps effective
Memory
Memory Size
24 GB
20 GB
VRAM (MB)
24,576
20,480
-16.7%
Memory Type
GDDR6X
GDDR6
Memory Bus
384 bit
160 bit
Bandwidth
1.01 TB/s
280.0 GB/s
Cache
L1 Cache
128 KB (per SM)
128 KB (per SM)
L2 Cache
6 MB
48 MB
Performance
Pixel Rate
208.3 GPixel/s
99.84 GPixel/s
Texture Rate
625.0 GTexel/s
299.5 GTexel/s
FP32 (TFLOPS)
40.00 TFLOPS
19.17 TFLOPS
FP64 (TFLOPS)
625.0 GFLOPS (1:64)
299.5 GFLOPS (1:64)
FP16 (TFLOPS)
40.00 TFLOPS (1:1)
19.17 TFLOPS (1:1)
AI/RT
RT Cores
84
48
-42.9%
Tensor Cores
336
192
-42.9%
Power
TDP
450 W
70 W
TDP (W)
450
70
-84.4%
Suggested PSU
850 W
250 W
Power Connectors
1x 16-pin
None
Architecture
Architecture
Ampere
Ada Lovelace
GPU Name
GA102
AD104
Generation
GeForce 30
Workstation Ada
(x000A)
Process Size
8 nm
5 nm
Transistors
28,300 million
35,800 million
Die Size
628 mm²
294 mm²
Foundry
Samsung
TSMC
Density
45.1M / mm²
121.8M / mm²
API Support
DirectX
12 Ultimate (12_2)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.4
1.4
OpenCL
3.0
3.0
CUDA
8.6
8.9
Shader Model
6.8
6.8
Physical
Slot Width
Triple-slot
Dual-slot
Length
336 mm
13.2 inches
168 mm
6.6 inches
Height
140 mm
5.5 inches
69 mm
2.7 inches
Outputs
1x HDMI 2.13x DisplayPort 1.4a
4x mini-DisplayPort 1.4a
Bus Interface
PCIe 4.0 x16
PCIe 4.0 x16
Other
Launch Price
1,999 USD
—
Production
End-of-life
Active
Predecessor
GeForce 20
Workstation Ampere
Successor
GeForce 40
Blackwell PRO W