March 2022

Nvidia’s next generation is coming soon. The highly anticipated RTX 4000 series is expected to release in the fall of this year, likely late September. In preperation for the release, we have been scouring the internet searching for information about this new line up. Specifically looking for its exact release date, performance, specifications and price. Below are the expected specifications for the RTX 4090, RTX 4080 and RTX 4070!

Graphics Processor
GPU Name: AD102
Architecture: Lovelace
Foundry: TSMC
Process Size: 5 nm
Transistors: unknown
Die Size: 611 mm²

Graphics Card
Release Date: 2022
Availability: 2022
Generation: GeForce 40
Predecessor: GeForce 30
Production: Unreleased
Bus Interface: PCIe 4.0 x16

Clock Speeds
Base Clock: 1815 MHz
Boost Clock: 1925 MHz
Memory Clock: 1325 MHz / 21.2 Gbps effective

Memory
Memory Size: 24 GB
Memory Type: GDDR6X
Memory Bus: 384 bit
Bandwidth: 1,018 GB/s

Theoretical Performance
Pixel Rate: 369.6 GPixel/s
Texture Rate: 1,047 GTexel/s
FP16 (half) performance: 67.02 TFLOPS (1:1)
FP32 (float) performance: 67.02 TFLOPS
FP64 (double) performance: 1,047 GFLOPS (1:64)

Board Design
Slot Width: Triple-slot
Length: 336 mm / 13.2 inches
Width: 140 mm / 5.5 inches
Height: 61 mm / 2.4 inches
TDP: 450 W
Suggested PSU: 850 W
Outputs: 1x HDMI 2.1 / 3x DisplayPort 1.4a
Power Connectors: 1x 16-pin

Graphics Features
DirectX: 12 Ultimate (12_2)
OpenGL: 4.6
OpenCL: 3.0
Vulkan: 1.2
CUDA: 9.0
Shader Model: 6.6

Render Config
Shading Units: 17408
TMUs: 544
ROPs: 192
SM Count: 136
Tensor Cores: 544
RT Cores: 136
L1 Cache: 128 KB (per SM)
L2 Cache: 96 MB

Graphics Processor
GPU Name: AD102
Architecture: Lovelace
Foundry: TSMC
Process Size: 5 nm
Transistors: unknown
Die Size: 611 mm²

Graphics Card
Release Date: 2022
Availability: 2022
Generation: GeForce 40
Predecessor: GeForce 30
Production: Unreleased
Bus Interface: PCIe 4.0 x16

Clock Speeds
Base Clock: 1925 MHz
Boost Clock: 2075 MHz
Memory Clock: 2250 MHz / 18 Gbps effective

Memory
Memory Size: 16 GB
Memory Type: GDDR6
Memory Bus: 256 bit
Bandwidth: 576 GB/s

Theoretical Performance
Pixel Rate: 365.2 GPixel/s
Texture Rate: 913.0 GTexel/s
FP16 (half) performance: 58.43 TFLOPS (1:1)
FP32 (float) performance: 58.43 TFLOPS
FP64 (double) performance: 913.0 GFLOPS (1:64)

Board Design
Slot Width: Dual-slot
Length: 336 mm / 13.2 inches
Width: 140 mm / 5.5 inches
Height: 61 mm / 2.4 inches
TDP: 400 W
Suggested PSU: 800 W
Outputs: 1x HDMI 2.1 / 3x DisplayPort 1.4a
Power Connectors: 1x 16-pin

Graphics Features
DirectX: 12 Ultimate (12_2)
OpenGL: 4.6
OpenCL: 3.0
Vulkan: 1.2
CUDA: 9.0
Shader Model: 6.6

Render Config
Shading Units: 14080
TMUs: 440
ROPs: 176
SM Count: 110
Tensor Cores: 440
RT Cores: 110
L1 Cache: 128 KB (per SM)
L2 Cache: 64 MB

Graphics Processor
GPU Name: AD103
Architecture: Lovelace
Foundry: TSMC
Process Size: 5 nm
Transistors: unknown
Die Size: 380 mm²

Graphics Card
Release Date: 2022
Availability: 2022
Generation: GeForce 40
Predecessor: GeForce 30
Production: Unreleased
Bus Interface: PCIe 4.0 x16

Clock Speeds
Base Clock: 1925 MHz
Boost Clock: 2075 MHz
Memory Clock: 2250 MHz / 18 Gbps effective

Memory
Memory Size: 12 GB
Memory Type: GDDR6
Memory Bus: 192 bit
Bandwidth: 432 GB/s

Theoretical Performance
Pixel Rate: 232.4 GPixel/s
Texture Rate: 630.8 GTexel/s
FP16 (half) performance: 40.37 TFLOPS (1:1)
FP32 (float) performance: 40.37 TFLOPS
FP64 (double) performance: 630.8 GFLOPS (1:64)

Board Design
Slot Width: Dual-slot
Length: 336 mm / 13.2 inches
Width: 140 mm / 5.5 inches
Height: 61 mm / 2.4 inches
TDP: 300 W
Suggested PSU: 700 W
Outputs: 1x HDMI 2.1 / 3x DisplayPort 1.4a
Power Connectors: 1x 12-pin

Graphics Features
DirectX: 12 Ultimate (12_2)
OpenGL: 4.6
OpenCL: 3.0
Vulkan: 1.2
CUDA: 9.0
Shader Model: 6.6

Render Config
Shading Units: 9728
TMUs: 304
ROPs: 112
SM Count: 76
Tensor Cores: 304
RT Cores: 76
L1 Cache: 128 KB (per SM)
L2 Cache: 48 MB

We are also hoping to see the RTX 4060 and RTX 4050’s available before the end of 2022! And it is likely the launch will be huge, with NVIDIA reported to have spent roughly $10 billion on RTX 4000 components from TSMC already!

Graphics Processor
GPU Name: AD104
Architecture: Lovelace
Foundry: TSMC
Process Size: 5 nm
Transistors: unknown
Die Size: 300 mm²

Graphics Card
Release Date: 2022
Availability: 2022
Generation: GeForce 40
Predecessor: GeForce 30
Production: Unreleased
Bus Interface: PCIe 4.0 x16

Clock Speeds
Base Clock: 1925 MHz
Boost Clock: 2075 MHz
Memory Clock: 2250 MHz / 18 Gbps effective

Memory
Memory Size: 12 GB
Memory Type: GDDR6
Memory Bus: 192 bit
Bandwidth: 432 GB/s

Theoretical Performance
Pixel Rate: 132.8 GPixel/s
Texture Rate: 381.8 GTexel/s
FP16 (half) performance: 424.44 TFLOPS (1:1)
FP32 (float) performance: 24.44 TFLOPS
FP64 (double) performance: 381.8 GFLOPS (1:64)

Board Design
Slot Width: Dual-slot
TDP: 200 W
Suggested PSU: 550 W
Outputs: 1x HDMI 2.1 / 3x DisplayPort 1.4a
Power Connectors: 1x 12-pin

Graphics Features
DirectX: 12 Ultimate (12_2)
OpenGL: 4.6
OpenCL: 3.0
Vulkan: 1.2
CUDA: 9.0
Shader Model: 6.6

Render Config
Shading Units: 5888
TMUs: 184
ROPs: 64
SM Count: 46
Tensor Cores: 184
RT Cores: 46
L1 Cache: 128 KB (per SM)
L2 Cache: 48 MB

Graphics Processor
GPU Name: AD106
Architecture: Lovelace
Foundry: TSMC
Process Size: 5 nm
Transistors: unknown
Die Size: 203 mm²

Graphics Card
Release Date: 2022
Availability: 2022
Generation: GeForce 40
Predecessor: GeForce 30
Production: Unreleased
Bus Interface: PCIe 4.0 x16

Clock Speeds
Base Clock: 1925 MHz
Boost Clock: 2075 MHz
Memory Clock: 2250 MHz / 18 Gbps effective

Memory
Memory Size: 8 GB
Memory Type: GDDR6
Memory Bus: 128 bit
Bandwidth: 288 GB/s

Theoretical Performance
Pixel Rate: 99.6 GPixel/s
Texture Rate: 249 GTexel/s
FP16 (half) performance: 15.94 TFLOPS (1:1)
FP32 (float) performance: 15.94 TFLOPS
FP64 (double) performance: 249 GFLOPS (1:64)

Board Design
Slot Width: Dual-slot
TDP: 150 W
Suggested PSU: 450 W
Outputs: 1x HDMI 2.1 / 3x DisplayPort 1.4a
Power Connectors: 1x 12-pin

Graphics Features
DirectX: 12 Ultimate (12_2)
OpenGL: 4.6
OpenCL: 3.0
Vulkan: 1.2
CUDA: 9.0
Shader Model: 6.6

Render Config
Shading Units: 3840
TMUs: 120
ROPs: 48
SM Count: 30
Tensor Cores: 120
RT Cores: 30
L1 Cache: 128 KB (per SM)
L2 Cache: 32 MB

We will keep you updated on all the news we see about the upcoming NVIDIA RTX 4000 series right here, so please check back with us soon for more information. Have a great week!