GPUSPECINDEX

NVIDIA · Ada Lovelace (derived from codename prefix) · 2024

GeForce RTX 4070 Super

GeForce RTX 4070 Super: 12 GB GDDR6X, 220 W, 7,168 CUDA cores. Launched January 17, 2024 at $599.

Memory
12 GB
Power
220 W
Bus width
192 bit
Memory12 GB19 of 77 in GeForce RTX
Power220 W25 of 66 by higher draw in GeForce RTX
CUDA cores7,16822 of 78 in GeForce RTX
Launch price$59922 of 54 in GeForce RTX

Affiliate links — we may earn a commission.

Memory

Capacity
12 GB
Memory type
GDDR6X
Bus width
192 bit
Memory clock
1,313 MHz
Bandwidth
504.2 GB/s

Compute

CUDA cores
7,168
SM count
56
TMUs
224
ROPs
80
FP32
28,385 GFLOPS
FP64
444 GFLOPS
Pixel fill rate
154.8 GP/s
Texture fill rate
443.52 GT/s

Clocks

Boost clock
1,980 MHz

Board

TDP
220 W
Bus interface
PCIe 4 x16
Form factor
desktop
Launch MSRP
599 USD

Physical

Architecture
Ada Lovelace (derived from codename prefix)
Codename
AD104-350-A1
Transistors
35.8 billion
Die size
294.5 mm²
Series
GeForce RTX
Launch date
January 17, 2024

Not recorded by the source: generation, process, base clock and external power required.

Silicon, made visible

Die area is drawn against the largest recorded die in the catalog. Shader dots use a labeled, card-specific ratio so both early and modern GPUs remain legible.

294.5 mm² die · 7,168 shadersDie area drawn to scale against the catalog maximum of 1,540 square millimeters. Each dot = 64 shaders.294.5 mm² die · 7,168 shadersEach dot = 64 shadersDashed frame = largest catalog die (1,540 mm²)

Within its series, the GeForce RTX 4070 Super ranks 16th of 66 for FP32 throughput and 19th of 77 for memory capacity. Its 220 W TDP is above the 75 W PCIe slot-power ceiling.

Source: Wikipedia, revision 1369488585 · catalog loaded 2026-08-16