The Nvidia GeForce RTX 3080 Ti Laptop GPU, also known as the 3080 Ti Mobile (designated for laptops as GN20-E8), stands as the premier graphics card for notebooks as of early 2022. Anchored in the GA103S Ampere chip, it presents an impressive configuration with 7,424 shaders and is paired with a substantial 16 GB GDDR6 graphics memory operating through a 256-bit memory bus. Supporting PCIe 4.0, this GPU is available in various configurations, with Total Graphics Power (TGP) spanning from 80 to 175 Watts as currently known. The Boost clock ranges from 1125 MHz (at 80W) to 1590 MHz (at 150W and above).
While the Max-Q variant is no longer utilized (previously designated for low-power variants), OEMs retain the option to implement Max-Q technologies such as Dynamic Boost and Whispermode.
Performance is intricately tied to the TGP setting and the cooling capabilities of the laptop. With its elevated core count, the RTX 3080 Ti outpaces its predecessor, the RTX 3080 Mobile, making it well-suited for demanding 4K gaming experiences.
The GA103S chip boasts 7,424 FP32 ALUs, with half of them capable of executing INT32 instructions (3,712 INT32 ALUs). Unlike Turing, where all shaders could execute FP32 or INT32 instructions, the GA103S introduces this division. Integrating 58 Raytracing cores and 232 Tensor cores for DLSS, along with a 4 MB L2 cache, it offers a comprehensive feature set. The Ampere architecture also incorporates a refined 5th generation video encoder (NVENC for H.264 and H.265) and a 7th generation decoder, now supporting various formats, including AV1. Compared to the GA103, the GA103S features a 256-bit memory bus.
Manufactured by Samsung using the 8nm (8N) process, the GA103S chip, while not quite matching the 7nm node at TSMC (employed by AMD and the professional GA100 Ampere chip), delivers commendable performance in the laptop GPU landscape.
| Graphics Processor | |
|---|---|
| Chip (GA103) |  |
| Architecture | Ampere |
| Foundry | Samsung |
| Process Size | 8 nm |
| Transistor Count | 22 Billion |
| Density | 44.4M / mm² |
| Die Size | 469 mm² |
| Memory | |
|---|---|
| Memory Size | 16 GB |
| Memory Type | GDDR6 |
| Memory Bus | 256 bit |
| Bandwidth | 512.0 GB/s |
| Graphics Config | |
|---|---|
| Shading Units | 7424 |
| Base Clock/Boost Clock | 810 MHz/1260 MHz |
| Memory Clock | 2000 MHz = 16 Gbps effective |
| TMUs/ROPs | 232/96 |
| SM Count | 58 |
| RT Cores/Tensor Cores | 58/232 |
| L1 Cache/L2 Cache | 128 KB (per SM)/4 MB |
| DirectX/OpenGL | 12 Ultimate (12_2)/4.6 |
| Theoretical Performance | |
|---|---|
| FP32 (float) | 18.71 TFLOPS |
| Pixel Rate | 121.0 GPixel/s |
| Texture Rate | 292.3 GTexel/s |
| 3DMark Time Spy Graphics | 12267 |
|---|---|
| 3DM11 Performance GPU | 42275 |
| Score | 54,542 |