NVIDIA GeForce RTX 4070 AD103 vs NVIDIA GeForce RTX 4070 Mobile Comparison

Architecture Differences

The two GPUs share the Ada Lovelace architecture and TSMC 5 nm process, but they diverge significantly in silicon and configuration. The RTX 4070 AD103 uses the AD103 chip with 45,900 million transistors on a 379 mm² die, while the RTX 4070 Mobile uses the AD106 chip with 22,900 million transistors on a 188 mm² die. Transistor density is nearly identical at 121.1M per mm² for the desktop part versus 121.8M per mm² for the mobile part, confirming the same manufacturing process.

The desktop GPU houses 5,888 shading units, 184 texture mapping units, 64 raster output units, 46 ray tracing cores, and 184 tensor cores. The mobile GPU reduces these to 4,608 shading units, 144 TMUs, 48 ROPs, 36 RT cores, and 144 tensor cores. That represents a 21.7% reduction in shaders, a 21.7% reduction in TMUs, a 25% reduction in ROPs, and a 21.7% reduction in both RT and tensor cores. Clock speeds also differ markedly: the desktop runs at 1920 MHz base and 2475 MHz boost, while the mobile part runs at 1395 MHz base and 1695 MHz boost. The desktop boost clock is 46% higher.

Memory configurations are entirely different. The desktop card uses 12 GB of GDDR6X on a 192-bit bus with 504.2 GB/s bandwidth. The mobile chip uses 8 GB of GDDR6 on a 128-bit bus with 256.0 GB/s bandwidth. Effective memory speed is 21 Gbps for the desktop versus 16 Gbps for the mobile. The desktop has double the memory bus width and nearly double the bandwidth. Pixel rate is 158.4 GPixel/s for the desktop versus 81.36 GPixel/s for the mobile, and texture rate is 455.4 GTexel/s versus 244.1 GTexel/s. FP32 throughput is 29.15 TFLOPS for the desktop versus 15.62 TFLOPS for the mobile, a difference of 86.6%.

Power delivery also separates them. The desktop card carries a 200 W TDP, requires a 550 W suggested power supply, uses a single 16-pin connector, and occupies a dual-slot form factor. The mobile GPU has a 115 W TDP, uses no power connectors, and is an integrated chip package (IGP). The desktop uses a PCIe 4.0 x16 interface, while the mobile uses PCIe 4.0 x8. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Where Each One Wins

Benchmark data from the database shows the mobile RTX 4070 holds wins across multiple test categories. In Geekbench OpenCL, the mobile scores 109,197, and in Geekbench Vulkan it scores 108,367. These are compute-focused tests that show strong performance for the mobile part despite its lower power envelope. PassMark tests also favor the mobile GPU in legacy and modern APIs: DirectX 9 scores 223, DirectX 10 scores 116, DirectX 11 scores 179, DirectX 12 scores 85, G2D scores 763, G3D scores 19,587, and GPU compute scores 8,399. The aggregate average benchmark score for the mobile part is 27,435.

The desktop RTX 4070 AD103 has no recorded benchmark scores in the database, so direct numeric comparisons cannot be made from the recorded measurements. However, its architectural advantages point to clear wins in raw rasterization and ray tracing workloads. The FP32 throughput of 29.15 TFLOPS versus 15.62 TFLOPS suggests the desktop part delivers nearly double the compute throughput. The 504.2 GB/s bandwidth versus 256.0 GB/s gives the desktop a major advantage in memory-heavy scenes, high-resolution textures, and large geometry loads. The pixel rate of 158.4 GPixel/s versus 81.36 GPixel/s indicates the desktop fills frames faster at high resolutions.

The mobile part wins on portability and integration. It fits in laptops with no external power connectors and a 115 W TDP, making it suitable for thin chassis. The desktop requires a dual-slot cooler, a 550 W power supply, and a PCIe 4.0 x16 slot. The mobile GPU also holds a higher percentile ranking among all GPUs: 73rd percentile versus the desktop's 50th percentile, though this reflects the mobile part's benchmark presence in the database rather than a direct performance comparison.

FAQ

Q: Which GPU has more shading units?

A: The desktop RTX 4070 AD103 has 5,888 shading units, while the mobile RTX 4070 has 4,608 shading units.

Q: What is the memory bandwidth difference?

A: The desktop card offers 504.2 GB/s using GDDR6X on a 192-bit bus, while the mobile card offers 256.0 GB/s using GDDR6 on a 128-bit bus.

Q: Does the mobile GPU require a power supply recommendation?

A: No, the database lists no suggested PSU for the mobile part, and it uses no power connectors. The desktop card lists a 550 W suggested PSU.

Q: What is the TDP of each GPU?

A: The desktop RTX 4070 AD103 has a 200 W TDP, and the mobile RTX 4070 has a 115 W TDP.

Q: Which GPU has a higher boost clock?

A: The desktop part boosts to 2475 MHz, while the mobile part boosts to 1695 MHz.

Q: What benchmark results exist for the mobile GPU?

A: The mobile GPU has nine recorded benchmarks, including Geekbench OpenCL at 109,197 and PassMark G3D at 19,587, with an average score of 27,435. The desktop GPU has no recorded benchmarks in the database.

Head-to-Head Benchmarks

Direct head-to-head benchmark data is not available in the database for these two GPUs. However, the mobile RTX 4070 has extensive recorded measurements, and the desktop part's specifications allow for architectural comparison.

The mobile GPU's Geekbench OpenCL score of 109,197 and Vulkan score of 108,367 represent its strongest compute results. In PassMark tests, the highest score is DirectX 9 at 223, followed by DirectX 11 at 179, DirectX 10 at 116, and DirectX 12 at 85. The G3D score of 19,587 is the most meaningful gaming-oriented metric, and the GPU compute score of 8,399 measures general compute workloads. The G2D score of 763 covers 2D operations.

The desktop GPU's specifications indicate it would outperform the mobile part in most measured categories. The FP32 throughput of 29.15 TFLOPS versus 15.62 TFLOPS is a 86.6% advantage. The texture rate of 455.4 GTexel/s versus 244.1 GTexel/s is an 86.6% advantage. The pixel rate of 158.4 GPixel/s versus 81.36 GPixel/s is a 94.7% advantage. Bandwidth of 504.2 GB/s versus 256.0 GB/s is a 97% advantage.

The mobile GPU's nearest rivals in the database provide context. The AMD Radeon RX 6700 XT scores 27,425, which is 0% different from the mobile RTX 4070's 27,435 average. The NVIDIA GeForce RTX 3090 scores 27,565, which is 0.5% higher. The NVIDIA RTX PRO 4000 Blackwell scores 27,135, which is 1.1% lower. The AMD Radeon Pro Vega 20 scores 27,839, which is 1.5% higher. These small deltas show the mobile RTX 4070 sits in a tight performance band around these desktop-class cards, despite its lower power draw.

The Verdict

The database shows two GPUs with the same architecture name but very different execution profiles. The desktop RTX 4070 AD103 is the stronger part on paper, with more cores, higher clocks, more memory, and much higher bandwidth. Its FP32 throughput of 29.15 TFLOPS nearly doubles the mobile part's 15.62 TFLOPS. Its 504.2 GB/s bandwidth versus 256.0 GB/s gives it a clear edge for high-resolution textures and memory-intensive workloads. The desktop card also supports a 192-bit bus versus the mobile's 128-bit bus, and it has 12 GB of GDDR6X versus 8 GB of GDDR6.

The mobile RTX 4070 is the only part with recorded benchmark scores, and its average of 27,435 places it near the AMD Radeon RX 6700 XT (27,425), the NVIDIA GeForce RTX 3090 (27,565), and the NVIDIA RTX PRO 4000 Blackwell (27,135). These are desktop-class results from a 115 W mobile chip, which indicates strong efficiency. The mobile part also holds a 73rd percentile ranking versus the desktop's 50th percentile, though this reflects the mobile part's presence in the database and the desktop's lack of recorded benchmarks.

For a desktop tower with a PCIe 4.0 x16 slot, a 550 W power supply, and space for a dual-slot card, the RTX 4070 AD103 is the clear choice based on specifications. For a laptop or compact system with no power connectors and a 115 W budget, the mobile RTX 4070 delivers competitive performance within its power class. The production status confirms this split: the desktop part is end-of-life, while the mobile part is active. The desktop part launched on 2024-02-29 with a launch MSRP of 599 USD, while the mobile part launched on 2023-01-02 with no recorded MSRP.

Specification Differences

| Field | RTX 4070 AD103 | RTX 4070 Mobile |

|---|---|---|

| Chip | AD103 | AD106 |

| Transistors | 45,900 million | 22,900 million |

| Die size | 379 mm² | 188 mm² |

| Base clock | 1920 MHz | 1395 MHz |

| Boost clock | 2475 MHz | 1695 MHz |

| Memory clock | 1313 MHz, 21 Gbps effective | 2000 MHz, 16 Gbps effective |

| Memory size | 12 GB | 8 GB |

| Memory type | GDDR6X | GDDR6 |

| Bus width | 192 bit | 128 bit |

| Bandwidth | 504.2 GB/s | 256.0 GB/s |

| Shading units | 5888 | 4608 |

| TMUs | 184 | 144 |

| ROPs | 64 | 48 |

| RT cores | 46 | 36 |

| Tensor cores | 184 | 144 |

| Pixel rate | 158.4 GPixel/s | 81.36 GPixel/s |

| Texture rate | 455.4 GTexel/s | 244.1 GTexel/s |

| FP32 | 29.15 TFLOPS | 15.62 TFLOPS |

| FP16 | 29.15 TFLOPS (1:1) | 15.62 TFLOPS (1:1) |

| TDP | 200 W | 115 W |

| Slot width | Dual-slot | IGP |

| Power connectors | 1x 16-pin | None |

| Suggested PSU | 550 W | None |

| Bus interface | PCIe 4.0 x16 | PCIe 4.0 x8 |

| Display outputs | 1x HDMI 2.1, 3x DisplayPort 1.4a | Portable Device Dependent |

| Production status | End-of-life | Active |

| Release date | 2024-02-29 | 2023-01-02 |

| Launch MSRP | 599 USD | None |

| Dimensions | 240 mm x 110 mm x 40 mm | None recorded |