NVIDIA GeForce RTX 4090 Mobile vs NVIDIA GeForce RTX 5070 SUPER Comparison

FAQ

Q: What is the performance difference between the RTX 4090 Mobile and the RTX 5070 SUPER?

A: The database shows the RTX 4090 Mobile has an average benchmark score of 43,667, placing it in the 84th percentile of all GPUs. The RTX 5070 SUPER has an average score of 2,690, which places it in the 18th percentile. The gap is substantial, with the 4090 Mobile scoring roughly 16 times higher in the recorded aggregate.

Q: What are the nearest rivals for each card according to the database?

A: For the RTX 4090 Mobile, the nearest rivals are the NVIDIA RTX A6000 (0.9% faster), the NVIDIA Quadro M6000 (0.8% slower), the GeForce RTX 5050 Mobile (0.9% slower), and the Quadro M6000 24 GB (0.9% slower). For the RTX 5070 SUPER, the nearest rivals are the Quadro K1100M (1% slower), GeForce GT 1030 (1.1% slower), Intel Arc Pro B50 (1.1% slower), and GeForce GT 440 (1.7% slower).

Q: Which card has more shading units?

A: The RTX 4090 Mobile has 9,728 shading units, while the RTX 5070 SUPER has 6,400. The 4090 Mobile also has more texture mapping units (304 vs 200), more render output units (112 vs 80), more RT cores (76 vs 50), and more tensor cores (304 vs 200).

Q: How do the memory configurations differ?

A: The RTX 4090 Mobile uses 16 GB of GDDR6 on a 256-bit bus with 576.0 GB/s bandwidth. The RTX 5070 SUPER uses 18 GB of GDDR7 on a 192-bit bus with 672.0 GB/s bandwidth. The 5070 SUPER provides more capacity and higher bandwidth despite the narrower bus.

Q: What are the power requirements for each card?

A: The RTX 4090 Mobile has a TDP of 120 W and uses no power connectors, as it is an integrated graphics processor (IGP). The RTX 5070 SUPER has a TDP of 275 W and uses a single 16-pin power connector, with a dual-slot cooler.

Q: Which card has higher clock speeds?

A: The RTX 5070 SUPER has a base clock of 2325 MHz and a boost clock of 2512 MHz. The RTX 4090 Mobile has a base clock of 1335 MHz and a boost clock of 1695 MHz. The 5070 SUPER runs at significantly higher frequencies.

Architecture Differences

The RTX 4090 Mobile uses the AD103 chip based on the Ada Lovelace architecture, while the RTX 5070 SUPER uses the GB205 chip based on the Blackwell 2.0 architecture. Both are fabricated by TSMC on a 5 nm process node, but the transistor counts differ considerably. The AD103 contains 45,900 million transistors on a 379 mm² die, giving a transistor density of 121.1M per mm². The GB205 contains 31,100 million transistors on a 263 mm² die, with a density of 118.3M per mm².

The core configurations show a clear divergence in design philosophy. The 4090 Mobile deploys 9,728 shading units, 304 TMUs, 112 ROPs, 76 RT cores, and 304 tensor cores. The 5070 SUPER uses 6,400 shading units, 200 TMUs, 80 ROPs, 50 RT cores, and 200 tensor cores. This means the 4090 Mobile has roughly 52% more shading units and the same proportional advantage across all core types.

Clock speeds tell a different story. The 5070 SUPER operates at a base clock of 2325 MHz and boosts to 2512 MHz, compared to the 4090 Mobile's 1335 MHz base and 1695 MHz boost. The 5070 SUPER's higher frequencies partially compensate for its lower core count, but not enough to close the overall gap.

The memory subsystems differ in both type and configuration. The 4090 Mobile uses 16 GB of GDDR6 on a 256-bit interface with 576.0 GB/s bandwidth. The 5070 SUPER uses 18 GB of GDDR7 on a 192-bit interface with 672.0 GB/s bandwidth. The newer GDDR7 memory runs at 1750 MHz with 28 Gbps effective speed, while the GDDR6 runs at 2250 MHz with 18 Gbps effective speed.

Interface capabilities also differ. The 4090 Mobile uses PCIe 4.0 x16, while the 5070 SUPER uses PCIe 5.0 x16. The 5070 SUPER provides display outputs including 1x HDMI 2.1b and 3x DisplayPort 2.1b, whereas the 4090 Mobile's display outputs are portable device dependent. The 4090 Mobile is an IGP with no power connectors and no slot width, while the 5070 SUPER is a dual-slot card measuring 245 mm by 115 mm by 40 mm and requires a single 16-pin connector.

Both cards support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The production status for both is Active. The 4090 Mobile was released on 2023-01-02, while the 5070 SUPER has a release date of 2025-12-31.

Head-to-Head Benchmarks

The database contains multiple benchmark entries for the RTX 4090 Mobile across several test suites, while the RTX 5070 SUPER has only a single recorded benchmark. This asymmetry requires careful interpretation of the aggregate scores.

For the RTX 4090 Mobile, the recorded benchmark results show strong performance across different APIs. In Geekbench OpenCL, it scores 180,831. In Geekbench Vulkan, it scores 170,774. Passmark results include 173 for DirectX 10, 262 for DirectX 11, 107 for DirectX 12, 310 for DirectX 9, 984 for G2D, 27,212 for G3D, and 12,347 for GPU compute. The average benchmark score is 43,667.

The RTX 5070 SUPER has one recorded benchmark: 3DMark Steel Nomad DX12 with a score of 2,690. Its average benchmark score is identical at 2,690. The percentile ranking for the 5070 SUPER is 18, while the 4090 Mobile sits at the 84th percentile.

Comparing the average scores directly, the 4090 Mobile delivers 43,667 versus 2,690 for the 5070 SUPER. The nearest rivals for the 4090 Mobile include the RTX A6000 at 44,075 (0.9% faster) and the Quadro M6000 at 43,301 (0.8% slower). The 5070 SUPER's nearest rivals are considerably lower performers: the Quadro K1100M at 2,664 (1% slower) and the GeForce GT 1030 at 2,662 (1.1% slower).

The pixel rate figures show the 5070 SUPER at 201.0 GPixel/s versus 189.8 GPixel/s for the 4090 Mobile, indicating a 5.9% advantage for the newer card in this specific metric. The texture rates are nearly identical: 515.3 GTexel/s for the 4090 Mobile and 502.4 GTexel/s for the 5070 SUPER, a 2.6% difference in favor of the 4090 Mobile.

FP32 compute performance is close as well. The 4090 Mobile delivers 32.98 TFLOPS, while the 5070 SUPER delivers 32.15 TFLOPS. Both cards achieve a 1:1 ratio for FP16 performance, meaning the half-precision throughput matches the single-precision figures.

The wins are split by metric rather than by overall benchmark suite. The 4090 Mobile wins on shading throughput, texture rate, and FP32 compute. The 5070 SUPER wins on pixel rate, memory bandwidth, and clock speed. In aggregate benchmark scoring, the 4090 Mobile is the clear leader.

Specification Differences

| Specification | RTX 4090 Mobile | RTX 5070 SUPER |

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

| Architecture | Ada Lovelace | Blackwell 2.0 |

| Chip | AD103 | GB205 |

| Transistors | 45,900 million | 31,100 million |

| Die Size | 379 mm² | 263 mm² |

| Transistor Density | 121.1M / mm² | 118.3M / mm² |

| Base Clock | 1335 MHz | 2325 MHz |

| Boost Clock | 1695 MHz | 2512 MHz |

| Memory Clock | 2250 MHz, 18 Gbps effective | 1750 MHz, 28 Gbps effective |

| Memory Size | 16 GB | 18 GB |

| Memory Type | GDDR6 | GDDR7 |

| Bus Width | 256 bit | 192 bit |

| Bandwidth | 576.0 GB/s | 672.0 GB/s |

| Shading Units | 9728 | 6400 |

| TMUs | 304 | 200 |

| ROPs | 112 | 80 |

| RT Cores | 76 | 50 |

| Tensor Cores | 304 | 200 |

| Pixel Rate | 189.8 GPixel/s | 201.0 GPixel/s |

| Texture Rate | 515.3 GTexel/s | 502.4 GTexel/s |

| FP32 | 32.98 TFLOPS | 32.15 TFLOPS |

| FP16 | 32.98 TFLOPS (1:1) | 32.15 TFLOPS (1:1) |

| TDP | 120 W | 275 W |

| Slot Width | IGP | Dual-slot |

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

| Bus Interface | PCIe 4.0 x16 | PCIe 5.0 x16 |

| Display Outputs | Portable Device Dependent | 1x HDMI 2.1b, 3x DisplayPort 2.1b |

| Dimensions | N/A | 245 mm x 115 mm x 40 mm |

| Release Date | 2023-01-02 | 2025-12-31 |

Where Each One Wins

The RTX 4090 Mobile demonstrates dominance in compute-heavy workloads that leverage its larger core configuration. With 9,728 shading units, 304 TMUs, and 112 ROPs, it provides substantially higher texture throughput at 515.3 GTexel/s and marginally higher FP32 compute at 32.98 TFLOPS versus 32.15 TFLOPS for the 5070 SUPER. The 4090 Mobile's 76 RT cores and 304 tensor cores suggest stronger ray tracing and AI acceleration capabilities based on raw hardware counts, though direct comparative benchmarks for these specific workloads are not recorded in the database.

The 4090 Mobile's average benchmark score of 43,667 places it at the 84th percentile of all GPUs, with nearest rivals like the RTX A6000 and Quadro M6000. This positions it as a high-end performer in the database's ranking system. The card's 120 W TDP and IGP form factor indicate it is designed for mobile integration where power efficiency is prioritized.

The RTX 5070 SUPER wins on memory bandwidth with 672.0 GB/s against 576.0 GB/s for the 4090 Mobile, a 16.7% advantage. This comes from GDDR7 memory running at 28 Gbps effective speed on a 192-bit bus. The higher pixel rate of 201.0 GPixel/s versus 189.8 GPixel/s gives it a 5.9% edge in fill-rate-limited scenarios. Its boost clock of 2512 MHz versus 1695 MHz indicates responsiveness in lightly threaded workloads that depend on frequency rather than core count.

The 5070 SUPER's benchmark presence is limited to a single 3DMark Steel Nomad DX12 test with a score of 2,690, placing it at the 18th percentile. Its nearest rivals include the Quadro K1100M and GeForce GT 1030, which are entry-level components. This suggests the recorded data for the 5070 SUPER is not yet representative of its full performance envelope, as the card has only one data point in the database.

The 5070 SUPER offers 18 GB of memory versus 16 GB for the 4090 Mobile, providing additional capacity for large datasets or high-resolution textures. Its PCIe 5.0 x16 interface doubles the bandwidth available to the host system compared to the 4090 Mobile's PCIe 4.0 x16. The 5070 SUPER also provides fixed display outputs with HDMI 2.1b and DisplayPort 2.1b, whereas the 4090 Mobile's outputs depend on the portable device implementation.

For scenarios where the 4090 Mobile's 84th percentile ranking and extensive benchmark coverage matter, it is the stronger choice in aggregate performance. The 5070 SUPER's advantages lie in memory bandwidth, pixel throughput, and clock speed, which may translate to specific workload benefits not captured by the limited benchmark data available.