NVIDIA GeForce RTX 2080 SUPER vs NVIDIA GeForce RTX 3070 Mobile Comparison

FAQ

Q: Which GPU is faster in the 3DMark Steel Nomad DX12 test?

A: The NVIDIA GeForce RTX 3070 Mobile wins this one. It scores 2380 versus 1882 for the RTX 2080 SUPER, a 20.9% advantage for the mobile Ampere part.

Q: Does the older RTX 2080 SUPER win any benchmarks?

A: Yes, decisively. Out of 10 head-to-head tests, the RTX 2080 SUPER wins 9. Its largest margin is in Passmark G2D, where it scores 920 versus 641, a 43.5% lead.

Q: How do the two compare in raw compute workloads?

A: The RTX 2080 SUPER leads in Passmark GPU Compute with 8290 versus 6827, a 21.4% advantage. It also wins Geekbench OpenCL with 99226 versus 92939 (6.8% ahead).

Q: Are both GPUs based on the same architecture?

A: No. The RTX 2080 SUPER uses the Turing architecture on a 12 nm TSMC process. The RTX 3070 Mobile uses Ampere on Samsung's 8 nm node. Both support DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4.

Q: What is the memory configuration of each card?

A: Both have 8 GB of GDDR6 on a 256-bit bus. The RTX 2080 SUPER has higher bandwidth at 495.9 GB/s, while the RTX 3070 Mobile delivers 448.0 GB/s.

Q: Which GPU has the higher average benchmark score?

A: The RTX 2080 SUPER has an average benchmark score of 24170, placing it in the 69th percentile of all GPUs. The RTX 3070 Mobile averages 20534, putting it in the 65th percentile.

Where Each One Wins

The RTX 2080 SUPER is the clear winner in legacy DirectX and general-purpose workloads. It beats the RTX 3070 Mobile in Passmark DirectX 9 by 41.9% (227 vs 160), in DirectX 10 by 23.9% (140 vs 113), in DirectX 11 by 19.6% (165 vs 138), and in DirectX 12 by 17.2% (75 vs 64). These are not small margins; they indicate a consistent advantage across API generations. The desktop card also dominates in Vulkan, scoring 111284 versus 86768, a 28.3% lead. If the workload is compute-heavy or relies on older APIs, the RTX 2080 SUPER is the stronger choice.

The RTX 3070 Mobile has one clear territory: modern DX12 gaming at high load, as shown by the 3DMark Steel Nomad result. Its 2380 score versus 1882 is a 20.9% win. This is the only head-to-head test the mobile GPU takes, but it is a meaningful one for current titles that stress DX12 rendering. The mobile part also has a higher pixel rate (124.8 GPixel/s versus 116.2 GPixel/s), which helps fill-rate-bound scenes.

For pure rasterization performance in Passmark G3D, the RTX 2080 SUPER leads with 19490 versus 15309, a 27.3% advantage. The desktop card also wins in 2D performance by a massive 43.5% (920 vs 641), which matters for multi-monitor desktop work and light 2D acceleration tasks.

Architecture Differences

The RTX 2080 SUPER is built on the Turing architecture using the TU104 chip, manufactured by TSMC on a 12 nm process. It packs 13,600 million transistors on a 545 mm² die, giving a transistor density of 25.0M per mm². The RTX 3070 Mobile uses the Ampere architecture with the GA104 chip, fabricated by Samsung on an 8 nm process. It has 17,400 million transistors on a much smaller 392 mm² die, yielding a density of 44.4M per mm².

The Ampere chip is not just denser, it is also organized differently. The RTX 3070 Mobile has 5120 shading units versus 3072 on the RTX 2080 SUPER, a 67% increase in raw shader count. However, the desktop Turing card has more texture units (192 versus 160) and more ray tracing cores (48 versus 40). The RTX 3070 Mobile has more ROPs (80 versus 64) but fewer tensor cores (160 versus 384).

FP32 throughput tells a similar story: the RTX 3070 Mobile reaches 15.97 TFLOPS, while the RTX 2080 SUPER delivers 11.15 TFLOPS. Interestingly, the FP16 figures diverge: the Turing card hits 22.30 TFLOPS via its 2:1 ratio, while the Ampere mobile part stays at 15.97 TFLOPS with a 1:1 ratio. The RTX 2080 SUPER's FP16 advantage makes it better suited for workloads that leverage half-precision compute.

The memory subsystems differ in clock and bandwidth. The RTX 2080 SUPER runs memory at 1937 MHz with 15.5 Gbps effective speed, producing 495.9 GB/s. The RTX 3070 Mobile runs at 1750 MHz with 14 Gbps effective, yielding 448.0 GB/s. Both use 8 GB GDDR6 on a 256-bit bus.

Specification Differences

| Specification | RTX 2080 SUPER | RTX 3070 Mobile |

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

| Architecture | Turing | Ampere |

| Process Node | 12 nm (TSMC) | 8 nm (Samsung) |

| Transistors | 13,600 million | 17,400 million |

| Die Size | 545 mm² | 392 mm² |

| Transistor Density | 25.0M / mm² | 44.4M / mm² |

| Base Clock | 1650 MHz | 1110 MHz |

| Boost Clock | 1815 MHz | 1560 MHz |

| Memory Clock | 1937 MHz / 15.5 Gbps effective | 1750 MHz / 14 Gbps effective |

| Memory Bandwidth | 495.9 GB/s | 448.0 GB/s |

| Shading Units | 3072 | 5120 |

| TMUs | 192 | 160 |

| ROPs | 64 | 80 |

| RT Cores | 48 | 40 |

| Tensor Cores | 384 | 160 |

| Pixel Rate | 116.2 GPixel/s | 124.8 GPixel/s |

| Texture Rate | 348.5 GTexel/s | 249.6 GTexel/s |

| FP32 | 11.15 TFLOPS | 15.97 TFLOPS |

| FP16 | 22.30 TFLOPS (2:1) | 15.97 TFLOPS (1:1) |

| TDP | 250 W | 115 W |

| Bus Interface | PCIe 3.0 x16 | PCIe 4.0 x16 |

| Display Outputs | 1x HDMI 2.0, 3x DisplayPort 1.4a, 1x USB Type-C | Portable Device Dependent |

| Slot Width | Dual-slot | Not specified |

| Power Connectors | 1x 6-pin + 1x 8-pin | None |

| Suggested PSU | 600 W | Not specified |

| Release Date | 2019-07-22 | 2021-01-11 |

The RTX 2080 SUPER is a dual-slot desktop card measuring 267 mm in length, 116 mm in height, and 35 mm in width. The RTX 3070 Mobile has no listed dimensions, as it is a mobile part. The desktop card requires a 600 W suggested power supply and uses one 6-pin and one 8-pin connector. The mobile GPU lists no power connectors and has a TDP of 115 W versus 250 W for the desktop card.

Head-to-Head Benchmarks

The RTX 3070 Mobile takes the single most modern test: 3DMark Steel Nomad DX12. It scores 2380 against 1882, a 20.9% margin. This is the only benchmark where the mobile Ampere part leads, and it suggests that under DX12 rendering loads with modern geometry, the newer architecture compensates for its lower clocks and mobile power envelope.

Everywhere else, the RTX 2080 SUPER leads, often by large margins. The biggest win is in Passmark G2D, where the desktop card scores 920 versus 641, a 43.5% advantage. That is followed by Passmark DirectX 9, where it wins 227 to 160, a 41.9% lead. In Geekbench Vulkan, the RTX 2080 SUPER scores 111284 versus 86768, a 28.3% win. Passmark G3D shows a 27.3% lead (19490 vs 15309). Passmark DirectX 10 gives a 23.9% advantage (140 vs 113). Passmark GPU Compute shows a 21.4% lead (8290 vs 6827). Passmark DirectX 11 is 19.6% in favor of the desktop card (165 vs 138). Passmark DirectX 12 is 17.2% (75 vs 64). Geekbench OpenCL is the closest win for the RTX 2080 SUPER at 6.8% (99226 vs 92939).

The pattern is clear: the RTX 2080 SUPER dominates in legacy APIs, compute, and 2D tasks, while the RTX 3070 Mobile excels specifically in modern DX12 rendering as measured by Steel Nomad.

The Verdict

The data points to two different use cases. If the workload involves legacy DirectX titles, Vulkan applications, OpenCL compute, or any 2D acceleration, the RTX 2080 SUPER is the better performer. Its wins range from 6.8% to 43.5% across nine benchmarks, and its average benchmark score of 24170 places it in the 69th percentile of all GPUs, above the RTX 3070 Mobile's 65th percentile and 20534 average.

The RTX 3070 Mobile, however, wins the one test that matters most for current-generation DX12 gaming: 3DMark Steel Nomad. Its 20.9% lead there shows that Ampere's architecture, despite lower boost clocks (1560 MHz versus 1815 MHz) and a lower 115 W TDP, is more efficient at handling modern DX12 workloads. The mobile part also has a higher pixel rate at 124.8 GPixel/s and more ROPs, which helps in fill-rate-bound scenarios.

For a desktop user with a fixed power budget, the RTX 2080 SUPER is the safer choice for broad compatibility and compute performance. For someone prioritizing modern DX12 gaming in a mobile form factor, the RTX 3070 Mobile's Steel Nomad result is the more relevant data point. The RTX 2080 SUPER launched at an MSRP of 699 USD, while the RTX 3070 Mobile has no listed launch MSRP. Choose based on the workload: legacy and compute favor Turing, modern DX12 favors Ampere.