AMD Radeon HD 7790 vs NVIDIA GeForce RTX 3070 Mobile Comparison

Where Each One Wins

The recorded data paints a stark picture of two GPUs from completely different eras. The NVIDIA GeForce RTX 3070 Mobile is the clear winner in nearly every measurable category, but the AMD Radeon HD 7790 holds a niche advantage in one specific area.

The RTX 3070 Mobile dominates in raw compute, modern API support, and memory bandwidth. Its benchmark scores across DirectX 10, 11, 12, and 9 tests, plus its OpenCL and Vulkan results, all land far above anything the HD 7790 can produce. The database shows the RTX 3070 Mobile achieving an average benchmark score of 20,534, while the HD 7790 sits at 17,666. That is a substantial gap, roughly 16% higher average performance for the NVIDIA part.

The HD 7790, however, wins in one specific test category: Apple's Metal API. The database records a Geekbench Metal score of 17,666 for the AMD card. The RTX 3070 Mobile has no recorded Metal benchmark in the database, which means the AMD card is the only one of the two with any measurable presence in that ecosystem. For anyone running Metal-based workloads, typically on macOS or iOS development environments, the HD 7790 is the only option with data backing it.

Beyond that single API, the RTX 3070 Mobile wins everywhere else. The Passmark G3D score of 15,309 for the NVIDIA part versus no comparable G3D score for the AMD part shows a massive performance gulf. The RTX 3070 Mobile also records a Passmark GPU Compute score of 6,827, while the HD 7790 has no such entry. The 3DMark Steel Nomad DX12 score of 2,380 for the RTX 3070 Mobile further cements its position as the modern performer.

The wins break down by use case. For gaming, content creation, machine learning inference, or any DirectX 12 Ultimate workload, the RTX 3070 Mobile is the only sensible choice. For legacy Metal compatibility or low-power embedded applications, the HD 7790 has a place, but it is a narrow one.

Architecture Differences

The architectural gap between these two GPUs spans two console generations and a fundamental shift in graphics processing design.

The RTX 3070 Mobile uses the GA104 chip built on NVIDIA's Ampere architecture. It is fabricated on an 8 nm process at Samsung, with 17,400 million transistors packed into a 392 mm² die. That works out to a transistor density of 44.4 million per square millimeter. The memory subsystem uses 8 GB of GDDR6 on a 256-bit bus, delivering 448.0 GB/s of bandwidth. The chip carries 5,120 shading units, 160 texture mapping units, 80 ROPs, 40 ray tracing cores, and 160 tensor cores. Pixel rate is 124.8 GPixel/s, texture rate is 249.6 GTexel/s, and FP32 compute reaches 15.97 TFLOPS. FP16 runs at the same 15.97 TFLOPS on a 1:1 basis. The card supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

The HD 7790 uses the Bonaire chip based on AMD's GCN 2.0 architecture. It is built on a 28 nm process at TSMC, with 2,080 million transistors on a 160 mm² die. Transistor density is 13.0 million per square millimeter. The memory subsystem is far more modest: 1024 MB of GDDR5 on a 128-bit bus, yielding 96.00 GB/s of bandwidth. The chip has 896 shading units, 56 TMUs, and 16 ROPs. There are no ray tracing cores and no tensor cores. Pixel rate is 16.00 GPixel/s and texture rate is 56.00 GTexel/s. FP32 compute is 1.792 TFLOPS. There is no recorded FP16 performance. The HD 7790 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.2.170.

The differences are enormous. The RTX 3070 Mobile has roughly 8.9 times the FP32 compute, 4.7 times the memory bandwidth, and 7.4 times the texture fill rate. It also has hardware ray tracing and tensor cores, neither of which exist on the HD 7790. The process node advantage (8 nm versus 28 nm) explains much of the efficiency and density gap, but the architectural leap from GCN 2.0 to Ampere is the bigger story.

The HD 7790 does have one thing the RTX 3070 Mobile lacks: a documented Metal API score. That suggests AMD's GCN architecture retains some compatibility with Apple's graphics stack, which the NVIDIA Ampere mobile part does not appear to have in the database.

Head-to-Head Benchmarks

Direct head-to-head benchmark comparisons are limited because the two GPUs share almost no common test entries. The database lists no head-to-head results, so the comparison relies on the individual benchmark lists for each card.

The RTX 3070 Mobile delivers a Passmark G3D score of 15,309. The HD 7790 has no G3D score recorded, so the NVIDIA part wins that category by default and by a massive margin. The RTX 3070 Mobile also posts a Passmark GPU Compute score of 6,827, which is more than 3.8 times the HD 7790's entire average benchmark score of 17,666. That compute score reflects the tensor core and FP32 advantage of the Ampere architecture.

In DirectX 11, the RTX 3070 Mobile scores 138 on Passmark, while the HD 7790 has no DirectX 11 entry. In DirectX 12, the RTX 3070 Mobile scores 64, and again the HD 7790 has no comparable entry. The same applies to DirectX 9 (160) and DirectX 10 (113). The RTX 3070 Mobile also records a 3DMark Steel Nomad DX12 score of 2,380, an OpenCL score of 92,939, and a Vulkan score of 86,768.

The HD 7790's only benchmark entry is Geekbench Metal at 17,666. That score sits right at its average benchmark score, which makes sense given it has only one recorded test. The RTX 3070 Mobile has no Metal score, so on that specific API, the AMD card is the only data point.

When comparing average benchmark scores, the RTX 3070 Mobile at 20,534 is 16.2% ahead of the HD 7790 at 17,666. But that average includes the NVIDIA card's multiple DirectX and compute tests, while the AMD card's average is based solely on its Metal score. A more apples-to-apples comparison would require the same test suite run on both cards, which the database does not provide.

The nearest rivals in the database put both cards in context. The RTX 3070 Mobile sits at the 65th percentile among all GPUs, with its closest competitor being the Intel Arc B570 at 20,556 (0.1% faster) and the NVIDIA Quadro M4000M at 20,480 (0.3% slower). The HD 7790 sits at the 61st percentile, with its nearest rival being the NVIDIA GeForce RTX 4060 at 17,639 (0.2% faster) and the AMD Radeon 780M at 17,588 (0.4% faster). That is a remarkable pairing: the HD 7790's closest rivals are modern integrated and entry-level discrete GPUs, not the high-end mobile part it is being compared against here.

FAQ

Q: Is the RTX 3070 Mobile faster than the HD 7790 in all tests?

A: In every shared or individually recorded test category, the RTX 3070 Mobile wins. The only exception is Metal, where the HD 7790 has a recorded score (17,666) and the RTX 3070 Mobile has none.

Q: What is the biggest performance gap between the two cards?

A: The Passmark GPU Compute score shows the largest gap: the RTX 3070 Mobile scores 6,827, while the HD 7790 has no recorded compute score. In raw FP32, the RTX 3070 Mobile delivers 15.97 TFLOPS versus 1.792 TFLOPS for the HD 7790, an 8.9x advantage.

Q: Does the HD 7790 support ray tracing?

A: No. The HD 7790 has no ray tracing cores and no tensor cores in its Bonaire chip. The RTX 3070 Mobile has 40 ray tracing cores and 160 tensor cores.

Q: Which card has better memory bandwidth?

A: The RTX 3070 Mobile has 448.0 GB/s of bandwidth over a 256-bit GDDR6 interface. The HD 7790 has 96.00 GB/s over a 128-bit GDDR5 interface. The NVIDIA part has 4.7 times the bandwidth.

Q: Are these cards still in production?

A: Both are marked as end-of-life in the database. The RTX 3070 Mobile was released in January 2021, and the HD 7790 was released in March 2013.

Q: Which card has better driver and API support?

A: The RTX 3070 Mobile supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The HD 7790 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.2.170. The NVIDIA part has the newer DirectX and Vulkan versions, but the HD 7790 has Metal support that the RTX 3070 Mobile lacks.

The Verdict

The data is unambiguous. The RTX 3070 Mobile is the superior GPU for virtually every modern workload. Its 15.97 TFLOPS of FP32 compute, 448.0 GB/s of memory bandwidth, and 40 ray tracing cores place it in a different performance class entirely. The HD 7790, with 1.792 TFLOPS and 96.00 GB/s, is a relic from the early GCN era.

For gamers, the RTX 3070 Mobile is the only option with meaningful DirectX 12 Ultimate support. For compute workloads, the tensor cores and 6,827 Passmark GPU Compute score make it the clear choice. For anyone needing Vulkan 1.4 or DirectX 12_2 features, the RTX 3070 Mobile wins outright.

The HD 7790's only justification is Metal compatibility. If a system requires that specific Apple API and no other GPU is available, the HD 7790 at least has a recorded score of 17,666. But that is a narrow edge case, not a recommendation.

The percentile rankings confirm the gap: the RTX 3070 Mobile sits at the 65th percentile of all GPUs, while the HD 7790 sits at the 61st. The average benchmark score difference (20,534 versus 17,666) is real but understates the practical difference, because the RTX 3070 Mobile's scores come from a much more demanding and diverse test suite.

Anyone building a modern system should pick the RTX 3070 Mobile. Anyone maintaining a legacy Metal-based workflow might consider the HD 7790, but only if the software stack cannot tolerate anything else.

Specification Differences

| Specification | NVIDIA GeForce RTX 3070 Mobile | AMD Radeon HD 7790 |

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

| Chip | GA104 | Bonaire |

| Architecture | Ampere | GCN 2.0 |

| Process Node | 8 nm | 28 nm |

| Foundry | Samsung | TSMC |

| Transistors | 17,400 million | 2,080 million |

| Die Size | 392 mm² | 160 mm² |

| Transistor Density | 44.4M / mm² | 13.0M / mm² |

| Memory Size | 8 GB | 1024 MB |

| Memory Type | GDDR6 | GDDR5 |

| Memory Bus Width | 256 bit | 128 bit |

| Memory Bandwidth | 448.0 GB/s | 96.00 GB/s |

| Memory Clock | 1750 MHz, 14 Gbps effective | 1500 MHz, 6 Gbps effective |

| Shading Units | 5120 | 896 |

| TMUs | 160 | 56 |

| ROPs | 80 | 16 |

| RT Cores | 40 | None |

| Tensor Cores | 160 | None |

| Pixel Rate | 124.8 GPixel/s | 16.00 GPixel/s |

| Texture Rate | 249.6 GTexel/s | 56.00 GTexel/s |

| FP32 | 15.97 TFLOPS | 1.792 TFLOPS |

| FP16 | 15.97 TFLOPS (1:1) | None recorded |

| TDP | 115 W | 85 W |

| Slot Width | Not recorded | Dual-slot |

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

| Suggested PSU | Not recorded | 250 W |

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

| Display Outputs | Portable Device Dependent | 1x DVI, 1x HDMI 1.4a, 2x mini-DisplayPort 1.2 |

| DirectX | 12 Ultimate (12_2) | 12 (12_0) |

| OpenGL | 4.6 | 4.6 |

| Vulkan | 1.4 | 1.2.170 |

| Release Date | 2021-01-11 | 2013-03-21 |

| Production Status | End-of-life | End-of-life |

| Predecessor | GeForce 20 Mobile | Northern Islands |

| Successor | None recorded | Sea Islands |

| Launch MSRP | None recorded | 149 USD |

| Average Benchmark Score | 20,534 | 17,666 |

| Percentile | 65 | 61 |