AMD Radeon Pro Vega 20 vs AMD Radeon RX 470 Comparison

The AMD Radeon RX 470 and the AMD Radeon Pro Vega 20 are both end-of-life GPUs from AMD, but they target very different segments and use different architectural generations. The RX 470 is a discrete, dual-slot desktop card built on GCN 4.0, while the Pro Vega 20 is an integrated-class (IGP) mobile processor designed for Apple Mac systems, built on the newer GCN 5.0 architecture. Despite the Pro Vega 20’s architectural advantage, benchmark data shows the RX 470 delivers significantly higher performance in every common workload, with the data indicating a clear, though not uniform, victory for the older discrete card.

Where Each One Wins

The performance split is entirely one-sided in favor of the AMD Radeon RX 470. Across all three head-to-head benchmarks available—Geekbench Metal, OpenCL, and Vulkan—the RX 470 records a win, giving it a 3-0 sweep over the Pro Vega 20. This means there is no single benchmark category where the Pro Vega 20 can claim an advantage; its strengths lie elsewhere, primarily in its form factor and power envelope.

The RX 470’s wins are substantial and consistent. Its largest margin comes in the Vulkan test, where it scores 39,884 versus the Pro Vega 20’s 26,410, a 51% advantage. In the Metal API, the RX 470 leads by 37%, and in OpenCL, it is ahead by 25.8%. These margins indicate that the RX 470 is the stronger compute and graphics performer across different API ecosystems, making it the better choice for raw throughput in gaming or general-purpose GPU workloads.

The Pro Vega 20, conversely, wins in the categories not captured by raw benchmark scores: it is an IGP with a 100 W TDP, no external power connectors, and portable-device-dependent display outputs. This makes it a fit for thin-and-light laptops where the RX 470’s dual-slot, 120 W design and 1x 6-pin power connector would be impractical. The data shows the Pro Vega 20’s advantage is not in speed but in integration and power efficiency, though it still offers 73rd percentile performance among all GPUs, placing it in a competitive position for its class.

Architecture Differences

The two GPUs represent different generations of AMD’s Graphics Core Next (GCN) architecture. The RX 470 is built on GCN 4.0, using the Ellesmere chip, and belongs to the Arctic Islands (RX 400) generation. The Pro Vega 20 uses the newer GCN 5.0 architecture with the Vega 12 chip, part of the Radeon Pro Mac (Vega Series) generation. Both are manufactured on a 14 nm process at GlobalFoundries, so the process node is identical.

The chip configurations differ significantly. The RX 470 has 2,048 shading units, 128 texture mapping units (TMUs), and 32 ROPs. The Pro Vega 20 has fewer execution resources: 1,280 shading units, 80 TMUs, and 32 ROPs. This 60% difference in shading units and TMUs explains much of the RX 470’s performance lead, despite its older architecture. The Pro Vega 20 compensates with a higher boost clock of 1283 MHz versus the RX 470’s 1206 MHz, but this does not overcome the core count deficit.

Memory architecture is a major differentiator. The RX 470 uses 4 GB of GDDR5 on a 256-bit bus, delivering 211.2 GB/s of bandwidth. The Pro Vega 20 uses 4 GB of HBM2 on a much wider 1024-bit bus, but its memory clock is only 740 MHz (1480 Mbps effective), yielding 189.4 GB/s. Thus, the RX 470 actually has higher memory bandwidth despite the narrower bus, a key factor in its benchmark wins. The Pro Vega 20’s HBM2 does offer a 2:1 FP16 ratio (6.569 TFLOPS) versus the RX 470’s 1:1 ratio (4.940 TFLOPS), making it potentially faster for half-precision compute, though no benchmark data confirms this advantage.

Head-to-Head Benchmarks

The Geekbench Metal test shows the RX 470 at 41,690 points against the Pro Vega 20’s 30,427, a 37% delta. This is a significant margin for a graphics API commonly used in macOS environments, where the Pro Vega 20 is designed to operate. The RX 470’s higher shading unit count and memory bandwidth likely drive this outcome, as Metal workloads often scale with these resources.

In Geekbench OpenCL, the gap narrows but remains decisive. The RX 470 scores 33,568 versus 26,679, a 25.8% advantage. OpenCL is a general-purpose compute API, and here the Pro Vega 20’s newer GCN 5.0 architecture with its FP16 capabilities does not translate into a win. The RX 470’s higher FP32 throughput (4.940 TFLOPS versus 3.284 TFLOPS) is the likely explanation, as most OpenCL workloads target FP32 precision.

The largest delta appears in Geekbench Vulkan, where the RX 470 achieves 39,884 against the Pro Vega 20’s 26,410, a 51% lead. This is the most lopsided result and suggests that the RX 470’s architecture handles Vulkan’s explicit, low-overhead model substantially better. The RX 470’s nearest rival in the average benchmark score list is the Intel Arc A370M with a delta of -0.6%, meaning the RX 470’s Vulkan performance is not an outlier but part of a broader pattern of strength. The Pro Vega 20’s nearest rival, the NVIDIA GeForce RTX 3090, shows a delta of 1% in its favor, indicating the Pro Vega 20’s average score is very close to that high-end card, yet it cannot match the RX 470 in any direct test.

FAQ

Q: Which GPU is faster in Geekbench Vulkan?

A: The AMD Radeon RX 470 wins decisively with a score of 39,884 versus the Pro Vega 20’s 26,410, a 51% advantage.

Q: Does the Pro Vega 20 have a higher memory bandwidth than the RX 470?

A: No. Despite using HBM2 on a 1024-bit bus, the Pro Vega 20’s bandwidth is 189.4 GB/s, which is lower than the RX 470’s 211.2 GB/s from GDDR5 on a 256-bit bus.

Q: What is the difference in shading unit counts?

A: The RX 470 has 2,048 shading units, while the Pro Vega 20 has 1,280, a difference of 768 units in favor of the RX 470.

Q: Which GPU has a higher TDP?

A: The RX 470 has a 120 W TDP, while the Pro Vega 20 is rated at 100 W, making the Pro Vega 20 the lower-power part.

Q: Are both GPUs based on the same architecture?

A: No. The RX 470 uses GCN 4.0, while the Pro Vega 20 uses GCN 5.0, though both are fabricated on a 14 nm process.

Q: How do the average benchmark scores compare?

A: The RX 470 has an average benchmark score of 28,996, placing it in the 74th percentile of all GPUs. The Pro Vega 20 averages 27,839, in the 73rd percentile.

Specification Differences

| Specification | AMD Radeon RX 470 | AMD Radeon Pro Vega 20 |

|:---------------|:------------------|:-----------------------|

| Chip | Ellesmere | Vega 12 |

| Architecture | GCN 4.0 | GCN 5.0 |

| Generation | Arctic Islands (RX 400) | Radeon Pro Mac (Vega Series) |

| Transistors | 5,700 million | Not specified |

| Die Size | 232 mm² | Not specified |

| Transistor Density | 24.6M / mm² | Not specified |

| Base Clock | 926 MHz | 815 MHz |

| Boost Clock | 1206 MHz | 1283 MHz |

| Memory Clock | 1650 MHz / 6.6 Gbps effective | 740 MHz / 1480 Mbps effective |

| Memory Type | GDDR5 | HBM2 |

| Memory Bus Width | 256 bit | 1024 bit |

| Memory Bandwidth | 211.2 GB/s | 189.4 GB/s |

| Shading Units | 2048 | 1280 |

| TMUs | 128 | 80 |

| ROPs | 32 | 32 |

| Pixel Rate | 38.59 GPixel/s | 41.06 GPixel/s |

| Texture Rate | 154.4 GTexel/s | 102.6 GTexel/s |

| FP32 Performance | 4.940 TFLOPS | 3.284 TFLOPS |

| FP16 Performance | 4.940 TFLOPS (1:1) | 6.569 TFLOPS (2:1) |

| TDP | 120 W | 100 W |

| Slot Width | Dual-slot | IGP |

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

| Suggested PSU | 300 W | None |

| Display Outputs | 1x HDMI 2.0b, 3x DisplayPort 1.4a | Portable Device Dependent |

| DirectX Support | 12 (12_0) | 12 (12_1) |

| Dimensions (LxHxW) | 240 mm x 95 mm x 35 mm | Not specified |

| Release Date | 2016-08-03 | 2018-11-13 |

| Launch MSRP | 179 USD | None |

The specification table shows the RX 470’s clear resource advantages in shading units, TMUs, FP32 throughput, and memory bandwidth. The Pro Vega 20 counters with a higher boost clock, higher pixel rate, superior FP16 performance, and a lower TDP, but these attributes do not translate into benchmark wins. The RX 470 also offers a higher DirectX feature level (12_1 versus 12_0) and a full suite of display outputs, while the Pro Vega 20 is dependent on the host device. The data overall suggests that for any performance-sensitive task, the RX 470 is the superior part, while the Pro Vega 20 is optimized for integration and efficiency in a mobile context.