AMD Radeon Pro Vega 20 vs AMD Radeon RX 590 Comparison

The AMD Radeon Pro Vega 20 and AMD Radeon RX 590 launched one day apart in November 2018, yet they represent two entirely different philosophies from the same manufacturer: a power-frugal workstation part built on HBM2 memory, and a mainstream desktop card chasing raw throughput. The recorded data points to a clear split. The RX 590 dominates the one shared benchmark in the database, while the Pro Vega 20 counters with a newer architecture, a higher percentile ranking against all GPUs, and a dramatically lower power envelope. Which one makes sense depends entirely on the workload, because these two cards are barely rivals despite what their release dates suggest.

Head-to-Head Benchmarks

The head-to-head record is short but decisive. In Geekbench Metal, the only test both cards have recorded in the database, the Radeon RX 590 scores 48407 against 30427 for the Pro Vega 20, a gap of 37.1 percent in the RX 590's favor. That is not a marginal win, it is a rout in the single most directly comparable test the two share.

The disparity becomes more interesting when placed in context. The Pro Vega 20's Metal score of 30427 still lands it at the 73rd percentile against all GPUs in the database, with an average benchmark score of 27839. Its nearest rivals cluster remarkably tightly around it: the Radeon RX 7800M sits within 0.2 percent, the Radeon Pro W5500X within 0.5 percent, the GeForce GTX 980 Ti within 0.6 percent, and even the GeForce RTX 3090 appears in its rival list at just 1 percent behind. In other words, the database places this modest 100 W part in the same overall neighborhood as some genuinely heavyweight silicon.

The RX 590, despite its commanding head-to-head win, sits at the 70th percentile with an average benchmark score of 24744. Its recorded tests include the Metal win above plus a 3D Mark Steel Nomad DX12 score of 1081, a test the Pro Vega 20 does not appear in. Its rivals include the RTX A5000 Mobile at essentially a dead heat (-0.1 percent), the Arc A350M at 0.4 percent, the Radeon RX 6600 XT at 1.2 percent, and the GeForce GTX 1630 at 1.9 percent behind.

So the headline tension is this: the RX 590 wins the only direct matchup 48407 to 30427, yet the Pro Vega 20 holds the slightly higher all-GPU percentile, 73 versus 70. The Metal test rewards the RX 590's far greater shading resources; the aggregate database view, which leans on the Vega's stronger compute API results in OpenCL (26679) and Vulkan (26410), narrows the overall picture considerably.

Architecture Differences

These cards diverge at nearly every architectural level, and the differences explain the benchmark outcomes.

The Pro Vega 20 uses the Vega 12 chip on GCN 5.0, manufactured on a 14 nm GlobalFoundries process. The RX 590 uses Polaris 30 on GCN 4.0, built on a 12 nm GlobalFoundries process. Despite the older architecture, the RX 590's newer node lets it run far higher clocks: 1469 MHz base and 1545 MHz boost, against 815 MHz base and 1283 MHz boost for the Vega 20. That clock gap, combined with a large shader advantage, is the engine behind the RX 590's Metal win.

The shader counts tell the biggest story. The RX 590 fields 2304 shading units, 144 TMUs, and 32 ROPs. The Pro Vega 20 has just 1280 shading units and 80 TMUs, though it matches the 32 ROPs. The result shows up directly in the throughput figures: the RX 590 delivers 7.119 TFLOPS of FP32 compute, 222.5 GTexel/s of texture rate, and 49.44 GPixel/s of pixel rate. The Vega 20 manages 3.284 TFLOPS FP32, 102.6 GTexel/s, and 41.06 GPixel/s. The RX 590 more than doubles texture throughput.

But the Vega 20 has a trump card in half-precision work: 6.569 TFLOPS FP16 at a 2:1 ratio, double its FP32 rate. The RX 590's FP16 runs at 1:1, offering no half-precision benefit at 7.119 TFLOPS. For workloads that exploit FP16, the compute gap between the two cards shrinks from massive to nearly closed.

Memory design is where the philosophies split hardest. The Pro Vega 20 pairs 4 GB of HBM2 on a 1024-bit bus running at 740 MHz (1480 Mbps effective) for 189.4 GB/s of bandwidth. The RX 590 uses 8 GB of GDDR5 on a 256-bit bus at 2000 MHz (8 Gbps effective) for 256.0 GB/s. The RX 590 wins both capacity and raw bandwidth, but the Vega's HBM2 achieves its figure with a fraction of the power budget.

Power and form factor underline the difference. The Pro Vega 20 is an integrated-class part (IGP slot width) with a 100 W TDP and no power connectors. The RX 590 is a dual-slot card measuring 241 mm long, with a 175 W TDP, one 8-pin connector, and a suggested 450 W PSU. The RX 590 offers fixed display outputs (one HDMI 2.0b, three DisplayPort 1.4a) while the Vega 20's outputs are portable-device dependent. Both share PCIe 3.0 x16 and Vulkan 1.3 with OpenGL 4.6; the Vega 20 supports DirectX 12 at feature level 12_1, a step above the RX 590's 12_0. Neither card has RT or tensor cores.

Where Each One Wins

The RX 590 wins wherever raw shading and texture throughput matter. Its 37.1 percent Geekbench Metal advantage, built on 2304 shaders running at 1545 MHz boost, makes it the clear choice for gaming and general graphics work. The 222.5 GTexel/s texture rate and 49.44 GPixel/s pixel rate both comfortably exceed the Vega 20's figures, and its 8 GB frame buffer doubles the Vega's 4 GB, an important margin for memory-hungry scenes. Its 256.0 GB/s of bandwidth also outpaces the Vega's 189.4 GB/s.

The Pro Vega 20 wins on efficiency and specific compute paths. It delivers its 73rd percentile standing from a 100 W TDP inside an integrated form factor, needing no power connector and no dedicated PSU headroom. Its FP16 throughput of 6.569 TFLOPS, double its FP32 rate, means certain compute workloads run at nearly the RX 590's level despite the massive shader deficit. Its Geekbench OpenCL (26679) and Vulkan (26410) results show consistent compute performance across APIs. And its rival neighborhood, including cards like the GTX 980 Ti and RTX 3090 within roughly a percent, indicates well-rounded aggregate standing.

The RX 590's additional 3D Mark Steel Nomad DX12 result of 1081 gives it a modern DirectX 12 gaming data point the Vega 20 lacks entirely in the database.

The Verdict

For gaming, the data is unambiguous: the RX 590 is the pick. It wins the only shared benchmark by 37.1 percent, doubles the shader count, and leads in every throughput metric that matters for real-time graphics, all while carrying twice the VRAM. Its launch MSRP was 279 USD.

The Pro Vega 20 is not a gaming card, and treating it as one misreads it. It is a compute and efficiency play: a 100 W integrated part that reaches the 73rd percentile overall, beats the RX 590's percentile standing (70), and delivers doubled FP16 throughput that closes the compute gap in suited workloads. Its HBM2 subsystem, wide 1024-bit bus, and portable-device-dependent outputs mark it as a component for compact professional systems rather than desktops.

Both are end-of-life products, so this comparison is retrospective. On the evidence recorded in the database, the RX 590 was the stronger graphics card; the Pro Vega 20 was the more efficient and architecturally forward compute part.

FAQ

Q: Which card is faster in benchmarks?

A: The RX 590. In Geekbench Metal, the only head-to-head test recorded, it scores 48407 versus 30427 for the Pro Vega 20, a 37.1 percent advantage.

Q: How do their overall database rankings compare?

A: The Pro Vega 20 sits at the 73rd percentile with an average score of 27839; the RX 590 sits at the 70th percentile with an average score of 24744. The Vega's higher percentile comes from strong OpenCL and Vulkan compute results despite losing the Metal test.

Q: Which card uses less power?

A: The Pro Vega 20, by a wide margin: a 100 W TDP with no power connectors, versus a 175 W TDP with an 8-pin connector and a suggested 450 W PSU for the RX 590.

Q: How do the memory configurations differ?

A: The Pro Vega 20 has 4 GB of HBM2 on a 1024-bit bus delivering 189.4 GB/s. The RX 590 has 8 GB of GDDR5 on a 256-bit bus delivering 256.0 GB/s. The RX 590 offers more capacity and bandwidth.

Q: Does either card support ray tracing hardware?

A: No. Both the Pro Vega 20 and the RX 590 lack RT cores and tensor cores.

Q: Which cards are their closest rivals?

A: The Pro Vega 20's nearest rivals include the Radeon RX 7800M (-0.2 percent) and Pro W5500X (-0.5 percent). The RX 590's include the RTX A5000 Mobile (-0.1 percent) and Arc A350M (0.4 percent).

Specification Differences

| Field | AMD Radeon Pro Vega 20 | AMD Radeon RX 590 |

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

| Chip | Vega 12 | Polaris 30 |

| Architecture | GCN 5.0 | GCN 4.0 |

| Generation | Radeon Pro Mac (Vega Series) | Polaris (RX 500) |

| Process Node | 14 nm | 12 nm |

| Foundry | GlobalFoundries | GlobalFoundries |

| Base Clock | 815 MHz | 1469 MHz |

| Boost Clock | 1283 MHz | 1545 MHz |

| Memory Size | 4 GB HBM2 | 8 GB GDDR5 |

| Bus Width | 1024 bit | 256 bit |

| Memory Bandwidth | 189.4 GB/s | 256.0 GB/s |

| Memory Speed | 740 MHz (1480 Mbps effective) | 2000 MHz (8 Gbps effective) |

| Shading Units | 1280 | 2304 |

| TMUs | 80 | 144 |

| ROPs | 32 | 32 |

| Pixel Rate | 41.06 GPixel/s | 49.44 GPixel/s |

| Texture Rate | 102.6 GTexel/s | 222.5 GTexel/s |

| FP32 | 3.284 TFLOPS | 7.119 TFLOPS |

| FP16 | 6.569 TFLOPS (2:1) | 7.119 TFLOPS (1:1) |

| TDP | 100 W | 175 W |

| Slot Width | IGP | Dual-slot |

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

| Suggested PSU | None listed | 450 W |

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

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

| Length | Not listed | 241 mm (9.5 inches) |

| Transistors | Not listed | 5,700 million |

| Die Size | Not listed | 232 mm² |

| Launch MSRP | Not listed | 279 USD |

| Percentile vs All GPUs | 73 | 70 |

| Average Benchmark Score | 27839 | 24744 |