AMD Radeon RX 580 vs NVIDIA RTX PRO 6000 Blackwell Max-Q Comparison

FAQ

Q: How do the two cards compare in raw benchmark performance?

A: The database contains a shared DirectX 12 workload, 3DMark Steel Nomad. The AMD Radeon RX 580 scores 1005, while the NVIDIA RTX PRO 6000 Blackwell Max-Q scores 11088. That is a 90.9% difference in favor of the NVIDIA card.

Q: Which card has more memory, and does it matter?

A: The RX 580 has 8 GB of GDDR5 on a 256 bit bus, delivering 256.0 GB/s. The RTX PRO 6000 has 96 GB of GDDR7 on a 512 bit bus, delivering 1.79 TB/s. The bandwidth difference is massive, roughly 7 times higher on the NVIDIA card.

Q: Are these cards in the same performance class?

A: No. The RX 580 sits at the 53rd percentile of all GPUs in the database. The RTX PRO 6000 sits at the 50th percentile. However, that percentile is based on very different benchmark coverage, the RX 580 has 11 recorded tests while the RTX PRO 6000 has only one.

Q: What are the closest rivals for each card?

A: The RX 580's nearest rivals by average score are the GeForce RTX 3050 Ti Mobile (0.1% behind), the GTX 1660 SUPER (0.4% behind), the Radeon 740M (0.5% ahead), and the FirePro W5100 (0.6% ahead). The RTX PRO 6000's nearest rivals are the RTX PRO 6000D Blackwell Max-Q (identical score), the Radeon RX 550 (0.1% ahead), the GTX 1650 SUPER (0.4% ahead), and the FirePro W4300 (1.2% behind).

Q: Which card is still in production?

A: The RX 580 is listed as end-of-life. The RTX PRO 6000 Blackwell Max-Q is listed as active production.

Q: What interface does each card use?

A: The RX 580 uses PCIe 3.0 x16. The RTX PRO 6000 uses PCIe 5.0 x16.

The Verdict

The data presents a clear split. Anyone running modern DirectX 12 workloads, the only test shared by both cards, should pick the NVIDIA RTX PRO 6000 Blackwell Max-Q. Its 11088 score dwarfs the RX 580's 1005, a 90.9% gap. The NVIDIA card also brings 96 GB of GDDR7 memory, 1.79 TB/s of bandwidth, and 188 RT cores, none of which the RX 580 has.

The RX 580 still has a role. It is a 14 nm Polaris part from 2017, end-of-life, with 8 GB of GDDR5. Its 53rd percentile placement and 12928 average benchmark score across 11 tests put it near cards like the GTX 1660 SUPER and RTX 3050 Ti Mobile. For older DirectX workloads, PassMark scores show strength: DirectX 9 at 124, DirectX 11 at 60, and DirectX 10 at 46. Those are legacy scenarios. The RTX PRO 6000 has no recorded scores for those tests, so comparisons there are impossible.

The verdict is straightforward. The RTX PRO 6000 is the choice for current-generation rendering, ray tracing, and large memory workloads. The RX 580 is a legacy card that still competes in its own class, but it is not a competitor to the Blackwell part in any shared modern benchmark.

Head-to-Head Benchmarks

The database records exactly one shared test between these two cards: 3DMark Steel Nomad under DirectX 12. The RX 580 scores 1005. The RTX PRO 6000 scores 11088. The delta is 90.9% in favor of NVIDIA.

That single result defines the head-to-head. The RX 580 wins zero shared tests. The RTX PRO 6000 wins one. There is no shared workload where the AMD card comes out ahead.

Looking at the RX 580's own benchmark suite, its strongest results are in legacy APIs. PassMark DirectX 9 gives it 124, DirectX 11 gives 60, DirectX 10 gives 46, and DirectX 12 gives 43. Its Geekbench scores are 45235 in Metal, 45173 in Vulkan, and 37453 in OpenCL. Its PassMark G3D score is 8813, and its compute score is 3488. Average across all 11 tests is 12928.

The RTX PRO 6000 has only the one recorded benchmark. That makes direct comparison limited. What the data does show is that in the one test both cards ran, the NVIDIA card is an order of magnitude faster.

For context, the RX 580's nearest rival, the GTX 1660 SUPER, averages 12986, only 0.4% ahead. The RTX PRO 6000's nearest rival, the RTX PRO 6000D Blackwell Max-Q, matches its 11088 exactly. The RX 550 sits 0.1% ahead of the RTX PRO 6000 in average score, and the GTX 1650 SUPER sits 0.4% ahead. Those deltas look odd on paper, but they reflect very different benchmark sets, not real performance equivalence.

Specification Differences

The specification gap is enormous. The RX 580 uses the Polaris 14 nm chip from GlobalFoundries with 5,700 million transistors on a 232 mm² die. The RTX PRO 6000 uses the GB202 chip from TSMC on a 5 nm node, with 92,200 million transistors on a 750 mm² die. Transistor density is 24.6M per mm² for AMD and 122.9M per mm² for NVIDIA.

Memory is 8 GB GDDR5 versus 96 GB GDDR7. Bus width is 256 bit versus 512 bit. Bandwidth is 256.0 GB/s versus 1.79 TB/s. Clocks differ: the RX 580 runs 1257 MHz base and 1340 MHz boost, while the RTX PRO 6000 runs 1035 MHz base and 2280 MHz boost.

Compute resources differ sharply. The RX 580 has 2304 shading units, 144 TMUs, and 32 ROPs. The RTX PRO 6000 has 24064 shading units, 752 TMUs, and 192 ROPs. It also has 188 RT cores and 752 tensor cores, features absent from the RX 580 entirely.

Pixel rate is 42.88 GPixel/s for AMD versus 437.8 GPixel/s for NVIDIA. Texture rate is 193.0 GTexel/s versus 1,714.6 GTexel/s. FP32 is 6.175 TFLOPS versus 6.175 TFLOPS? No, that is wrong. The RX 580 is 6.175 TFLOPS, the RTX PRO 6000 is 109.7 TFLOPS. FP16 is 6.175 TFLOPS (1:1) for AMD and 109.7 TFLOPS (1:1) for NVIDIA.

Power figures also differ. The RX 580 has a 185 W TDP with a single 8-pin connector and a suggested 450 W PSU. The RTX PRO 6000 has a 300 W TDP with a single 16-pin connector and a suggested 700 W PSU.

Interfaces and outputs differ. The RX 580 uses PCIe 3.0 x16 and outputs 1x HDMI 2.0b plus 3x DisplayPort 1.4a. The RTX PRO 6000 uses PCIe 5.0 x16 and outputs 4x DisplayPort 2.1b.

Physical dimensions differ. The RX 580 is 241 mm long, 9.5 inches. The RTX PRO 6000 is 267 mm long, 10.5 inches, 111 mm tall, 4.4 inches, and 40 mm wide, 1.6 inches. Both are dual-slot.

Architecture Differences

The RX 580 is built on GCN 4.0, the Polaris 20 chip, using a 14 nm process from GlobalFoundries. It belongs to the Polaris RX 500 generation. The RTX PRO 6000 is built on Blackwell 2.0, the GB202 chip, using a 5 nm process from TSMC. It belongs to the Blackwell PRO W (x000) generation.

The RX 580 has no ray tracing cores and no tensor cores. The RTX PRO 6000 has 188 RT cores and 752 tensor cores. This is a fundamental architectural split. The NVIDIA card is designed for hardware-accelerated ray tracing and AI-accelerated compute, the AMD card has neither.

Cache and memory architecture also differ. The RX 580 uses GDDR5 with a 256 bit bus. The RTX PRO 6000 uses GDDR7 with a 512 bit bus. The NVIDIA card also supports a much newer PCIe interface, 5.0 versus 3.0.

API support differs. The RX 580 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.3. The RTX PRO 6000 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The DirectX 12_2 feature level enables ray tracing and mesh shaders, which the RX 580 cannot do.

Release status differs. The RX 580 launched on April 17, 2017, with a launch MSRP of 229 USD. It is end-of-life. Its predecessor is Arctic Islands, its successor is Vega. The RTX PRO 6000 launched on March 17, 2025, with a launch MSRP of 8,565 USD. It is active production. Its predecessor is Workstation Ada, and it has no successor listed.

Where Each One Wins

The RTX PRO 6000 wins every shared modern workload. In 3DMark Steel Nomad DirectX 12, it beats the RX 580 by 90.9%. It wins on memory capacity, 96 GB versus 8 GB. It wins on bandwidth, 1.79 TB/s versus 256.0 GB/s. It wins on FP32, 109.7 TFLOPS versus 6.175 TFLOPS. It wins on pixel rate, texture rate, and ROP count. It has RT and tensor cores, the RX 580 has none. It supports PCIe 5.0, DirectX 12 Ultimate, and Vulkan 1.4.

The RX 580 wins in legacy API benchmarks, but only in its own test suite. Its PassMark DirectX 9 score of 124 is its strongest API result. Its DirectX 11 score of 60 and DirectX 10 score of 46 are also respectable for a card of its era. Its Geekbench Metal score of 45235 and Vulkan score of 45173 show it remains viable in older or lighter workloads. The RTX PRO 6000 has no recorded scores in any of those tests, so no direct comparison exists.

The RX 580 also wins on power draw. Its 185 W TDP is lower than the RTX PRO 6000's 300 W. Its suggested PSU of 450 W is much lower than 700 W. Its single 8-pin connector is simpler than the 16-pin connector on the NVIDIA card. For a builder with an older system, that matters.

The RTX PRO 6000 is the pick for anyone doing ray-traced rendering, AI inference, or working with very large datasets that need 96 GB of VRAM. The RX 580 is the pick for legacy systems, older DirectX 9 and DirectX 11 games, or anyone who needs a low-power card with a simple power connector. The data does not support any other conclusion. In the one test both cards share, the NVIDIA card is 90.9% faster. That is not a close race. It is a generational gap made visible in a single number.