AMD Radeon RX 590 GME vs NVIDIA CMP 70HX Comparison

The AMD Radeon RX 590 GME and NVIDIA CMP 70HX represent two divergent approaches to GPU design, separated by architecture, process node, and intended purpose. The benchmark data reveals a clear performance hierarchy, but the underlying specification differences tell a more nuanced story about what each card is built to accomplish. The RX 590 GME, a 2020 Polaris refresh, dominates the available benchmark comparisons, while the CMP 70HX, an Ampere-based mining card, lags significantly despite possessing a more modern and complex silicon. This analysis examines the data to determine where each card excels, what the architecture differences imply, and who should consider each product based strictly on the numbers.

Where Each One Wins

The benchmark results are decisively one-sided. The AMD Radeon RX 590 GME wins both head-to-head comparisons, taking 2 wins against 0 for the NVIDIA CMP 70HX. In the Geekbench OpenCL test, the RX 590 GME scores 36,294, while the CMP 70HX manages only 25,135. This is a 44.4% advantage for AMD. The Vulkan benchmark shows an even larger gap: the RX 590 GME posts 60,508 against the CMP 70HX's 35,817, a 68.9% lead. For any user looking at raw compute performance in these specific APIs, the choice is unambiguous.

However, the CMP 70HX is not without its own statistical standing. While it loses the head-to-head, its average benchmark score of 30,476 places it in the 75th percentile of all GPUs, only slightly below the RX 590 GME's 77th percentile. The CMP 70HX's nearest rival is the NVIDIA Tesla M60, with a deltaPct of 0%, meaning the two perform nearly identically on average. This suggests the CMP 70HX is not a weak card in absolute terms, but it is outclassed by the RX 590 GME in the specific tests where both were evaluated. The RX 590 GME's average score of 32,601 puts it just 0.2% ahead of the AMD FirePro S9300 X2 and 0.4% ahead of the AMD Radeon RX 7900 GRE, indicating it sits in a competitive mid-range tier. The CMP 70HX, by contrast, sits 0.1% ahead of the AMD Radeon RX 6700 and 1.3% ahead of the AMD Radeon RX 6800, showing it is closer to those mainstream parts than to the RX 590 GME.

Architecture Differences

The architectural gap between these two cards is vast and explains the benchmark results. The RX 590 GME uses the Polaris 20 chip based on GCN 4.0 architecture, built on a 14 nm process at GlobalFoundries. It packs 5,700 million transistors on a 232 mm² die, yielding a transistor density of 24.6 million per mm². This is a mature design from the Polaris generation, released in early 2020 as an end-of-life product. Its predecessor is Arctic Islands, and its successor is Vega, situating it in a well-trodden lineage.

The CMP 70HX, on the other hand, is built on the GA104 chip using Nvidia's Ampere architecture, fabricated on an 8 nm process at Samsung. It contains 17,400 million transistors on a 392 mm² die, resulting in a density of 44.4 million per mm². This is over three times the transistor count and nearly double the density of the RX 590 GME. The CMP 70HX is classified under "Mining GPUs" generation, a purpose-built category with no display outputs. Its architecture supports DirectX 12 Ultimate (12_2), while the RX 590 GME only reaches DirectX 12 (12_0). Both support OpenGL 4.6, but the CMP 70HX has Vulkan 1.4 support versus the RX 590 GME's Vulkan 1.3.

The memory subsystems differ dramatically as well. The RX 590 GME uses 8 GB of GDDR5 on a 256-bit bus, providing 256.0 GB/s of bandwidth. The CMP 70HX also has 8 GB, but it uses GDDR6X on a 256-bit bus, delivering 608.3 GB/s—more than double the bandwidth. This is a critical distinction, as higher bandwidth typically benefits compute workloads, yet the RX 590 GME still wins in the benchmarks. The CMP 70HX also features 30 RT cores and 120 tensor cores, which the RX 590 GME lacks entirely. These are meant for ray tracing and AI acceleration, but they do not appear in the benchmark data provided.

Head-to-Head Benchmarks

The Geekbench OpenCL test shows the RX 590 GME scoring 36,294 against the CMP 70HX's 25,135. The 44.4% delta is substantial and indicates a fundamental advantage in general-purpose compute. The RX 590 GME has 2,304 shading units, 144 TMUs, and 32 ROPs, while the CMP 70HX has 3,840 shading units, 120 TMUs, and 64 ROPs. Despite having fewer shading units and half the ROPs, the RX 590 GME achieves a higher texture rate of 204.5 GTexel/s versus the CMP 70HX's 167.4 GTexel/s. The pixel rate also favors the CMP 70HX at 89.28 GPixel/s versus the RX 590 GME's 45.44 GPixel/s, but this does not translate to a benchmark win.

In the Vulkan test, the RX 590 GME scores 60,508, which is 68.9% higher than the CMP 70HX's 35,817. This is the largest margin in the data. The RX 590 GME's fp32 throughput is 6.543 TFLOPS, while the CMP 70HX reaches 10.71 TFLOPS—a 64% advantage for Nvidia in raw floating-point math. Yet the benchmark results invert this expectation. The RX 590 GME's fp16 performance is also 6.543 TFLOPS (1:1), matching its fp32, while the CMP 70HX also offers 10.71 TFLOPS fp16 (1:1). The architectural efficiency of GCN 4.0 appears to outweigh the Ampere's raw compute ceiling in these specific workloads.

The clock speeds tell a similar story. The RX 590 GME has a base clock of 1257 MHz and a boost of 1420 MHz, while the CMP 70HX has a base of 1365 MHz and a boost of 1395 MHz. The CMP 70HX's boost is barely above its base, suggesting thermal or power constraints, while the RX 590 GME has a wider boost headroom. The CMP 70HX's memory clock is 1188 MHz with 19 Gbps effective speed, compared to the RX 590 GME's 2000 MHz and 8 Gbps effective. The GDDR6X bandwidth advantage is real, but it doesn't help in these tests.

The Verdict

The data leads to a clear conclusion: the AMD Radeon RX 590 GME is the superior performer in every benchmark where both were tested. It wins the OpenCL test by 44.4% and the Vulkan test by 68.9%, making it the choice for any workload relying on these APIs. Its average benchmark score of 32,601 is 7% higher than the CMP 70HX's 30,476, and it sits in the 77th percentile versus the CMP 70HX's 75th. For users seeking compute performance in OpenCL or Vulkan, the RX 590 GME is unequivocally the better option.

The NVIDIA CMP 70HX, however, should not be dismissed outright. Its architecture is far more advanced, with 17,400 million transistors, 30 RT cores, and 120 tensor cores, making it a more capable part for future-proofing in specific scenarios. Its 608.3 GB/s memory bandwidth is more than double the RX 590 GME's, and its fp32 throughput of 10.71 TFLOPS is significantly higher. The CMP 70HX also supports DirectX 12 Ultimate and Vulkan 1.4, while the RX 590 GME is limited to DirectX 12 (12_0) and Vulkan 1.3. For workloads that leverage these features—such as ray tracing or tensor-based AI—the CMP 70HX could theoretically excel, but the provided benchmarks do not capture that. Given the data, the RX 590 GME is the pick for general compute, while the CMP 70HX is a specialized tool with a different intended use case.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon RX 590 GME has an average benchmark score of 32,601, which is higher than the NVIDIA CMP 70HX's 30,476.

Q: How much faster is the RX 590 GME in the Vulkan benchmark?

A: The RX 590 GME scores 60,508 in Geekbench Vulkan, which is 68.9% higher than the CMP 70HX's 35,817.

Q: What are the memory bandwidth figures for each card?

A: The RX 590 GME has a bandwidth of 256.0 GB/s using GDDR5, while the CMP 70HX has 608.3 GB/s using GDDR6X.

Q: Does the CMP 70HX have any features the RX 590 GME lacks?

A: Yes, the CMP 70HX includes 30 RT cores and 120 tensor cores, which the RX 590 GME does not have, and it supports DirectX 12 Ultimate and Vulkan 1.4.

Q: What is the transistor count difference between the two?

A: The RX 590 GME has 5,700 million transistors, while the CMP 70HX has 17,400 million transistors.

Q: Which card has a higher boost clock?

A: The RX 590 GME has a boost clock of 1420 MHz, which is higher than the CMP 70HX's 1395 MHz boost clock.

Specification Differences

| Specification | AMD Radeon RX 590 GME | NVIDIA CMP 70HX |

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

| Architecture | GCN 4.0 | Ampere |

| Process Node | 14 nm | 8 nm |

| Transistors | 5,700 million | 17,400 million |

| Die Size | 232 mm² | 392 mm² |

| Base Clock | 1257 MHz | 1365 MHz |

| Boost Clock | 1420 MHz | 1395 MHz |

| Memory Type | GDDR5 | GDDR6X |

| Memory Bandwidth | 256.0 GB/s | 608.3 GB/s |

| Shading Units | 2304 | 3840 |

| TMUs | 144 | 120 |

| ROPs | 32 | 64 |

| RT Cores | None | 30 |

| Tensor Cores | None | 120 |

| Pixel Rate | 45.44 GPixel/s | 89.28 GPixel/s |

| Texture Rate | 204.5 GTexel/s | 167.4 GTexel/s |

| FP32 Performance | 6.543 TFLOPS | 10.71 TFLOPS |

| Power Connectors | 1x 8-pin | 1x 12-pin |

| Suggested PSU | 450 W | 200 W |

| Bus Interface | PCIe 3.0 x16 | PCIe 1.0 x4 |

| Display Outputs | 1x HDMI 2.0b, 3x DisplayPort 1.4a | No outputs |

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

| Vulkan Support | 1.3 | 1.4 |

| Length | 241 mm (9.5 inches) | 267 mm (10.5 inches) |

| Height | Not specified | 112 mm (4.4 inches) |