AMD Radeon RX 6900 XT vs NVIDIA CMP 90HX Comparison

The NVIDIA CMP 90HX and the AMD Radeon RX 6900 XT are both end-of-life graphics cards, but they serve fundamentally different purposes. The data shows a clear separation: the CMP 90HX is a dedicated mining product with no display outputs, while the RX 6900 XT is a full-featured consumer card. The benchmark results provide a single head-to-head data point, but the architectural and specification differences paint a much larger picture of where each card excels.

Where Each One Wins

The recorded data gives the AMD Radeon RX 6900 XT a decisive win in the only shared benchmark. In Geekbench OpenCL, the RX 6900 XT scores 187,673 against the CMP 90HX's 69,000. That is a delta of -63.2% for the NVIDIA card, meaning the AMD part is roughly 172% faster in this compute test. For any general-purpose compute workload, the data indicates the RX 6900 XT is the superior choice.

The NVIDIA CMP 90HX, however, has no victories in the head-to-head benchmarks. Its advantage lies entirely outside of raw benchmark scores. The card is designed for a single purpose: mining. It has no display outputs, which means it cannot drive a monitor. The CMP 90HX wins in the niche scenario where a user needs a card dedicated solely to hashing, with no expectation of rendering an interface or gaming. Its 90th percentile ranking among all GPUs (based on its single Geekbench OpenCL score) shows it is a competent compute device, but its lack of outputs makes it useless for a desktop workstation.

The RX 6900 XT wins in every practical use case that involves interacting with the card. It has a full suite of display outputs, including 1x HDMI 2.1, 2x DisplayPort 1.4a, and 1x USB Type-C. Its benchmark suite covers DirectX 9, 10, 11, and 12, Vulkan, Metal, and OpenCL. The PassMark G3D score of 26,732 and the Geekbench Vulkan score of 148,525 confirm it is a capable gaming and rendering card. The architecture differences support this split: the CMP 90HX is a mining tool, and the RX 6900 XT is a complete graphics solution.

Architecture Differences

The two cards come from different foundries and process nodes. The NVIDIA CMP 90HX uses the GA102 chip on Samsung's 8 nm process, while the AMD Radeon RX 6900 XT uses the Navi 21 chip on TSMC's 7 nm process. The smaller process node gives AMD a density advantage: 51.5 million transistors per mm² versus 45.1 million for NVIDIA. However, the NVIDIA chip is physically larger at 628 mm² and contains 28,300 million transistors, compared to the AMD chip's 520 mm² die with 26,800 million transistors.

The compute architectures are fundamentally different. The CMP 90HX is built on Ampere, which includes dedicated tensor cores (200 of them) and RT cores (50). The RX 6900 XT is built on RDNA 2.0 and has no tensor cores but includes 80 RT cores. The shader configurations also diverge: the CMP 90HX has 6,400 shading units, 200 TMUs, and 80 ROPs, while the RX 6900 XT has 5,120 shading units, 320 TMUs, and 128 ROPs. This means the AMD card has more texture and pixel throughput potential, despite having fewer shaders. The data confirms this: the RX 6900 XT's pixel rate is 288.0 GPixel/s and texture rate is 720.0 GTexel/s, both far higher than the CMP 90HX's 136.8 GPixel/s and 342.0 GTexel/s.

Memory architecture also differs. The CMP 90HX uses 10 GB of GDDR6X on a 320-bit bus, delivering 760.3 GB/s of bandwidth. The RX 6900 XT uses 16 GB of GDDR6 on a 256-bit bus, delivering 512.0 GB/s. The NVIDIA card has more bandwidth, but the AMD card has more capacity and a wider effective feature set for consumer workloads. The CMP 90HX's PCIe 1.0 x4 interface is a clear indication of its mining focus, as it does not need high bandwidth to the host system. The RX 6900 XT uses PCIe 4.0 x16, which is standard for a modern consumer card.

FAQ

Q: Can the NVIDIA CMP 90HX be used to output video to a monitor?

A: No. The database records "No outputs" for the CMP 90HX. It is a mining-only product with no display connections. The AMD Radeon RX 6900 XT, by contrast, has 1x HDMI 2.1, 2x DisplayPort 1.4a, and 1x USB Type-C.

Q: Which card has more memory bandwidth?

A: The NVIDIA CMP 90HX has 760.3 GB/s of bandwidth, which is higher than the AMD Radeon RX 6900 XT's 512.0 GB/s. The NVIDIA card uses GDDR6X memory on a 320-bit bus, while the AMD card uses GDDR6 on a 256-bit bus.

Q: How do their compute scores compare in OpenCL?

A: The AMD Radeon RX 6900 XT scores 187,673 in Geekbench OpenCL, while the NVIDIA CMP 90HX scores 69,000. This results in a -63.2% delta for the NVIDIA card, showing a significant advantage for AMD in this specific test.

Q: What is the physical size difference between the two cards?

A: The NVIDIA CMP 90HX is 285 mm long and 112 mm tall. The AMD Radeon RX 6900 XT is 267 mm long, 120 mm tall, and 50 mm wide. The AMD card is shorter but taller and adds width, while the NVIDIA card is longer.

Q: Which card has a higher boost clock?

A: The AMD Radeon RX 6900 XT has a boost clock of 2250 MHz, which is higher than the NVIDIA CMP 90HX's boost clock of 1710 MHz. The AMD card also has a higher base clock at 1825 MHz versus 1500 MHz.

Q: Do both cards support the same API feature level?

A: Yes. Both the NVIDIA CMP 90HX and the AMD Radeon RX 6900 XT support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Specification Differences

| Specification | NVIDIA CMP 90HX | AMD Radeon RX 6900 XT |

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

| Process Node | 8 nm (Samsung) | 7 nm (TSMC) |

| Transistors | 28,300 million | 26,800 million |

| Die Size | 628 mm² | 520 mm² |

| Transistor Density | 45.1M / mm² | 51.5M / mm² |

| Base Clock | 1500 MHz | 1825 MHz |

| Boost Clock | 1710 MHz | 2250 MHz |

| Game Clock | Not specified | 2015 MHz |

| Memory Size | 10 GB | 16 GB |

| Memory Type | GDDR6X | GDDR6 |

| Memory Bus Width | 320 bit | 256 bit |

| Memory Bandwidth | 760.3 GB/s | 512.0 GB/s |

| Memory Clock | 19 Gbps effective | 16 Gbps effective |

| Shading Units | 6,400 | 5,120 |

| TMUs | 200 | 320 |

| ROPs | 80 | 128 |

| RT Cores | 50 | 80 |

| Tensor Cores | 200 | None |

| Pixel Rate | 136.8 GPixel/s | 288.0 GPixel/s |

| Texture Rate | 342.0 GTexel/s | 720.0 GTexel/s |

| FP32 Performance | 21.89 TFLOPS | 23.04 TFLOPS |

| FP16 Performance | 21.89 TFLOPS (1:1) | 46.08 TFLOPS (2:1) |

| TDP | 320 W | 300 W |

| Slot Width | Dual-slot | Triple-slot |

| Bus Interface | PCIe 1.0 x4 | PCIe 4.0 x16 |

| Display Outputs | No outputs | 1x HDMI 2.1, 2x DP 1.4a, 1x USB-C |

| Dimensions | 285 mm x 112 mm | 267 mm x 120 mm x 50 mm |

| Release Date | 2021-07-27 | 2020-10-27 |

Head-to-Head Benchmarks

The only direct benchmark comparison in the database is Geekbench OpenCL. The AMD Radeon RX 6900 XT scores 187,673, while the NVIDIA CMP 90HX scores 69,000. The delta is -63.2%, indicating a massive performance gap in favor of AMD. This is a compute-oriented test, which should theoretically favor the mining-focused CMP 90HX, yet the RX 6900 XT outperforms it by a wide margin.

The CMP 90HX's nearest rivals in the database, such as the Intel Arc A770 (68,809, +0.3%) and the AMD Radeon Instinct MI25 (68,562, +0.6%), show it sits in a narrow performance band. It is slightly above the Arc A770 and MI25, but below the AMD Radeon Pro WX 8200 (69,870, -1.2%) and the NVIDIA Quadro P6000 (69,986, -1.4%). This confirms the CMP 90HX is a mid-tier compute card, not a high-end performer.

The RX 6900 XT's nearest rivals include the NVIDIA CMP 50HX (51,790, -1.6%) and the AMD Radeon RX Vega 64 (50,001, +1.9%). This places the RX 6900 XT in a higher performance class, with an average benchmark score of 50,951 across all its tests. The CMP 90HX's average score is 69,000, but that is based on a single test, which makes a direct average comparison misleading. The RX 6900 XT's suite includes multiple DirectX and PassMark tests, showing consistent capability across different APIs.

The Verdict

The data makes the choice clear for most users. The AMD Radeon RX 6900 XT is the only viable option for any task that requires a display output, gaming, or general-purpose graphics rendering. It wins the sole head-to-head benchmark by a wide margin, has a higher boost clock, more memory capacity, and a more complete feature set including display outputs and PCIe 4.0 x16 connectivity.

The NVIDIA CMP 90HX is a specialized tool. Its lack of display outputs and its PCIe 1.0 x4 interface restrict it to mining or headless compute tasks. Its higher memory bandwidth (760.3 GB/s) and larger die size (628 mm²) are advantages in that narrow context, but its single benchmark score of 69,000 is far below the RX 6900 XT's OpenCL result. The CMP 90HX also has a higher TDP of 320 W versus 300 W, making it less power-efficient on paper.

For a consumer building a gaming PC or a workstation, the RX 6900 XT is the clear winner. Its 16 GB of memory, higher pixel rate, and superior compute scores make it a versatile card. The CMP 90HX should only be considered by someone who specifically needs a dedicated mining card and has no requirement to connect a monitor. The benchmark data does not support the CMP 90HX as a general-purpose alternative. The RX 6900 XT is the right choice for nearly every scenario.