AMD Radeon RX 6800 XT vs NVIDIA CMP 90HX Comparison

Where Each One Wins

The recorded data splits these two cards into very different roles. The AMD Radeon RX 6800 XT wins the only direct head-to-head benchmark in the database, and it does so by a massive margin. In the Geekbench OpenCL test, the RX 6800 XT scores 171,304 against the NVIDIA CMP 90HX's 69,000, a delta of 59.7% in AMD's favor. That is not a narrow victory; it is a category-level gap.

The NVIDIA CMP 90HX has zero wins in the head-to-head comparison. Its single benchmark entry, a Geekbench OpenCL score of 69,000, places it at the 90th percentile of all GPUs in the database. That percentile is actually higher than the RX 6800 XT's 85th percentile, which is a useful nuance: the CMP 90HX is a strong card overall, but its measured workload in this database is narrow. The RX 6800 XT, by contrast, has a much wider benchmark footprint, covering DirectX 9 through DirectX 12, Metal, Vulkan, OpenCL, and multiple Passmark tests.

For general compute workloads, the RX 6800 XT is the clear winner. For the specific mining-oriented use case the CMP 90HX was built around, the database does not contain a direct mining benchmark, so no conclusion can be drawn there. What the data does show is that in the one test where both cards appear, AMD wins outright. If the workload is OpenCL compute, the choice is obvious.

FAQ

Q: Which card has the higher average benchmark score?

A: The AMD Radeon RX 6800 XT has an average benchmark score of 48,477, while the NVIDIA CMP 90HX has an average of 69,000. That said, the averages are not directly comparable because the CMP 90HX has only one benchmark entry, while the RX 6800 XT has eleven entries across different test suites.

Q: Does the NVIDIA CMP 90HX outperform the RX 6800 XT in any recorded test?

A: No. In the only head-to-head benchmark, Geekbench OpenCL, the RX 6800 XT wins with a delta of 59.7%. The CMP 90HX records zero wins in the head-to-head comparison.

Q: How does the CMP 90HX compare to its nearest rivals?

A: The CMP 90HX sits slightly ahead of the Intel Arc A770 and AMD Radeon Instinct MI25. It is 0.3% ahead of the Arc A770 and 0.6% ahead of the MI25. It trails the AMD Radeon Pro WX 8200 by 1.2% and the NVIDIA Quadro P6000 by 1.4%.

Q: What is the RX 6800 XT's position relative to its nearest rivals?

A: The RX 6800 XT is 1.5% ahead of the NVIDIA RTX A1000 Mobile, and it trails the Intel Arc A550M by 2.5%, the NVIDIA GeForce RTX 5070 Ti by 3%, and the AMD Radeon RX Vega 64 by 3%.

Q: Do both cards support the same DirectX and Vulkan versions?

A: Yes. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: Which card has a higher percentile ranking across all GPUs?

A: The NVIDIA CMP 90HX ranks at the 90th percentile, while the AMD Radeon RX 6800 XT ranks at the 85th percentile. This means the CMP 90HX is in a higher overall percentile, despite losing the direct head-to-head benchmark.

Head-to-Head Benchmarks

The database contains exactly one head-to-head benchmark between these two cards, and it is decisive. In Geekbench OpenCL, the AMD Radeon RX 6800 XT scores 171,304, while the NVIDIA CMP 90HX scores 69,000. The delta is 59.7%, meaning the RX 6800 XT is nearly 60% faster in this test. That is not a small edge; it is a dominant result.

The scale of the difference is worth emphasizing. The RX 6800 XT's OpenCL score is roughly 2.5 times higher than the CMP 90HX's. In practical terms, if a workload scales linearly with OpenCL performance, the AMD card would finish in less than half the time.

The CMP 90HX does have a higher percentile versus all GPUs, 90th versus the RX 6800 XT's 85th, but that ranking is based on a single benchmark. The RX 6800 XT's 85th percentile is derived from a much broader set of tests, including 3DMark Steel Nomad DX12, Geekbench Metal, Geekbench Vulkan, and several Passmark entries. The CMP 90HX's percentile is high, but it is also narrow.

Looking at the nearest rivals provides context for both cards. The CMP 90HX's score of 69,000 is essentially tied with the Intel Arc A770 at 68,809 and the AMD Radeon Instinct MI25 at 68,562, and it is just below the AMD Radeon Pro WX 8200 at 69,870 and the NVIDIA Quadro P6000 at 69,986. The RX 6800 XT's average score of 48,477 is close to the NVIDIA RTX A1000 Mobile at 47,743, slightly above the Intel Arc A550M at 49,737, and a bit below the NVIDIA GeForce RTX 5070 Ti at 49,957 and the AMD Radeon RX Vega 64 at 50,001.

The takeaway from the head-to-head data is straightforward: in the one test where both are measured, the RX 6800 XT is dramatically faster. The CMP 90HX may hold a higher percentile in the overall database, but that does not translate into a win in this matchup.

Specification Differences

The two cards differ in nearly every major specification category. Memory is a clear split: the CMP 90HX has 10 GB of GDDR6X on a 320-bit bus, while the RX 6800 XT has 16 GB of GDDR6 on a 256-bit bus. Bandwidth favors NVIDIA at 760.3 GB/s versus AMD's 512.0 GB/s, but the RX 6800 XT's larger frame buffer capacity is better suited to high-resolution textures.

The compute configuration also diverges. The CMP 90HX has 6,400 shading units, 200 TMUs, and 80 ROPs. The RX 6800 XT has fewer shading units at 4,608, but more TMUs at 288 and more ROPs at 128. The CMP 90HX has 50 ray tracing cores and 200 tensor cores, while the RX 6800 XT has 72 ray tracing cores and no tensor cores.

Clock speeds favor AMD. The RX 6800 XT has a base clock of 1825 MHz and a boost clock of 2250 MHz, with a game clock of 2015 MHz. The CMP 90HX has a base clock of 1500 MHz and a boost clock of 1710 MHz. Memory clocks are also different: the CMP 90HX runs at 1188 MHz with 19 Gbps effective, while the RX 6800 XT runs at 2000 MHz with 16 Gbps effective.

Pixel and texture rates are higher on the RX 6800 XT, at 288.0 GPixel/s and 648.0 GTexel/s, compared to 136.8 GPixel/s and 342.0 GTexel/s on the CMP 90HX. FP32 throughput is close, with the CMP 90HX at 21.89 TFLOPS and the RX 6800 XT at 20.74 TFLOPS. FP16 differs significantly: the CMP 90HX delivers 21.89 TFLOPS at a 1:1 ratio, while the RX 6800 XT delivers 41.47 TFLOPS at a 2:1 ratio.

Power draw is similar: 320 W for the CMP 90HX and 300 W for the RX 6800 XT, both with dual-slot coolers and dual 8-pin connectors. Both suggest a 700 W power supply. Physical dimensions differ, with the CMP 90HX at 285 mm long and 112 mm tall, and the RX 6800 XT at 267 mm long, 120 mm tall, and 50 mm wide.

The bus interface is a major difference. The CMP 90HX uses PCIe 1.0 x4, a legacy interface, while the RX 6800 XT uses PCIe 4.0 x16. Display outputs also differ: the CMP 90HX has no outputs, while the RX 6800 XT has 1x HDMI 2.1, 2x DisplayPort 1.4a, and 1x USB Type-C.

Architecture Differences

The CMP 90HX is built on NVIDIA's Ampere architecture with the GA102 chip, fabricated on an 8 nm process at Samsung. The RX 6800 XT is built on AMD's RDNA 2.0 architecture with the Navi 21 chip, fabricated on a 7 nm process at TSMC. The process node difference is significant: 8 nm versus 7 nm, which contributes to the RX 6800 XT's higher transistor density of 51.5M per mm², versus 45.1M per mm² for the CMP 90HX.

Transistor counts are close. The CMP 90HX has 28,300 million transistors on a 628 mm² die, while the RX 6800 XT has 26,800 million on a 520 mm² die. The smaller die with fewer transistors but a denser process means AMD packs more compute into less silicon area.

The CMP 90HX is a mining-specific product. It has no display outputs, which is consistent with its purpose. The RX 6800 XT is a full consumer GPU with video outputs and a launch date of October 2020, while the CMP 90HX launched in July 2021. The RX 6800 XT has a predecessor (Navi) and a successor (Navi III); the CMP 90HX has neither in the database.

Ray tracing and tensor cores differentiate the two. The CMP 90HX has 50 RT cores and 200 tensor cores, reflecting NVIDIA's architecture focus on hybrid rendering and AI acceleration. The RX 6800 XT has 72 RT cores and no tensor cores, relying on AMD's RDNA 2 design for ray tracing and compute. The RX 6800 XT's FP16 throughput is double its FP32, while the CMP 90HX runs FP16 at a 1:1 ratio with FP32.

The CMP 90HX's PCIe 1.0 x4 interface is a notable limitation for a card of this performance class. The RX 6800 XT's PCIe 4.0 x16 interface provides far more bandwidth for host communication, though mining workloads may not be heavily dependent on that interface.

The Verdict

The data points to a clear conclusion for most users: the AMD Radeon RX 6800 XT is the better card for general and compute-oriented workloads. It wins the only head-to-head benchmark by 59.7%, has a broader benchmark suite covering DirectX 9 through 12, Vulkan, Metal, and OpenCL, and offers 16 GB of memory versus 10 GB. Its higher pixel rate, texture rate, and FP16 performance make it more versatile across different types of workloads.

The NVIDIA CMP 90HX is not without merit. It holds a higher percentile ranking across all GPUs at 90th versus 85th, and its FP32 performance is marginally higher at 21.89 TFLOPS versus 20.74 TFLOPS. It also has higher memory bandwidth at 760.3 GB/s, which could be relevant for specific bandwidth-bound tasks. Its nearest rivals are all within 1.5% of its score, indicating it is a solid performer in its measured test.

However, the CMP 90HX is a mining card with no display outputs, a legacy PCIe 1.0 x4 interface, and a single benchmark entry in the database. The RX 6800 XT is a fully featured consumer GPU with modern connectivity, a 7 nm process, and a much wider software ecosystem. For anyone building a system that needs to output video, game, or run a variety of compute tasks, the RX 6800 XT is the practical choice.

For a niche scenario where raw OpenCL performance is the only metric, the RX 6800 XT still wins decisively. The CMP 90HX's higher percentile does not compensate for a 59.7% deficit in the direct comparison. The verdict is straightforward: pick the RX 6800 XT unless the specific mining use case justifies the CMP 90HX's design, and the database provides no evidence that it does.