NVIDIA CMP 90HX vs NVIDIA GeForce RTX 5090 D Comparison

The NVIDIA GeForce RTX 5090 D and the NVIDIA CMP 90HX occupy opposite ends of NVIDIA’s hardware spectrum. One is a flagship Blackwell consumer GPU aimed at gaming and creation; the other is a dedicated Ampere mining part with no display outputs. The data reveals an extraordinary performance gulf, but also a surprising closeness in overall percentile ranking. This analysis compares the two strictly on benchmark scores, specifications, and architectural facts.

Head-to-Head Benchmarks

The only common benchmark between the two cards is Geekbench OpenCL, and the result is a landslide. The RTX 5090 D scores 310674 points, while the CMP 90HX manages 69000 points. That translates to a deltaPct of 350.3% in favor of the RTX 5090 D. To put this in perspective, the RTX 5090 D is roughly 4.5 times faster in raw compute throughput as measured by this test. This is not a close contest; it is a generation-defining gap.

The RTX 5090 D’s average benchmark score across all its tested workloads is 77712, while the CMP 90HX’s average is exactly its single Geekbench score of 69000. The RTX 5090 D’s average is dragged down by its low Passmark DirectX scores (e.g., 231 in DirectX 10, 219 in DirectX 12), which are likely outlier results from synthetic legacy tests. Still, the Geekbench OpenCL delta is the single most telling number.

Looking at the RTX 5090 D’s other benchmarks, it achieves 14326 in 3DMark Steel Nomad DX12, 376915 in Geekbench Vulkan, and 44065 in Passmark G3D. The CMP 90HX has no such results, so cross-comparison is impossible. However, the CMP 90HX’s percentile rank of 90 versus the RTX 5090 D’s 92 suggests that despite the massive raw compute gap, the mining card still sits near the top of the overall GPU distribution — proof of how powerful even a crippled Ampere part is.

In terms of nearest rivals, the RTX 5090 D is 1.1% ahead of the AMD Radeon RX 6650M XT (avg score 76904) and 1.6% behind the AMD Radeon RX 6850M XT (78940). The CMP 90HX is 0.3% ahead of the Intel Arc A770 (68809) and 1.2% behind the AMD Radeon Pro WX 8200 (69870). These deltas show that the RTX 5090 D’s average is competitive with high-end laptop GPUs, while the CMP 90HX sits in a similar class to midrange desktop parts.

Where Each One Wins

The RTX 5090 D wins everywhere that compute and graphics matter. In the single head-to-head test, it wins outright. Its shading units (21760) dwarf the CMP 90HX’s 6400, and its texture rate of 1,636.8 GTexel/s versus 342.0 GTexel/s means it can fill geometry far faster. Pixel rate is also lopsided: 423.6 GPixel/s versus 136.8 GPixel/s. For any workload involving 3D rendering, ray tracing, or high-resolution game assets, the RTX 5090 D is the only choice.

The CMP 90HX’s sole win is in its niche: mining. It has 10 GB of GDDR6X memory on a 320 bit bus, yielding 760.3 GB/s of bandwidth. That is substantial, but the RTX 5090 D doubles it with 32 GB of GDDR7 on a 512 bit bus, reaching 1.79 TB/s. The CMP 90HX’s advantage is its power efficiency per mining hash, but that is not measurable from the provided benchmark data. Its 320 W TDP and 700 W suggested PSU are lower than the RTX 5090 D’s 575 W and 950 W, respectively, but the data does not show actual mining performance.

The RTX 5090 D also wins on API support parity. Both cards support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, but the RTX 5090 D’s Blackwell architecture includes 170 RT cores and 680 tensor cores, versus the CMP 90HX’s 50 RT cores and 200 tensor cores. For machine learning inference or ray-traced workloads, the RTX 5090 D has a clear hardware advantage.

The Verdict

Pick the NVIDIA GeForce RTX 5090 D if you need any form of graphics output, gaming, or professional compute. Its benchmark data shows a 350.3% lead over the CMP 90HX in Geekbench OpenCL, and it has a display output (1x HDMI 2.1b, 3x DisplayPort 2.1b) while the CMP 90HX has none. The RTX 5090 D is also an active product with a release date of 2025-01-29, whereas the CMP 90HX is end-of-life, released 2021-07-27. The RTX 5090 D’s launch MSRP is 2,299 USD.

Pick the NVIDIA CMP 90HX only if you have a use case that requires no display output and you are constrained by power or PCIe interface limitations. It runs on PCIe 1.0 x4, which is bizarrely ancient, but its 69000 Geekbench OpenCL score places it at the 90th percentile of all GPUs — it is not slow. However, its nearest rivals (Intel Arc A770, AMD Radeon Instinct MI25) are all modern parts with display capabilities. The CMP 90HX’s only practical advantage is its lower 320 W TDP and 700 W PSU requirement, but the data does not support any performance-per-watt claim.

For almost any reader, the RTX 5090 D is the obvious choice. The CMP 90HX is a historical artifact from the mining boom, and its single benchmark result cannot justify its existence in a modern system.

FAQ

Q: How much faster is the RTX 5090 D than the CMP 90HX in Geekbench OpenCL?

A: The RTX 5090 D scores 310674 versus the CMP 90HX’s 69000, a deltaPct of 350.3%.

Q: Do both cards have the same API support?

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

Q: What is the memory configuration difference?

A: The RTX 5090 D has 32 GB of GDDR7 on a 512 bit bus with 1.79 TB/s bandwidth, while the CMP 90HX has 10 GB of GDDR6X on a 320 bit bus with 760.3 GB/s.

Q: Which card has a higher percentile rank among all GPUs?

A: The RTX 5090 D is at the 92nd percentile, while the CMP 90HX is at the 90th percentile.

Q: Does the CMP 90HX have any display outputs?

A: No, the CMP 90HX has no display outputs, while the RTX 5090 D has 1x HDMI 2.1b and 3x DisplayPort 2.1b.

Q: What are the transistor counts and process nodes?

A: The RTX 5090 D uses 92,200 million transistors on a 5 nm TSMC process, while the CMP 90HX uses 28,300 million transistors on an 8 nm Samsung process.

Architecture Differences

The RTX 5090 D is built on the GB202 chip using Blackwell 2.0 architecture, fabricated by TSMC on a 5 nm process. It packs 92,200 million transistors onto a 750 mm² die, yielding a transistor density of 122.9M / mm². The CMP 90HX uses the GA102 chip with Ampere architecture, fabricated by Samsung on an 8 nm process. It has 28,300 million transistors on a 628 mm² die, giving a density of 45.1M / mm². This density gap is a direct result of the process node difference — the RTX 5090 D fits over three times more transistors per square millimeter.

The RTX 5090 D has 170 RT cores and 680 tensor cores, while the CMP 90HX has 50 RT cores and 200 tensor cores. The Blackwell part also has 21760 shading units, 680 TMUs, and 176 ROPs, versus the Ampere part’s 6400 shading units, 200 TMUs, and 80 ROPs. The RTX 5090 D’s FP32 throughput is 104.8 TFLOPS, and its FP16 is also 104.8 TFLOPS (1:1). The CMP 90HX offers 21.89 TFLOPS in both FP32 and FP16, again at 1:1 ratio.

Memory architecture differs fundamentally. The RTX 5090 D uses GDDR7 at 1750 MHz (28 Gbps effective) with a 512 bit bus, while the CMP 90HX uses GDDR6X at 1188 MHz (19 Gbps effective) with a 320 bit bus. This results in 1.79 TB/s versus 760.3 GB/s of bandwidth. The RTX 5090 D also supports PCIe 5.0 x16, whereas the CMP 90HX is limited to PCIe 1.0 x4 — a severe bottleneck for any data transfer.

Specification Differences

The two cards differ in nearly every measurable specification. The RTX 5090 D has a base clock of 2017 MHz and a boost clock of 2407 MHz, while the CMP 90HX runs at 1500 MHz base and 1710 MHz boost. The RTX 5090 D’s TDP is 575 W with a 950 W suggested PSU and a single 16-pin power connector; the CMP 90HX draws 320 W with a 700 W suggested PSU and two 8-pin connectors.

Memory size: 32 GB versus 10 GB. Memory type: GDDR7 versus GDDR6X. Bus width: 512 bit versus 320 bit. Bandwidth: 1.79 TB/s versus 760.3 GB/s. Pixel rate: 423.6 GPixel/s versus 136.8 GPixel/s. Texture rate: 1,636.8 GTexel/s versus 342.0 GTexel/s.

Physical dimensions also diverge. The RTX 5090 D is 304 mm long, 137 mm tall, and 48 mm wide. The CMP 90HX is 285 mm long and 112 mm tall, with no width listed. Both are dual-slot. The RTX 5090 D has display outputs; the CMP 90HX has none. The RTX 5090 D is an active product with a successor (GeForce 60), while the CMP 90HX is end-of-life with no successor. Finally, the RTX 5090 D has a launch MSRP of 2,299 USD, while the CMP 90HX has no launch MSRP listed.