NVIDIA CMP 90HX vs NVIDIA RTX 5880 Ada Generation Comparison

Head-to-Head Benchmarks

The recorded data contains a single direct comparison between the NVIDIA CMP 90HX and the NVIDIA RTX 5880 Ada Generation: the Geekbench OpenCL benchmark. This test measures general-purpose compute performance across a wide range of workloads, and the result is decisive. The RTX 5880 Ada Generation scores 326,898 points, while the CMP 90HX scores 69,000 points. The delta is -78.9% from the perspective of the CMP 90HX, meaning the Ada card delivers approximately 4.7 times the raw OpenCL compute throughput of the Ampere mining card. In absolute terms, the RTX 5880 leads by 257,898 points, a margin that dwarfs the entire score of the CMP 90HX.

This is not a close contest. The CMP 90HX sits at the 90th percentile among all GPUs in the database, with an average benchmark score of 69,000. Its nearest rivals in that metric include the Intel Arc A770, which scores 68,809 (a 0.3% difference), and the AMD Radeon Instinct MI25 at 68,562 (0.6% higher). Meanwhile, the RTX 5880 Ada Generation holds an 85th percentile rank, with an average score of 45,972 across its full benchmark suite, a figure depressed by several low DirectX and 2D results. Its nearest rivals are the NVIDIA RTX A2000 at 46,043 (-0.2% relative to the 5880) and the Intel Arc A730M at 45,592 (0.8% higher). The percentile rankings do not tell the full story: the CMP 90HX has a higher percentile despite a far lower raw OpenCL score, because the Ada card's average is pulled down by its Passmark results, which include a DirectX 12 score of 70 and a G2D score of 777. Nevertheless, in the one benchmark where both cards were measured under identical conditions, the RTX 5880 utterly outperforms the CMP 90HX.

Where Each One Wins

The CMP 90HX has no wins in the head-to-head benchmark data. Every recorded comparison favors the RTX 5880 Ada Generation. The CMP 90HX does hold a higher percentile ranking against the broader GPU population (90th versus 85th), which suggests that its single benchmark result places it in a strong position relative to the full database, but that is not a direct victory over the Ada card.

The RTX 5880 Ada Generation wins the only direct test, and its advantage is not marginal. The Geekbench OpenCL score of 326,898 against 69,000 is a 373.8% improvement for the Ada card, a gap that indicates a fundamentally different compute capability. Beyond that single comparison, the Ada card's additional benchmark results paint a more nuanced picture of its strengths. Its Passmark G3D score of 25,096 and GPU compute score of 14,208 are solid, while its legacy DirectX 9 score of 335 and DirectX 10 score of 167 show that its dominance is concentrated in modern workloads, not in older API paths. The CMP 90HX, by design, has no display outputs and no gaming-oriented benchmark data, so it cannot be evaluated in DirectX or 2D contexts at all. For compute-centric tasks measured by OpenCL, the RTX 5880 is the clear winner. For any workload that requires graphics output, the CMP 90HX cannot participate, making the Ada card the only functional choice regardless of raw performance.

Architecture Differences

The two GPUs come from different architectural generations entirely. The CMP 90HX is built on the GA102 chip using the Ampere architecture, manufactured on an 8 nm process at Samsung. The RTX 5880 Ada Generation uses the AD102 chip with the Ada Lovelace architecture, built on a 5 nm process at TSMC. The transistor counts reflect this generational leap: the Ampere chip contains 28,300 million transistors on a 628 mm² die, while the Ada chip packs 76,300 million transistors into a slightly smaller 609 mm² die. The transistor density tells the same story, with the Ada chip reaching 125.3 million transistors per square millimeter versus 45.1 million for the Ampere chip. This is a manufacturing and design advantage that enables the Ada card's higher throughput despite its lower power draw.

The compute resources differ substantially. The CMP 90HX has 6,400 shading units, 200 texture mapping units, and 80 ROPs. The RTX 5880 Ada Generation more than doubles those figures with 14,080 shading units, 440 TMUs, and 176 ROPs. Ray tracing cores follow the same pattern: 50 on the Ampere card versus 110 on the Ada card. Tensor cores scale from 200 to 440. The resulting peak rates are starkly different. The Ada card reaches 69.27 TFLOPS of FP32 and FP16 performance, while the CMP 90HX manages 21.89 TFLOPS in both. Pixel fill rate jumps from 136.8 GPixel/s to 433.0 GPixel/s, and texture fill rate rises from 342.0 GTexel/s to 1,082.4 GTexel/s.

Memory configurations also diverge sharply. The CMP 90HX uses 10 GB of GDDR6X on a 320-bit bus, yielding 760.3 GB/s of bandwidth. The RTX 5880 Ada Generation uses 48 GB of GDDR6 on a 384-bit bus, providing 864.0 GB/s. The Ada card has nearly five times the memory capacity and 13.6% more bandwidth. Clock behavior differs too: the Ampere card runs at a 1500 MHz base and 1710 MHz boost, while the Ada card has a much lower 975 MHz base but a far higher 2460 MHz boost. The memory clocks are 1188 MHz (19 Gbps effective) for the CMP 90HX versus 2250 MHz (18 Gbps effective) for the RTX 5880. Despite the Ada card's greater compute resources, its TDP is lower at 285 W versus 320 W, and its suggested PSU is 600 W versus 700 W. The power connector requirements also differ: the CMP 90HX needs two 8-pin connectors, while the RTX 5880 uses a single 16-pin connector.

The two cards serve different roles at the interface and output level. The CMP 90HX has no display outputs, which is consistent with its mining-focused design, and it uses a PCIe 1.0 x4 bus interface, which severely limits host communication. The RTX 5880 Ada Generation provides four DisplayPort 1.4a outputs and uses a full PCIe 4.0 x16 interface. The CMP 90HX is 285 mm long and 112 mm tall, while the RTX 5880 is slightly shorter at 267 mm with the same 112 mm height. Both are dual-slot cards. The CMP 90HX is end-of-life, released in July 2021, while the RTX 5880 is active, released in January 2024, with a predecessor in the Workstation Ampere line and a successor in Blackwell PRO W.

FAQ

Q: Which GPU has the higher Geekbench OpenCL score?

A: The NVIDIA RTX 5880 Ada Generation, with a score of 326,898 compared to 69,000 for the NVIDIA CMP 90HX, a difference of -78.9% relative to the Ada card.

Q: How much memory does each card have?

A: The CMP 90HX has 10 GB of GDDR6X on a 320-bit bus, while the RTX 5880 Ada Generation has 48 GB of GDDR6 on a 384-bit bus.

Q: What is the TDP of each card?

A: The CMP 90HX has a TDP of 320 W with a suggested PSU of 700 W, while the RTX 5880 Ada Generation has a TDP of 285 W with a suggested PSU of 600 W.

Q: Do both cards support the same APIs?

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

Q: Can the CMP 90HX be used for display output?

A: No, the CMP 90HX has no display outputs. The RTX 5880 Ada Generation has 4x DisplayPort 1.4a outputs.

Q: Which card has a higher production status?

A: The CMP 90HX is end-of-life, while the RTX 5880 Ada Generation is active.

Specification Differences

The two cards differ across nearly every major specification category. The process node is 8 nm for the CMP 90HX (Samsung) versus 5 nm for the RTX 5880 Ada Generation (TSMC). Transistor count is 28,300 million versus 76,300 million, and die size is 628 mm² versus 609 mm². Transistor density is 45.1 million per mm² versus 125.3 million per mm².

Clock speeds differ significantly: the CMP 90HX has a 1500 MHz base and 1710 MHz boost, while the RTX 5880 has a 975 MHz base and 2460 MHz boost. Memory clocks are 1188 MHz (19 Gbps effective) versus 2250 MHz (18 Gbps effective). Memory capacity is 10 GB versus 48 GB, with memory type GDDR6X versus GDDR6, bus width 320-bit versus 384-bit, and bandwidth 760.3 GB/s versus 864.0 GB/s.

Compute units scale accordingly: shading units are 6,400 versus 14,080, TMUs are 200 versus 440, ROPs are 80 versus 176, RT cores are 50 versus 110, and tensor cores are 200 versus 440. Pixel rate is 136.8 GPixel/s versus 433.0 GPixel/s, texture rate is 342.0 GTexel/s versus 1,082.4 GTexel/s, and FP32/FP16 performance is 21.89 TFLOPS versus 69.27 TFLOPS.

Power and connectivity also differ. TDP is 320 W versus 285 W, power connectors are 2x 8-pin versus 1x 16-pin, and suggested PSU is 700 W versus 600 W. The bus interface is PCIe 1.0 x4 on the CMP 90HX versus PCIe 4.0 x16 on the RTX 5880. Display outputs are absent on the CMP 90HX, while the RTX 5880 has 4x DisplayPort 1.4a. Physical dimensions are 285 mm by 112 mm for the CMP 90HX versus 267 mm by 112 mm for the RTX 5880, with both being dual-slot. Release dates are July 2021 for the CMP 90HX and January 2024 for the RTX 5880. Production status is end-of-life versus active, and the RTX 5880 lists a predecessor in Workstation Ampere and a successor in Blackwell PRO W.