GPU Comparison

The NVIDIA RTX 4000 SFF Ada Generation and the NVIDIA CMP 40HX occupy opposite ends of the GPU spectrum: one is a modern, power-efficient workstation accelerator, the other a legacy mining part. Benchmark data shows a decisive performance gap, but the architectural and specification differences explain why these cards are not direct substitutes. The RTX 4000 SFF Ada Generation delivers significantly higher compute scores while consuming a fraction of the power, whereas the CMP 40HX offers greater memory bandwidth and raw pixel throughput despite its older design. Below is a detailed comparison based strictly on the data provided.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA RTX 4000 SFF Ada Generation has an average benchmark score of 117,088, compared to the CMP 40HX’s 85,637. The RTX 4000 SFF Ada Generation also ranks in the 95th percentile of all GPUs, while the CMP 40HX sits in the 93rd percentile.

Q: How large is the performance gap in the head-to-head tests?

A: In Geekbench OpenCL, the RTX 4000 SFF Ada Generation scores 124,812 versus 93,395 for the CMP 40HX, a 33.6% advantage. In Geekbench Vulkan, the margin widens to 40.4%, with scores of 109,364 and 77,879, respectively.

Q: What are the memory specifications for each card?

A: The RTX 4000 SFF Ada Generation has 20 GB of GDDR6 memory on a 160-bit bus, delivering 280.0 GB/s bandwidth. The CMP 40HX has 8 GB of GDDR6 on a 256-bit bus, delivering 448.0 GB/s bandwidth. Both use 1750 MHz memory clocks with 14 Gbps effective speed.

Q: Which card is more power-efficient?

A: The RTX 4000 SFF Ada Generation has a TDP of 70 W and requires a 250 W suggested PSU, whereas the CMP 40HX has a TDP of 185 W and requires a 450 W suggested PSU. The RTX 4000 SFF Ada Generation also uses no power connectors, while the CMP 40HX needs one 8-pin connector.

Q: Are there differences in display output capabilities?

A: Yes. The RTX 4000 SFF Ada Generation provides four mini-DisplayPort 1.4a outputs. The CMP 40HX has no display outputs at all, reflecting its mining-specific design.

Q: What is the production status of each card?

A: The RTX 4000 SFF Ada Generation is listed as "Active" in production, with a release date of 2023-03-20. The CMP 40HX is marked "End-of-life," with a release date of 2021-02-24.

Architecture Differences

The two GPUs are built on fundamentally different architectures and process nodes. The RTX 4000 SFF Ada Generation uses the AD104 chip based on Ada Lovelace architecture, fabricated on a 5 nm process at TSMC. It contains 35,800 million transistors on a 294 mm² die, yielding a transistor density of 121.8M per mm². In contrast, the CMP 40HX relies on the TU106 chip with Turing architecture, built on a 12 nm process at TSMC. This older design packs 10,800 million transistors onto a much larger 445 mm² die, resulting in a transistor density of just 24.3M per mm².

These architectural differences translate directly into compute resources. The RTX 4000 SFF Ada Generation features 6,144 shading units, 192 TMUs, 48 RT cores, and 192 tensor cores. The CMP 40HX has 2,304 shading units, 144 TMUs, 36 RT cores, and 288 tensor cores. Notably, the CMP 40HX has more tensor cores than the RTX 4000 SFF Ada Generation, but this does not compensate for its far lower shading unit count. Both GPUs support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, so API-level feature parity exists despite the generational gap.

Clock speeds also differ significantly. The CMP 40HX runs at a 1470 MHz base and 1650 MHz boost, while the RTX 4000 SFF Ada Generation operates at a 720 MHz base and 1560 MHz boost. Despite the lower base clock, the Ada Lovelace chip achieves higher FP32 throughput: 19.17 TFLOPS versus 7.603 TFLOPS for the Turing part. The RTX 4000 SFF Ada Generation also delivers FP16 at a 1:1 ratio (19.17 TFLOPS), while the CMP 40HX offers FP16 at 2:1 (15.21 TFLOPS), meaning the Ada card is faster in both precision modes.

The physical design further separates these products. The RTX 4000 SFF Ada Generation measures 168 mm in length and 69 mm in height, fitting in a small form factor. The CMP 40HX is larger at 229 mm by 111 mm by 35 mm. The RTX 4000 SFF Ada Generation uses a PCIe 4.0 x16 interface, while the CMP 40HX is limited to PCIe 1.0 x4, a severe bottleneck for data transfer. The CMP 40HX also lacks any display outputs, reinforcing its purpose as a compute-only mining card.

Head-to-Head Benchmarks

The benchmark data shows a clear and consistent winner: the RTX 4000 SFF Ada Generation outperforms the CMP 40HX in both available tests. In Geekbench OpenCL, the RTX 4000 SFF Ada Generation scores 124,812 against 93,395 for the CMP 40HX, a 33.6% advantage. This is a substantial margin, reflecting the Ada card’s superior shading unit count and modern architecture. The CMP 40HX’s higher boost clock (1650 MHz vs 1560 MHz) and larger memory bus cannot offset the massive difference in compute resources.

The Vulkan test shows an even wider gap. The RTX 4000 SFF Ada Generation scores 109,364, while the CMP 40HX manages 77,879, a 40.4% difference. This suggests that the Ada Lovelace architecture handles modern graphics APIs more efficiently, likely due to its newer design and better driver optimization for compute workloads. The CMP 40HX’s PCIe 1.0 x4 interface may also hamper its performance in certain workloads, though the benchmark scores themselves are primarily compute-bound.

Looking at the nearest rivals provides additional context. The RTX 4000 SFF Ada Generation’s average score of 117,088 places it just 0.3% below the NVIDIA GB10 (117,393) and 1.6% below the AMD Radeon PRO W7700 (118,976). It sits 2.4% above the NVIDIA Tesla V100 SXM2 16 GB (114,395) and 2.8% above the RTX A5500 Mobile (113,944). This positions the RTX 4000 SFF Ada Generation as a strong mid-range workstation card, competitive with much larger and more power-hungry parts.

The CMP 40HX’s average score of 85,637 puts it 1.7% below the AMD Radeon PRO W7600 (87,108) and 2.1% below the NVIDIA Quadro GP100 (87,445). It is 4.4% above the AMD Radeon PRO W6600 (81,995) and 5.8% above the AMD Radeon Pro Vega 64X (80,959). While the CMP 40HX is not a weak performer in absolute terms, it trails the RTX 4000 SFF Ada Generation by roughly 27% in average score, a gap that no amount of memory bandwidth can bridge.

In both head-to-head tests, the RTX 4000 SFF Ada Generation wins decisively, securing 2 wins out of 2 possible. The CMP 40HX fails to claim a single victory, and its deltas of 33.6% and 40.4% indicate that it is simply outclassed in compute-heavy scenarios. The data does not support any scenario where the CMP 40HX would be the preferred choice for general-purpose or graphics workloads.

The Verdict

Based strictly on the benchmark data, the NVIDIA RTX 4000 SFF Ada Generation is the superior GPU for virtually any compute or graphics task. Its average score of 117,088 versus 85,637 for the CMP 40HX represents a 36.8% overall advantage, and it wins both head-to-head tests by margins of 33.6% and 40.4%. The RTX 4000 SFF Ada Generation achieves this while consuming only 70 W of power, compared to 185 W for the CMP 40HX, making it dramatically more efficient. It also offers 20 GB of memory versus 8 GB, double the shading units (6,144 vs 2,304), and a modern PCIe 4.0 x16 interface instead of the CMP 40HX’s legacy PCIe 1.0 x4.

The only area where the CMP 40HX holds an advantage is memory bandwidth (448.0 GB/s vs 280.0 GB/s) and pixel rate (105.6 GPixel/s vs 99.84 GPixel/s). However, these metrics are insufficient to overcome the RTX 4000 SFF Ada Generation’s massive lead in FP32 throughput (19.17 TFLOPS vs 7.603 TFLOPS) and texture rate (299.5 GTexel/s vs 237.6 GTexel/s). For users who need display outputs, the RTX 4000 SFF Ada Generation is the only viable option, as the CMP 40HX has none. The RTX 4000 SFF Ada Generation is also an active product, while the CMP 40HX is end-of-life.

Who should pick which? The RTX 4000 SFF Ada Generation is the clear choice for workstation users, small-form-factor builds, or anyone requiring a modern, power-efficient GPU with robust compute performance. The CMP 40HX might appeal to those with legacy workloads that prioritize raw memory bandwidth over compute, or who have specific mining-oriented needs that match its 8 GB memory and 448.0 GB/s bandwidth. But for nearly all practical purposes, the data points unambiguously to the RTX 4000 SFF Ada Generation as the better investment in performance and longevity.

Specification Differences

The following table highlights only the fields where the two GPUs differ:

| Specification | RTX 4000 SFF Ada Generation | CMP 40HX |

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

| Chip | AD104 | TU106 |

| Architecture | Ada Lovelace | Turing |

| Generation | Workstation Ada | Mining GPUs |

| Process Node | 5 nm | 12 nm |

| Transistors | 35,800 million | 10,800 million |

| Die Size | 294 mm² | 445 mm² |

| Transistor Density | 121.8M / mm² | 24.3M / mm² |

| Base Clock | 720 MHz | 1470 MHz |

| Boost Clock | 1560 MHz | 1650 MHz |

| Memory Size | 20 GB | 8 GB |

| Memory Bus Width | 160 bit | 256 bit |

| Memory Bandwidth | 280.0 GB/s | 448.0 GB/s |

| Shading Units | 6144 | 2304 |

| TMUs | 192 | 144 |

| RT Cores | 48 | 36 |

| Tensor Cores | 192 | 288 |

| Pixel Rate | 99.84 GPixel/s | 105.6 GPixel/s |

| Texture Rate | 299.5 GTexel/s | 237.6 GTexel/s |

| FP32 | 19.17 TFLOPS | 7.603 TFLOPS |

| FP16 | 19.17 TFLOPS (1:1) | 15.21 TFLOPS (2:1) |

| TDP | 70 W | 185 W |

| Power Connectors | None | 1x 8-pin |

| Suggested PSU | 250 W | 450 W |

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

| Display Outputs | 4x mini-DisplayPort 1.4a | No outputs |

| Length | 168 mm (6.6 inches) | 229 mm (9 inches) |

| Height | 69 mm (2.7 inches) | 111 mm (4.4 inches) |

| Width | Not specified | 35 mm (1.4 inches) |

| Production Status | Active | End-of-life |

| Release Date | 2023-03-20 | 2021-02-24 |

| Launch MSRP | Not specified | 699 USD |

| Avg Benchmark Score | 117,088 | 85,637 |

| Percentile vs All GPUs | 95 | 93 |