AMD Radeon RX 7600M vs NVIDIA CMP 90HX Comparison

# NVIDIA CMP 90HX vs AMD Radeon RX 7600M

The NVIDIA CMP 90HX and AMD Radeon RX 7600M occupy starkly different corners of the GPU landscape—the former a desktop-oriented mining card with no display outputs, the latter a mobile gaming chip designed for laptops. Despite their contrasting purposes, benchmark data places them within striking distance of each other in raw compute performance. The CMP 90HX scores 69,000 in Geekbench OpenCL, edging out the RX 7600M's 63,775 by 8.2%. Yet percentile rankings tell a nuanced story: the CMP 90HX sits at the 90th percentile of all GPUs, while the RX 7600M trails only one point behind at 89th. This narrow percentile gap, despite an 8.2% score difference, suggests the field is densely packed at this performance tier.

FAQ

Q: Which GPU has the higher Geekbench OpenCL score?

A: The NVIDIA CMP 90HX leads with a score of 69,000 compared to the AMD Radeon RX 7600M's 63,775, a difference of 8.2%. The CMP 90HX won the only head-to-head benchmark test recorded.

Q: How do these GPUs compare to their nearest rivals?

A: The CMP 90HX is 0.3% ahead of the Intel Arc A770 (68,809) and 0.6% ahead of the AMD Radeon Instinct MI25 (68,562), but 1.2% behind the AMD Radeon Pro WX 8200 (69,870). The RX 7600M sits within 0.7% of three rivals—RX 9060 XT LP (63,830), CMP 30HX (63,842), and Radeon Pro WX 9100 (64,212)—and is just 0.1% ahead of the Radeon Pro Vega 56 (63,693).

Q: What are the architectural differences between the two chips?

A: The CMP 90HX uses NVIDIA's Ampere architecture (GA102 chip) built on Samsung's 8 nm process, while the RX 7600M uses AMD's RDNA 3.0 architecture (Navi 33 chip, codename "Hotpink Bonefish") on TSMC's 6 nm process. The CMP 90HX packs 28,300 million transistors on a 628 mm² die, whereas the RX 7600M has 13,300 million transistors on a 204 mm² die.

Q: How do memory subsystems differ?

A: The CMP 90HX features 10 GB of GDDR6X on a 320-bit bus with 760.3 GB/s bandwidth. The RX 7600M has 8 GB of GDDR6 on a 128-bit bus with 256.0 GB/s bandwidth—roughly one-third the memory bandwidth of the NVIDIA card.

Q: Which GPU has better pixel and texture throughput?

A: Despite having fewer cores, the RX 7600M achieves a higher pixel rate at 154.2 GPixel/s versus 136.8 GPixel/s for the CMP 90HX. However, the CMP 90HX leads in texture rate at 342.0 GTexel/s compared to 269.9 GTexel/s for the AMD chip.

Q: What is the production status and release timeline?

A: The CMP 90HX is end-of-life and was released on July 27, 2021. The RX 7600M is active and was released on January 3, 2023, with its predecessor listed as Polaris Mobile.

Architecture Differences

The CMP 90HX and RX 7600M represent fundamentally different design philosophies. The NVIDIA chip is built on the Ampere architecture using Samsung's 8 nm node, a process that allows for massive transistor counts—28,300 million across a 628 mm² die. This yields a transistor density of 45.1M per mm². The RX 7600M, by contrast, uses AMD's RDNA 3.0 architecture on TSMC's 6 nm node, achieving a higher density of 65.2M per mm² despite having only 13,300 million transistors on a 204 mm² die. The smaller, denser AMD chip demonstrates a more modern manufacturing approach, but the NVIDIA GPU compensates with sheer silicon area.

Core configurations diverge sharply. The CMP 90HX houses 6,400 shading units, 200 TMUs, and 80 ROPs, along with 50 RT cores and 200 tensor cores. The RX 7600M fields just 1,792 shading units, 112 TMUs, and 64 ROPs, with 28 RT cores and no tensor cores. This 3.6x difference in shading units explains why the NVIDIA card achieves 21.89 TFLOPS FP32 performance versus 17.27 TFLOPS for the AMD chip. Interestingly, the RX 7600M doubles its FP32 output in FP16 (34.55 TFLOPS at a 2:1 ratio), while the CMP 90HX maintains a 1:1 ratio, delivering 21.89 TFLOPS in both precisions.

Memory architecture reinforces the separation. The CMP 90HX's 320-bit bus with GDDR6X delivers 760.3 GB/s, while the RX 7600M's 128-bit bus with GDDR6 provides 256.0 GB/s. This 3x bandwidth advantage is critical for the NVIDIA card's compute-heavy workload. Conversely, the RX 7600M boosts to 2,410 MHz versus the CMP 90HX's 1,710 MHz, though both share a 1,500 MHz base clock. The AMD card's higher boost clock partially compensates for its fewer cores, but cannot overcome the bandwidth deficit.

Power and physical characteristics could not differ more. The CMP 90HX draws 320 W with a 700 W suggested PSU, uses dual-slot cooling with 2x 8-pin connectors, and measures 285 mm in length. The RX 7600M is an IGP (integrated graphics processor) with no power connectors, a 90 W TDP, and no listed dimensions. The NVIDIA card has no display outputs, while the RX 7600M's outputs are portable-device dependent. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, with the CMP 90HX also supporting PCIe 1.0 x4 while the RX 7600M uses PCIe 4.0 x16.

Head-to-Head Benchmarks

The only recorded head-to-head benchmark is Geekbench OpenCL, where the CMP 90HX scores 69,000 against the RX 7600M's 63,775, yielding an 8.2% victory for NVIDIA. This win aligns with the CMP 90HX's higher shading unit count and memory bandwidth, but the margin is narrower than the core disparity suggests. The RX 7600M's higher boost clock and improved pixel rate (154.2 GPixel/s vs 136.8 GPixel/s) hint at architectural efficiency that narrows the gap.

Context from nearest rivals illuminates the competitive landscape. The CMP 90HX's 69,000 sits just 0.3% above the Intel Arc A770 (68,809) and 0.6% above the AMD Radeon Instinct MI25 (68,562), but trails the AMD Radeon Pro WX 8200 (69,870) by 1.2% and the NVIDIA Quadro P6000 (69,986) by 1.4%. This clustering within a 1.5% range suggests the CMP 90HX is precisely at a performance plateau—neither dominant nor deficient among its peers.

The RX 7600M's 63,775 places it in an even tighter cluster. It is 0.1% behind the RX 9060 XT LP (63,830) and CMP 30HX (63,842), 0.1% ahead of the Radeon Pro Vega 56 (63,693), and 0.7% behind the Radeon Pro WX 9100 (64,212). All four rivals fall within a 0.8% band, making the RX 7600M's position virtually indistinguishable from its immediate competition. The 8.2% gap between the two headline GPUs, therefore, represents a meaningful but not overwhelming performance delta.

Specification Differences

| Specification | NVIDIA CMP 90HX | AMD Radeon RX 7600M |

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

| Architecture | Ampere | RDNA 3.0 |

| Process Node | 8 nm (Samsung) | 6 nm (TSMC) |

| Transistors | 28,300 million | 13,300 million |

| Die Size | 628 mm² | 204 mm² |

| Transistor Density | 45.1M / mm² | 65.2M / mm² |

| Boost Clock | 1710 MHz | 2410 MHz |

| Game Clock | None | 2070 MHz |

| Memory Size | 10 GB | 8 GB |

| Memory Type | GDDR6X | GDDR6 |

| Memory Bus | 320 bit | 128 bit |

| Memory Bandwidth | 760.3 GB/s | 256.0 GB/s |

| Shading Units | 6400 | 1792 |

| TMUs | 200 | 112 |

| ROPs | 80 | 64 |

| RT Cores | 50 | 28 |

| Tensor Cores | 200 | None |

| FP32 | 21.89 TFLOPS | 17.27 TFLOPS |

| FP16 | 21.89 TFLOPS (1:1) | 34.55 TFLOPS (2:1) |

| Pixel Rate | 136.8 GPixel/s | 154.2 GPixel/s |

| Texture Rate | 342.0 GTexel/s | 269.9 GTexel/s |

| TDP | 320 W | 90 W |

| Slot Width | Dual-slot | IGP |

| Power Connectors | 2x 8-pin | None |

| Suggested PSU | 700 W | None |

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

| Display Outputs | No outputs | Portable Device Dependent |

| Dimensions | 285 mm x 112 mm | Not listed |

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

| Release Date | 2021-07-27 | 2023-01-03 |

The Verdict

The data presents a clear performance hierarchy: the CMP 90HX wins the only benchmark, with an 8.2% lead in Geekbench OpenCL. Its advantages in shading units (6,400 vs 1,792), memory bandwidth (760.3 GB/s vs 256.0 GB/s), and FP32 throughput (21.89 vs 17.27 TFLOPS) are decisive for compute workloads. The RX 7600M counters with a higher boost clock (2,410 vs 1,710 MHz), superior pixel rate (154.2 vs 136.8 GPixel/s), and better FP16 performance (34.55 vs 21.89 TFLOPS), yet these strengths do not translate into a benchmark victory.

However, the verdict depends entirely on use case. The CMP 90HX is a mining GPU with no display outputs, a dual-slot 320 W design, and end-of-life status. It cannot render frames for gaming or professional visualization. The RX 7600M is an active mobile GPU with display outputs, a 90 W TDP, and integration into portable devices. For any application requiring visual output, the CMP 90HX is disqualified by design. For raw compute in a headless mining rig, the CMP 90HX's 8.2% advantage and 10 GB GDDR6X memory make it the stronger choice.

The percentile rankings—90th for CMP 90HX versus 89th for RX 7600M—confirm that both GPUs perform in the upper echelon of all graphics cards. The RX 7600M achieves this with one-third the power draw and a fraction of the die area, suggesting AMD's 6 nm process and RDNA 3.0 architecture deliver superior efficiency. The CMP 90HX's 320 W TDP versus the RX 7600M's 90 W TDP underscores this efficiency gap—the NVIDIA card consumes 3.6x more power to achieve just 8.2% more performance.

Where Each One Wins

NVIDIA CMP 90HX wins in compute-intensive scenarios. The 8.2% Geekbench OpenCL lead, combined with 3x memory bandwidth and 3.6x shading units, makes it the clear choice for headless compute tasks such as cryptocurrency mining, scientific simulations, or data processing where display output is unnecessary. Its 10 GB GDDR6X buffer and 200 tensor cores provide additional flexibility for AI-adjacent workloads, though the lack of display outputs limits its application scope. The 90th percentile ranking places it among the top tier of GPUs, and its nearest rivals—all within 1.4%—confirm it holds its own against professional-grade cards like the Quadro P6000 and Radeon Pro WX 8200.

AMD Radeon RX 7600M wins in mobile and efficiency-driven scenarios. As an IGP with a 90 W TDP and portable-device-dependent outputs, it is designed for laptops where power and space are constrained. Its higher boost clock and pixel rate suggest better responsiveness in graphics-heavy tasks at lower resolutions, and its FP16 throughput (34.55 TFLOPS) is 58% higher than the CMP 90HX, potentially benefiting applications that leverage half-precision arithmetic. The 89th percentile ranking shows it performs near the same tier as the CMP 90HX despite its mobile positioning. The RX 7600M also wins on architectural modernity—6 nm process, higher transistor density, and 2023 release date versus the CMP 90HX's 2021 launch and end-of-life status.

The tie-breaker is form factor and longevity. The CMP 90HX's dual-slot, 285 mm design and 700 W PSU requirement make it a desktop-only solution, while the RX 7600M's IGP form factor and active production status give it relevance in current mobile platforms. For buyers who need a GPU today in a laptop, the RX 7600M is the only viable option. For those building a headless compute rig with existing desktop infrastructure, the CMP 90HX's benchmark superiority—however modest—and its 10 GB memory capacity make it the data-backed winner. The 8.2% delta in OpenCL performance is real, but so is the 230 W TDP gap; the choice hinges on whether raw compute or system integration matters more.