AMD Radeon 660M vs AMD Radeon Pro 560X Comparison

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon Pro 560X holds the higher average, scoring 15,082 across all benchmarks, while the AMD Radeon 660M averages 13,812. This places the Pro 560X in the 57th percentile of all GPUs, versus the 660M's 55th percentile.

Q: How do the two compare in OpenCL performance?

A: The AMD Radeon 660M wins decisively in the Geekbench OpenCL test, scoring 12,876 against the Pro 560X's 9,663. The deltaPct of -25% indicates the Pro 560X trails by 25% in this workload.

Q: Which GPU wins in Vulkan benchmarks?

A: The AMD Radeon Pro 560X takes the Vulkan test with a score of 16,819, beating the 660M's 14,748. The Pro 560X leads by 14% in this API, according to the deltaPct.

Q: What are the memory configurations of these two GPUs?

A: The AMD Radeon Pro 560X features 4 GB of dedicated GDDR5 memory on a 128-bit bus, delivering 94.08 GB/s bandwidth. The AMD Radeon 660M uses System Shared memory, with its bandwidth listed as System Dependent.

Q: Do both GPUs support the same level of DirectX?

A: No. The AMD Radeon Pro 560X supports DirectX 12 (12_0), while the AMD Radeon 660M supports DirectX 12 Ultimate (12_2), a newer and more feature-rich specification.

Q: What is the power draw difference between the two?

A: The AMD Radeon Pro 560X has a TDP of 75 W, whereas the AMD Radeon 660M is rated at 40 W. Both are integrated-class parts (IGP) with no power connectors.

Architecture Differences

The two GPUs come from fundamentally different design eras and philosophies. The AMD Radeon Pro 560X is built on the Polaris 21 chip using the GCN 4.0 architecture, fabricated on a 14 nm process by GlobalFoundries. It contains 3,000 million transistors on a 123 mm² die, yielding a transistor density of 24.4 million per mm². The AMD Radeon 660M, by contrast, uses the Rembrandt chip with RDNA 2.0 architecture, manufactured on TSMC's 6 nm node. This newer chip packs 13,100 million transistors across a 208 mm² die, achieving a density of 63.0 million per mm² — over 2.5 times the density of the older Polaris part.

The compute configurations diverge sharply. The Pro 560X carries 1,024 shading units, 64 TMUs, and 16 ROPs. The 660M has only 384 shading units and 24 TMUs, but maintains the same 16 ROPs. Critically, the 660M adds 6 ray tracing cores, a feature entirely absent from the GCN-based Pro 560X. This reflects the architectural shift: RDNA 2.0 is designed for modern rendering pipelines with hardware-accelerated ray tracing, while GCN 4.0 predates that capability.

Clock speeds and memory hierarchy differ as well. The 660M runs at a base clock of 1500 MHz with a boost up to 1900 MHz, while the Pro 560X's base and boost clocks are not specified in the data. The Pro 560X uses dedicated GDDR5 memory at 1470 MHz (5.9 Gbps effective), while the 660M relies on system shared memory. The bus interface also advances: the Pro 560X uses PCIe 3.0 x8, while the 660M uses PCIe 4.0 x8.

The 660M also supports a newer feature set, including DirectX 12 Ultimate (12_2) and Vulkan 1.4, versus the Pro 560X's DirectX 12 (12_0) and Vulkan 1.3. Both support OpenGL 4.6. The manufacturing process and transistor density differences point to the 660M being a much more modern and efficient design.

Head-to-Head Benchmarks

The two GPUs split their direct benchmark comparisons evenly, with one win apiece. In the Geekbench OpenCL test, the AMD Radeon 660M posts a score of 12,876, while the AMD Radeon Pro 560X manages only 9,663. The deltaPct of -25% confirms the Pro 560X is a full quarter behind in this compute-oriented workload. This is a substantial margin, likely reflecting the 660M's newer RDNA 2.0 architecture and its ability to execute certain compute tasks more efficiently despite having fewer shading units.

The Vulkan benchmark flips the result. The AMD Radeon Pro 560X scores 16,819, beating the 660M's 14,748 by a 14% margin. This is a notable victory for the older GCN part, suggesting that in this particular graphics API workload, the Pro 560X's higher shading unit count (1,024 vs. 384) and dedicated memory bandwidth provide a decisive advantage. The 660M's reliance on system shared memory and its lower TMU count may hold it back in this scenario.

Looking at the broader benchmark landscape, the Pro 560X's average score of 15,082 puts it within 0.1% of the NVIDIA GeForce GTX 660 Ti (15,063) and 0.6% behind the AMD Radeon RX 7600 (15,171). The 660M's average of 13,812 places it 0.1% ahead of the NVIDIA RTX A2000 Mobile (13,821) and 0.5% above the AMD Radeon RX 7900 XT (13,745). These rival comparisons show that while the Pro 560X generally sits in a slightly higher performance tier, the 660M is competitive within its own class, and the head-to-head results demonstrate that neither GPU is universally superior.

The Verdict

The data paints a picture of two GPUs with complementary strengths. The AMD Radeon Pro 560X is the better choice for Vulkan-based workloads, where it outperforms the 660M by 14%, and it holds a higher overall average benchmark score (15,082 vs. 13,812). It also ranks higher in the global percentile (57th vs. 55th). For users prioritizing raw graphics API performance in Vulkan or seeking a GPU with dedicated 4 GB GDDR5 memory and 94.08 GB/s bandwidth, the Pro 560X is the data-backed selection.

The AMD Radeon 660M is the superior option for OpenCL compute tasks, beating the Pro 560X by 25% in that benchmark. It also offers modern features that the older GCN part lacks: ray tracing cores, DirectX 12 Ultimate support, Vulkan 1.4 compatibility, and a significantly more power-efficient design at 40 W versus 75 W. The 660M's 6 nm process and RDNA 2.0 architecture represent a generational leap in efficiency, and its higher transistor density (63.0M / mm² vs. 24.4M / mm²) indicates a more advanced manufacturing technology.

Neither GPU is a clear overall winner — the split is exactly 1 win each in head-to-head tests. The Pro 560X appeals to those who need dedicated memory and stronger Vulkan performance; the 660M appeals to those who prioritize compute throughput, modern API support, and lower power consumption. The choice hinges entirely on the workload.

Specification Differences

The two GPUs differ across nearly every major specification category. The process node advances from 14 nm (GlobalFoundries) on the Pro 560X to 6 nm (TSMC) on the 660M. Transistor count jumps from 3,000 million to 13,100 million, while die size grows from 123 mm² to 208 mm². Transistor density more than doubles, from 24.4M / mm² to 63.0M / mm².

The compute units are starkly different: the Pro 560X has 1,024 shading units and 64 TMUs, while the 660M has 384 shading units and 24 TMUs. Both have 16 ROPs. The 660M adds 6 ray tracing cores; the Pro 560X has none. Clock speeds are only specified for the 660M (1500 MHz base, 1900 MHz boost), with no base or boost data for the Pro 560X.

Memory configurations are fundamentally different. The Pro 560X uses 4 GB of GDDR5 with a 128-bit bus and 94.08 GB/s bandwidth, while the 660M uses System Shared memory with System Dependent bandwidth. The memory clock is 1470 MHz (5.9 Gbps effective) for the Pro 560X; the 660M's memory clock is not applicable.

Pixel and texture rates favor different aspects: the Pro 560X has a higher texture rate at 64.26 GTexel/s versus 45.60 GTexel/s, but the 660M has a higher pixel rate at 30.40 GPixel/s versus 16.06 GPixel/s. FP32 performance is 2.056 TFLOPS for the Pro 560X versus 1,459.2 GFLOPS for the 660M, while FP16 performance is 2.056 TFLOPS (1:1) for the Pro 560X versus 2.918 TFLOPS (2:1) for the 660M.

TDP differs significantly: 75 W for the Pro 560X versus 40 W for the 660M. The bus interface advances from PCIe 3.0 x8 to PCIe 4.0 x8. API support differs in DirectX (12_0 vs. 12 Ultimate) and Vulkan (1.3 vs. 1.4). The Pro 560X launched in 2018, while the 660M launched in 2022. The 660M has a predecessor (Vega II IGP) and successor (Navi III IGP); the Pro 560X has neither listed. Both are IGP slot width with no power connectors.

Where Each One Wins

AMD Radeon Pro 560X wins in:

• Vulkan graphics performance, leading by 14% (16,819 vs. 14,748)
• Overall average benchmark score (15,082 vs. 13,812) and global percentile ranking (57th vs. 55th)
• Texture throughput, with a texture rate of 64.26 GTexel/s versus 45.60 GTexel/s
• FP32 compute, at 2.056 TFLOPS versus 1,459.2 GFLOPS
• Dedicated memory bandwidth, with 94.08 GB/s versus system dependent shared memory

AMD Radeon 660M wins in:

• OpenCL compute, beating the Pro 560X by 25% (12,876 vs. 9,663)
• Pixel fill rate, at 30.40 GPixel/s versus 16.06 GPixel/s
• FP16 performance, at 2.918 TFLOPS versus 2.056 TFLOPS
• Power efficiency, with a 40 W TDP versus 75 W
• Feature set, including 6 ray tracing cores, DirectX 12 Ultimate, and Vulkan 1.4
• Manufacturing technology, with a 6 nm process and 63.0M / mm² density versus 14 nm and 24.4M / mm²

For users running Vulkan-based applications or workloads that favor dedicated memory and high texture throughput, the Pro 560X is the stronger option. For those running OpenCL compute tasks, needing ray tracing support, or prioritizing power efficiency and modern API compatibility, the 660M is the clear pick. The split benchmark results confirm that the choice depends entirely on the specific application and workload requirements.