AMD Radeon 660M vs AMD Radeon Vega 11 Comparison

# AMD Radeon Vega 11 vs AMD Radeon 660M

The AMD Radeon Vega 11 and AMD Radeon 660M are two integrated GPUs from different eras, and the benchmark data reveals a clear split: the Vega 11 wins in OpenCL compute workloads, while the 660M dominates in Vulkan graphics performance. With an average benchmark score of 14,352 versus 13,812, the Vega 11 holds a 3.9% overall edge, but the 660M’s architectural advantages in modern APIs make it the stronger choice for graphics-heavy tasks. Both GPUs occupy similar percentile positions — 56th and 55th respectively — but their strengths are sharply delineated across different test types.

Head-to-Head Benchmarks

The direct comparison between these two IGPs is defined by a single decisive split. In Geekbench OpenCL, the Vega 11 scores 13,392 against the 660M’s 12,876, giving the older part a 4% victory. This is a meaningful margin, especially considering that the 660M runs at a higher base clock (1,500 MHz vs 300 MHz) and boost clock (1,900 MHz vs 1,400 MHz). The Vega 11’s advantage in raw compute throughput — 1.971 TFLOPS FP32 versus 1,459.2 GFLOPS — translates directly into OpenCL performance, where the 704 shading units outperform the 660M’s 384 shading units despite the latter’s newer architecture.

The Vulkan results tell the opposite story. The 660M delivers 14,748 points, crushing the Vega 11’s 12,273 by a substantial 16.8% margin. This is a larger gap than the OpenCL difference, and it highlights how the RDNA 2.0 architecture excels in modern graphics APIs. The 660M’s 16 ROPs (double the Vega 11’s 8) and its pixel rate of 30.40 GPixel/s versus 11.20 GPixel/s contribute heavily to this outcome. The Vega 11’s higher texture rate (61.60 GTexel/s vs 45.60 GTexel/s) helps in some workloads, but Vulkan clearly favors the newer design.

When looking at the broader competitive landscape, the Vega 11 sits within 0.4% of the NVIDIA GeForce GTX 965M (14,404) and 0.1% of the NVIDIA GeForce GTX TITAN (14,373). The 660M, meanwhile, trails the NVIDIA RTX A2000 Mobile by just 0.1% (13,821 vs 13,821) and the AMD Radeon RX 570X by 0.4% (13,871). Interestingly, the 660M is actually 0.5% ahead of the AMD Radeon RX 7900 XT (13,745), a flagship discrete card — though that delta is within noise. The overall averages — 14,352 for Vega 11 and 13,812 for 660M — place them in nearly identical performance tiers, but the distribution of scores is what matters.

Architecture Differences

The two GPUs represent fundamentally different design philosophies. The Vega 11 uses the GCN 5.0 architecture, fabricated on GlobalFoundries’ 12 nm process, while the 660M leverages RDNA 2.0 on TSMC’s 6 nm node. This process shrink is dramatic: the 660M’s chip packs 13,100 million transistors into a 208 mm² die, yielding a transistor density of 63.0M per mm². The Vega 11, by contrast, has only 4,940 million transistors on a 210 mm² die, for a density of 23.5M per mm². The 660M achieves nearly triple the density, which explains how it can deliver competitive performance with far fewer shading units.

Shader configurations diverge sharply. The Vega 11 fields 704 shading units, 44 TMUs, and 8 ROPs. The 660M counters with just 384 shading units and 24 TMUs, but doubles the ROP count to 16. Crucially, the 660M adds 6 ray tracing cores — a feature the Vega 11 lacks entirely, as it has no RT cores. This makes the 660M DirectX 12 Ultimate (12_2) compliant, while the Vega 11 only reaches DirectX 12 (12_1). The 660M also supports Vulkan 1.4 versus the Vega 11’s Vulkan 1.3, and both offer OpenGL 4.6.

Clock behavior highlights the efficiency gains of RDNA 2.0. The 660M operates at a 1,500 MHz base and 1,900 MHz boost, significantly higher than the Vega 11’s 300 MHz base and 1,400 MHz boost. Despite these higher clocks, the 660M’s FP32 throughput is lower (1,459.2 GFLOPS vs 1.971 TFLOPS), because fewer shading units are active. FP16 performance follows a similar pattern: the Vega 11 delivers 3.942 TFLOPS (2:1) versus the 660M’s 2.918 TFLOPS (2:1). The 660M’s pixel rate of 30.40 GPixel/s is nearly triple the Vega 11’s 11.20 GPixel/s, but the Vega 11’s texture rate of 61.60 GTexel/s beats the 660M’s 45.60 GTexel/s.

Power envelopes differ considerably. The Vega 11 is rated at 15 W TDP, while the 660M draws 40 W. Both are IGP-class parts with no power connectors, and both use system-shared memory with system-dependent bandwidth. The bus interface also differs: the Vega 11 uses IGP, while the 660M connects via PCIe 4.0 x8. The 660M’s release date of January 2022 comes more than two years after the Vega 11’s September 2019 launch, and both are now end-of-life products. The Vega 11 succeeds GCN 3.0 IGP and is succeeded by Vega II IGP, while the 660M follows Vega II IGP and leads to Navi III IGP.

The Verdict

The data points to a clear conclusion: choose the AMD Radeon 660M for modern graphics workloads, and the AMD Radeon Vega 11 for raw compute. The 660M’s 16.8% Vulkan win is decisive and reflects its RDNA 2.0 architecture, ray tracing support, and higher pixel throughput. The Vega 11’s 4% OpenCL advantage, while real, is narrower and less impactful for gaming or DirectX 12 Ultimate applications. The 660M also matches or beats its rivals more convincingly: it sits 0.5% ahead of the RX 7900 XT, whereas the Vega 11 trails the GTX TITAN by 0.1%.

For users prioritizing longevity and API support, the 660M is the obvious pick. DirectX 12 Ultimate, Vulkan 1.4, and 6 RT cores make it future-proof in ways the Vega 11 cannot match. The Vega 11’s higher TMU count and texture rate are useful for specific compute tasks, but they do not compensate for the 660M’s modern feature set. The 660M’s 40 W TDP is higher than the Vega 11’s 15 W, but that power budget enables significantly higher clocks and better graphics performance. If you are choosing between these two for a new system, the 660M is the stronger all-around part unless your workload is exclusively OpenCL compute.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon Vega 11 scores 14,352 on average, while the AMD Radeon 660M scores 13,812 — a 3.9% difference in favor of the Vega 11.

Q: How big is the Vulkan performance gap between the two?

A: The 660M scores 14,748 in Geekbench Vulkan versus the Vega 11’s 12,273, giving the 660M a 16.8% lead.

Q: Does the AMD Radeon 660M support ray tracing?

A: Yes, the 660M includes 6 ray tracing cores, while the Vega 11 has no RT cores at all.

Q: What are the DirectX version differences?

A: The 660M supports DirectX 12 Ultimate (12_2), whereas the Vega 11 is limited to DirectX 12 (12_1).

Q: Which GPU has more shading units?

A: The Vega 11 has 704 shading units, compared to the 660M’s 384 shading units.

Q: What are the TDP ratings for each GPU?

A: The Vega 11 is rated at 15 W, while the 660M has a 40 W TDP.

Where Each One Wins

AMD Radeon Vega 11 wins in OpenCL compute, leveraging its 704 shading units and 1.971 TFLOPS FP32 to outpace the 660M by 4% (13,392 vs 12,876). Its texture rate of 61.60 GTexel/s exceeds the 660M’s 45.60 GTexel/s, making it better suited for texture-heavy compute workloads. The Vega 11 also holds a higher average benchmark score (14,352 vs 13,812) and a slightly better percentile ranking (56th vs 55th). It achieves this with a much lower 15 W TDP, making it the efficiency choice for compute tasks.

AMD Radeon 660M wins decisively in Vulkan, posting 14,748 versus 12,273 — a 16.8% advantage. Its 16 ROPs and 30.40 GPixel/s pixel rate (nearly triple the Vega 11’s 11.20 GPixel/s) make it superior for graphics rendering. The 660M also brings ray tracing, DirectX 12 Ultimate, Vulkan 1.4, and a PCIe 4.0 x8 interface. Its 6 nm process and 63.0M / mm² transistor density allow higher clocks (1,900 MHz boost) with fewer shaders. For gaming or any modern API workload, the 660M is the clear winner.

Specification Differences

| Specification | AMD Radeon Vega 11 | AMD Radeon 660M |

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

| Architecture | GCN 5.0 | RDNA 2.0 |

| Process Node | 12 nm | 6 nm |

| Foundry | GlobalFoundries | TSMC |

| Transistors | 4,940 million | 13,100 million |

| Die Size | 210 mm² | 208 mm² |

| Transistor Density | 23.5M / mm² | 63.0M / mm² |

| Base Clock | 300 MHz | 1,500 MHz |

| Boost Clock | 1,400 MHz | 1,900 MHz |

| Shading Units | 704 | 384 |

| TMUs | 44 | 24 |

| ROPs | 8 | 16 |

| RT Cores | None | 6 |

| Pixel Rate | 11.20 GPixel/s | 30.40 GPixel/s |

| Texture Rate | 61.60 GTexel/s | 45.60 GTexel/s |

| FP32 | 1.971 TFLOPS | 1,459.2 GFLOPS |

| FP16 | 3.942 TFLOPS (2:1) | 2.918 TFLOPS (2:1) |

| TDP | 15 W | 40 W |

| Bus Interface | IGP | PCIe 4.0 x8 |

| DirectX | 12 (12_1) | 12 Ultimate (12_2) |

| Vulkan | 1.3 | 1.4 |

| Release Date | 2019-09-29 | 2022-01-03 |