AMD Radeon 660M vs AMD Radeon RX 5500M Comparison

The Verdict

The data is unambiguous: the AMD Radeon RX 5500M is the far faster part, and the AMD Radeon 660M is a compact integrated solution that trades heavily on efficiency. In the two head-to-head benchmark comparisons available, the RX 5500M wins both, with margins of 66.8% in Geekbench OpenCL and 58.7% in Geekbench Vulkan. The RX 5500M’s average benchmark score of 13356 puts it at the 54th percentile of all GPUs, while the 660M sits at 13356? No — the 660M’s average score is 13812, placing it at the 55th percentile. That makes the 660M nominally higher in percentile ranking, but this is a statistical quirk driven by the different suites of tests run on each card. The RX 5500M’s raw compute advantage is decisive. The 660M’s only real argument is its 40 W TDP versus 85 W for the RX 5500M, plus integrated simplicity with no separate memory of its own.

Pick the RX 5500M if you need discrete-level performance for gaming or compute workloads. Its 4 GB of dedicated GDDR6 memory and 224.0 GB/s bandwidth are far beyond the 660M’s system-shared memory. Pick the 660M if you want a low-power integrated GPU that can handle light tasks and some gaming at reduced settings. The 660M is end-of-life, as is the RX 5500M, so neither is a future-proof purchase. The RX 5500M’s nearest rival comparisons show it trading blows with the AMD FirePro M6100 (0% delta), AMD Radeon HD 8950M (-0.1%), and AMD Radeon Pro 555X (0.3%). The 660M’s rival set is more exotic, including the NVIDIA RTX A2000 Mobile (-0.1%), AMD Radeon RX 570X (-0.4%), and even the AMD Radeon RX 7900 XT (0.5%) — though those deltas are within a 2% band, meaning the 660M’s average score is essentially tied with that disparate group.

Architecture Differences

The two GPUs come from different generations of AMD’s RDNA architecture. The Radeon 660M is built on RDNA 2.0, while the Radeon RX 5500M uses the original RDNA 1.0. The 660M’s chip is codenamed Rembrandt, and its generation is listed as “Navi II IGP (Rembrandt Mobile).” The RX 5500M uses the Navi 14 chip, in the “Navi Mobile (RX 5000M)” generation. The 660M is fabricated on a 6 nm process at TSMC, with 13,100 million transistors on a 208 mm² die, yielding a transistor density of 63.0M per mm². The RX 5500M uses an older 7 nm process, also at TSMC, but packs just 6,400 million transistors onto a 158 mm² die, for a density of 40.5M per mm². The 660M’s newer node and higher transistor count explain how it fits more features into an integrated design.

The 660M has 384 shading units, 24 texture mapping units, 16 ROPs, and 6 ray tracing cores. It supports DirectX 12 Ultimate (12_2), which includes hardware ray tracing. The RX 5500M has 1408 shading units, 88 TMUs, and 32 ROPs, but no ray tracing cores at all. It supports DirectX 12 (12_1), a step below the 660M’s feature set. Both support OpenGL 4.6 and Vulkan 1.4. The 660M’s pixel rate is 30.40 GPixel/s and its texture rate is 45.60 GTexel/s; the RX 5500M’s pixel rate is 52.64 GPixel/s and texture rate is 144.8 GTexel/s. In FP32 compute, the 660M delivers 1,459.2 GFLOPS versus the RX 5500M’s 4.632 TFLOPS. FP16 figures are 2.918 TFLOPS (2:1) for the 660M and 9.265 TFLOPS (2:1) for the RX 5500M. The 660M’s memory is entirely system-shared, with bandwidth described as system dependent. The RX 5500M has 4 GB of GDDR6 on a 128-bit bus, with 224.0 GB/s of bandwidth. The 660M’s base clock is 1500 MHz with a boost of 1900 MHz; the RX 5500M’s base is 1375 MHz, boost 1645 MHz, and game clock 1448 MHz. The RX 5500M has a separate memory clock of 1750 MHz (14 Gbps effective). Both use a PCIe 4.0 x8 bus interface.

FAQ

Q: Which GPU is faster in raw compute benchmarks?

A: The AMD Radeon RX 5500M is dramatically faster. It wins Geekbench OpenCL with a score of 38725 versus 12876 for the 660M, a 66.8% delta. In Geekbench Vulkan, the RX 5500M scores 35693 against the 660M’s 14748, a 58.7% delta.

Q: Does the Radeon 660M support ray tracing?

A: Yes. The 660M has 6 ray tracing cores and supports DirectX 12 Ultimate (12_2). The RX 5500M has no ray tracing cores and is limited to DirectX 12 (12_1).

Q: How much memory does each GPU have?

A: The Radeon 660M has system-shared memory with system-dependent bandwidth. The Radeon RX 5500M has 4 GB of dedicated GDDR6 memory on a 128-bit bus, delivering 224.0 GB/s.

Q: Which GPU consumes less power?

A: The Radeon 660M has a 40 W TDP, less than half the RX 5500M’s 85 W TDP. This makes the 660M far more suited to thin-and-light designs.

Q: Are these GPUs still in production?

A: No. Both are listed as end-of-life. The 660M was released on 2022-01-03, and the RX 5500M on 2019-10-06.

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

A: The 660M’s average benchmark score of 13812 is within 0.5% of the NVIDIA RTX A2000 Mobile (13821), AMD Radeon RX 570X (13871), and AMD Radeon RX 7900 XT (13745). The RX 5500M’s average score of 13356 is essentially tied with the AMD FirePro M6100 (13354, 0% delta) and within 0.4% of the AMD Radeon HD 8950M and AMD Radeon Pro 555.

Specification Differences

The two GPUs differ across nearly every major specification field. The process node is 6 nm for the 660M versus 7 nm for the RX 5500M. Transistor count is 13,100 million versus 6,400 million, and die size is 208 mm² versus 158 mm², giving respective densities of 63.0M/mm² and 40.5M/mm². The 660M has 384 shading units, 24 TMUs, 16 ROPs, and 6 RT cores; the RX 5500M has 1408 shading units, 88 TMUs, and 32 ROPs, with no RT cores. Base clocks are 1500 MHz versus 1375 MHz, and boost clocks are 1900 MHz versus 1645 MHz; the RX 5500M adds a game clock of 1448 MHz and a memory clock of 1750 MHz (14 Gbps effective). Memory configuration is system-shared for the 660M versus 4 GB GDDR6 on a 128-bit bus with 224.0 GB/s for the RX 5500M. Pixel rate is 30.40 GPixel/s versus 52.64 GPixel/s, texture rate is 45.60 GTexel/s versus 144.8 GTexel/s, and FP32 compute is 1,459.2 GFLOPS versus 4.632 TFLOPS. FP16 is 2.918 TFLOPS versus 9.265 TFLOPS (both 2:1). TDP is 40 W versus 85 W. The 660M is an IGP with no slot width; the RX 5500M’s slot width is not listed. The 660M supports DirectX 12 Ultimate (12_2), while the RX 5500M is limited to DirectX 12 (12_1); both support OpenGL 4.6 and Vulkan 1.4. The 660M’s predecessor is Vega II IGP and its successor is Navi III IGP; the RX 5500M’s predecessor is Polaris Mobile with no successor listed. Release dates differ by over two years: 2022-01-03 for the 660M versus 2019-10-06 for the RX 5500M.

Head-to-Head Benchmarks

The head-to-head results are lopsided. In Geekbench OpenCL, the RX 5500M scores 38725 against the 660M’s 12876, a margin of 66.8% in favor of the discrete card. That is not a close race by any measure. The RX 5500M’s 1408 shading units and 224.0 GB/s of dedicated memory bandwidth simply overwhelm the 660M’s 384 shading units and system-shared memory. In Geekbench Vulkan, the gap narrows slightly but remains huge: the RX 5500M scores 35693 versus 14748 for the 660M, a 58.7% delta. The 660M’s newer RDNA 2.0 architecture and ray tracing cores help it close some of the gap in Vulkan, but the raw compute throughput difference is too large to overcome. The RX 5500M wins both tests, and the data shows no test where the 660M comes out ahead.

Looking at the broader benchmark picture, the RX 5500M has a much richer suite of results. It scores 50359 in Geekbench Metal, 38725 in OpenCL, and 35693 in Vulkan, plus a range of PassMark tests: 39 in DirectX 10, 35 in DirectX 11, 28 in DirectX 12, 100 in DirectX 9, 414 in G2D, 5848 in G3D, and 2316 in GPU compute. The 660M only has Geekbench OpenCL and Vulkan scores. The RX 5500M’s PassMark G3D score of 5848 and compute score of 2316 indicate solid directX performance, though its DirectX 9 score of 100 is notably high. The 660M’s average benchmark score of 13812 is actually higher than the RX 5500M’s 13356, but that is because the 660M’s two scores are both Geekbench results, while the RX 5500M’s average includes lower PassMark scores that drag it down. In the tests they share — OpenCL and Vulkan — the RX 5500M wins decisively. The percentile rankings confirm this: the 660M sits at the 55th percentile and the RX 5500M at the 54th, a negligible difference that does not reflect the head-to-head deltas. The RX 5500M’s nearest rivals include the AMD FirePro M6100 (13354, 0% delta) and AMD Radeon Pro 555X (13321, 0.3% delta), while the 660M’s nearest rivals are a mixed bag including the NVIDIA RTX A2000 Mobile (13821, -0.1%) and the flagship AMD Radeon RX 7900 XT (13745, 0.5%). Those rival sets show that the 660M’s average score is competitive with much more expensive discrete GPUs on paper, but the head-to-head benchmark data tells the real story: the RX 5500M is the performance winner by a wide margin.