AMD Radeon R9 M375 vs NVIDIA GeForce GTX 660 Comparison

The Verdict

The data separates these two clearly. The NVIDIA GeForce GTX 660 wins both recorded head-to-head benchmarks, and the margin is not trivial. In Geekbench OpenCL, the GTX 660 scores 11347 against the Radeon R9 M375's 10457, a 7.8% advantage. The Vulkan gap widens further: 11415 versus 9682, a 15.2% lead for NVIDIA. If raw compute performance is the priority, the GTX 660 is the pick.

However, the R9 M375 is not without merit. Its average benchmark score of 10070 actually exceeds the GTX 660's 9022 average, despite losing both direct comparisons. This happens because the GTX 660 also carries a Geekbench Metal score of 4305, which drags its average down. The R9 M375's percentile rank of 48 also sits above the GTX 660's 45. So the AMD part is more consistent across its tested workloads, while the NVIDIA part peaks higher in the two shared tests.

The GTX 660 targets desktop builders who need a proven, end-of-life card with a 140 W TDP and a 1x 6-pin power connector. The R9 M375 is a mobile part (Tropo chip, GCN 1.0) with no TDP listed, no power connector, and no dimensions in the database, suggesting it was designed for laptops. Choose the GTX 660 for higher peak compute in OpenCL and Vulkan. Choose the R9 M375 if you need a mobile GPU with a better average benchmark score and a higher percentile standing relative to all GPUs.

Architecture Differences

The two GPUs come from different architectural lineages. The AMD Radeon R9 M375 uses the Tropo chip built on GCN 1.0, fabricated by TSMC on a 28 nm process. It packs 1,500 million transistors on a 123 mm² die, yielding a transistor density of 12.2M per mm². The NVIDIA GeForce GTX 660 uses the GK106 chip with Kepler architecture, also on TSMC's 28 nm node. It carries 2,540 million transistors across a larger 221 mm² die, but its density is slightly lower at 11.5M per mm².

Shader resources differ significantly. The R9 M375 has 640 shading units, 40 texture mapping units, and 16 ROPs. The GTX 660 more than doubles the TMUs (80) and raises ROPs to 24, while shading units jump to 960. This explains the GTX 660's higher texture rate: 82.56 GTexel/s versus 40.60 GTexel/s. Pixel rates also favor NVIDIA: 20.64 GPixel/s versus 16.24 GPixel/s.

Memory configurations are completely different. The R9 M375 uses 2 GB of DDR3 on a 128-bit bus, delivering 28.80 GB/s of bandwidth. The GTX 660 uses 2 GB of GDDR5 on a 192-bit bus, delivering 144.2 GB/s, exactly five times the bandwidth. Clock speeds are close (1000 MHz base for AMD, 980 MHz base for NVIDIA), but memory clocks diverge: 900 MHz (1800 Mbps effective) for AMD versus 1502 MHz (6 Gbps effective) for NVIDIA.

Feature support shows minor API differences. Both support DirectX 12, but the R9 M375 lists 12 (11_1) while the GTX 660 lists 12 (11_0). OpenGL is identical at 4.6. Vulkan versions are near-identical: 1.2.170 for AMD, 1.2.175 for NVIDIA. The GTX 660 includes display outputs (2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2), while the R9 M375 lists no display outputs in the database.

Head-to-Head Benchmarks

The first shared test is Geekbench OpenCL. The GTX 660 scores 11347, the R9 M375 scores 10457. The delta is -7.8% for AMD, meaning NVIDIA wins by that margin. This is a meaningful gap, but not a rout. The R9 M375's lower bandwidth (28.80 GB/s) and fewer ROPs likely contribute to the deficit, though the database does not break down per-test factors.

The second shared test is Geekbench Vulkan. Here the GTX 660 extends its lead: 11415 versus 9682, a -15.2% delta for AMD. This is more than double the OpenCL gap. The Vulkan result suggests the NVIDIA architecture handles the newer API workload more efficiently, or the memory bandwidth advantage becomes more critical in that test. Either way, the GTX 660 wins decisively.

Looking at the broader benchmark picture, the R9 M375 has only two recorded scores: OpenCL (10457) and Vulkan (9682). The GTX 660 has three: OpenCL (11347), Vulkan (11415), and Metal (4305). The Metal score is dramatically lower, pulling the GTX 660's average down to 9022. The R9 M375's average of 10070 comes from its two scores alone. So while the GTX 660 wins both head-to-head tests, its overall average is lower because of that Metal outlier. This is a critical nuance for anyone comparing averages without looking at individual tests.

The R9 M375 sits at the 48th percentile of all GPUs, versus the GTX 660's 45th percentile. Both are mid-pack parts. The R9 M375's nearest rival, the NVIDIA Quadro K5100M, scores 10043 (0.3% behind), while the AMD Radeon Pro 5300M scores 10013 (0.6% behind). The GTX 660's nearest rivals include the NVIDIA TITAN V CEO Edition at 9037 (-0.2%) and the NVIDIA GeForce GTX 560 at 9058 (-0.4%). These comparisons show both cards are tightly clustered with their peers.

FAQ

Q: Which GPU is faster in Geekbench OpenCL?

A: The NVIDIA GeForce GTX 660 scores 11347, beating the AMD Radeon R9 M375's 10457 by 7.8%.

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon R9 M375 has an average score of 10070, which is higher than the GTX 660's 9022. This is because the GTX 660's Metal score of 4305 lowers its average.

Q: What is the memory bandwidth difference?

A: The GTX 660 has 144.2 GB/s of bandwidth (GDDR5, 192-bit bus), while the R9 M375 has 28.80 GB/s (DDR3, 128-bit bus). That is a five-fold advantage for NVIDIA.

Q: Do both cards support DirectX 12?

A: Yes, both support DirectX 12. The R9 M375 lists 12 (11_1), while the GTX 660 lists 12 (11_0). Both also support OpenGL 4.6.

Q: What is the GTX 660's power requirement?

A: The GTX 660 has a 140 W TDP, a 1x 6-pin power connector, and a suggested PSU of 300 W. The R9 M375 has no TDP or power connector listed in the database.

Q: Which GPU has a higher percentile ranking?

A: The R9 M375 ranks at the 48th percentile of all GPUs, while the GTX 660 ranks at the 45th percentile.

Where Each One Wins

The GTX 660 wins in every direct comparison recorded. In Geekbench OpenCL, it is 7.8% ahead. In Geekbench Vulkan, it is 15.2% ahead. It also has overwhelming advantages in raw throughput metrics: 1.981 TFLOPS FP32 versus 1,299.2 GFLOPS, 82.56 GTexel/s versus 40.60 GTexel/s, and 20.64 GPixel/s versus 16.24 GPixel/s. If your workload is OpenCL or Vulkan compute, the GTX 660 is the clear choice. It also offers more display outputs (2x DVI, HDMI, DisplayPort) and a longer track record as a desktop card with a 241 mm length (9.5 inches).

The R9 M375 wins on average benchmark score (10070 versus 9022) and percentile ranking (48 versus 45). It also has a smaller die (123 mm² versus 221 mm²) and fewer transistors (1,500 million versus 2,540 million), which suggests lower power draw and thermal output, though the database does not list a TDP for the AMD part. For a mobile GPU, the R9 M375's lower resource footprint may be an advantage in constrained chassis. Its Vulkan score of 9682 is respectable, and its OpenCL score of 10457 is within 8% of the GTX 660 despite the massive bandwidth deficit.

For gamers or compute users on a desktop with a 300 W PSU budget, the GTX 660 is straightforward. For laptop builders or those prioritizing a higher average score and percentile rank, the R9 M375 is the data-backed option. Neither card is current; both are end-of-life. But the GTX 660 is the stronger performer in the tests that matter most for modern compute APIs.

Specification Differences

| Specification | AMD Radeon R9 M375 | NVIDIA GeForce GTX 660 |

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

| Architecture | GCN 1.0 | Kepler |

| Chip | Tropo | GK106 |

| Process node | 28 nm | 28 nm |

| Transistors | 1,500 million | 2,540 million |

| Die size | 123 mm² | 221 mm² |

| Transistor density | 12.2M / mm² | 11.5M / mm² |

| Base clock | 1000 MHz | 980 MHz |

| Boost clock | 1015 MHz | 1032 MHz |

| Memory clock | 900 MHz (1800 Mbps effective) | 1502 MHz (6 Gbps effective) |

| Memory size | 2 GB | 2 GB |

| Memory type | DDR3 | GDDR5 |

| Memory bus width | 128 bit | 192 bit |

| Memory bandwidth | 28.80 GB/s | 144.2 GB/s |

| Shading units | 640 | 960 |

| TMUs | 40 | 80 |

| ROPs | 16 | 24 |

| Pixel rate | 16.24 GPixel/s | 20.64 GPixel/s |

| Texture rate | 40.60 GTexel/s | 82.56 GTexel/s |

| FP32 | 1,299.2 GFLOPS | 1.981 TFLOPS |

| TDP | Not listed | 140 W |

| Slot width | Not listed | Dual-slot |

| Power connectors | Not listed | 1x 6-pin |

| Suggested PSU | Not listed | 300 W |

| Display outputs | Not listed | 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 |

| DirectX | 12 (11_1) | 12 (11_0) |

| OpenGL | 4.6 | 4.6 |

| Vulkan | 1.2.170 | 1.2.175 |

| Length | Not listed | 241 mm (9.5 inches) |

| Release date | 2015-05-04 | 2012-09-05 |

| Production status | End-of-life | End-of-life |

The GTX 660 has a launch MSRP of 229 USD. The R9 M375 has no launch MSRP listed. Both are end-of-life products. The GTX 660 is a dual-slot desktop card requiring a 300 W PSU and one 6-pin connector, while the R9 M375 has no such requirements listed, consistent with its mobile positioning.