AMD Radeon R7 M370 vs NVIDIA GeForce 945M Comparison

The data places the NVIDIA GeForce 945M and the AMD Radeon R7 M370 as direct contemporaries in the mobile discrete GPU space, both from the 2015 era and both now listed as end-of-life products. The recorded benchmark results show a clear, single-point performance gap between them, but the underlying architectural differences explain not just the score delta, but also where each part still holds relevance for specific workloads.

Head-to-Head Benchmarks

The only directly comparable benchmark in the database is Geekbench OpenCL, and the NVIDIA GeForce 945M wins that test outright. The 945M scores 8,099 points, while the Radeon R7 M370 scores 7,063 points. That is a delta of 14.7% in favor of the NVIDIA part. In practical terms, this means that for any OpenCL-accelerated compute task, the 945M delivers roughly one-seventh more throughput than the R7 M370. The R7 M370 does have an additional recorded Vulkan benchmark score of 6,465, but since there is no corresponding Vulkan score for the 945M in the database, that result cannot be used for a direct comparison; it only serves as an absolute reference point for the AMD part's API performance.

Looking at the broader context, the 945M's score places it at the 42nd percentile among all GPUs in the database. Its nearest rivals include the NVIDIA GRID K2 (8,080, delta 0.2%), the AMD Radeon R9 M360 (8,129, delta -0.4%), the NVIDIA GeForce GTX 650 Ti Boost (8,067, delta 0.4%), and the NVIDIA GeForce GTX 950M (8,135, delta -0.4%). This clustering is tight: the 945M is effectively within a fraction of a percent of all four of those cards, meaning its OpenCL performance is right at the boundary of that performance tier. The R7 M370, by contrast, sits at the 38th percentile, with a different set of nearest rivals: the AMD FirePro M5100 (6,830, delta -1%), the NVIDIA GeForce GT 1010 (6,698, delta 1%), the AMD Radeon R7 M460 (6,612, delta 2.3%), and the NVIDIA GeForce GTX 675M (6,946, delta -2.6%). The R7 M370 is ahead of the GT 1010 and R7 M460, but behind the FirePro M5100 and GTX 675M, which shows it is positioned in a lower performance band than the 945M.

The win count in the head-to-head data is 1 for the NVIDIA part and 0 for the AMD part, which matches the single OpenCL test available. No other shared benchmarks exist in the database, so the outcome is decisive but limited in scope.

Architecture Differences

The two GPUs come from fundamentally different design philosophies. The NVIDIA GeForce 945M uses the GM107 chip, built on the Maxwell architecture, manufactured by TSMC on a 28 nm process. The chip contains 1,870 million transistors on a die size of 148 mm², resulting in a transistor density of 12.6 million per square millimeter. The AMD Radeon R7 M370 uses the Litho chip, based on GCN 1.0 architecture, also on TSMC's 28 nm process. However, the AMD chip is significantly smaller and less complex: 950 million transistors on a 77 mm² die, with a density of 12.3 million per square millimeter. The 945M has nearly double the transistor count and almost double the die area, which gives it more hardware resources to work with.

That resource advantage shows directly in the compute unit counts. The 945M has 640 shading units, 40 texture mapping units (TMUs), and 16 raster output units (ROPs). The R7 M370 has 384 shading units, 24 TMUs, and only 8 ROPs. The NVIDIA part has 66% more shading units, 67% more TMUs, and double the ROPs. Clock speeds are comparable, with the 945M running at a base of 928 MHz and a boost of 1,020 MHz, while the R7 M370 runs at 875 MHz base and 960 MHz boost. The higher core count combined with the higher boost clock means the 945M's raw throughput scales well beyond the AMD part. The pixel rate tells the story: the 945M produces 16.32 GPixel/s versus the R7 M370's 7.680 GPixel/s. The texture rate is 40.80 GTexel/s versus 23.04 GTexel/s. The FP32 compute is 1,305.6 GFLOPS versus 737.3 GFLOPS. In every raw throughput metric, the 945M is roughly 70% to 80% ahead.

Memory is where the R7 M370 makes a partial comeback. Both cards have 2 GB of VRAM and a 128-bit bus width, but the type differs. The 945M uses DDR3 at 900 MHz, providing 1,800 Mbps effective speed and a bandwidth of 28.80 GB/s. The R7 M370 uses GDDR5 at 900 MHz, providing 3.6 Gbps effective speed and a bandwidth of 57.60 GB/s. That is exactly double the memory bandwidth for the AMD part. This is a significant advantage for bandwidth-sensitive workloads, such as high-resolution texture streaming or certain compute kernels that are memory-bound rather than compute-bound.

The feature sets also differ. The 945M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. The R7 M370 supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The AMD part has a slightly higher DirectX feature level, but the NVIDIA part has a much newer Vulkan version. Both are end-of-life products, so neither will see further driver optimizations, but the API support is what it is. The 945M is an MXM Module with an MXM-B (3.0) bus interface, while the R7 M370 uses PCIe 3.0 x8. The 945M has a recorded TDP of 75 W, while the R7 M370's TDP is not listed in the database.

Where Each One Wins

The NVIDIA GeForce 945M is the clear winner for compute-heavy tasks. Its OpenCL score is 14.7% higher, and that advantage comes from having more shading units, more TMUs, more ROPs, and higher clock speeds. For any workload that is shader-bound, such as GPGPU compute, rendering, or physics simulation, the 945M will deliver measurably better results. Its FP32 throughput of 1,305.6 GFLOPS is nearly double the R7 M370's 737.3 GFLOPS, so any floating-point math will favor NVIDIA. The 945M's higher pixel rate also makes it better for fill-rate-limited scenarios, which can include certain post-processing effects or lower-resolution rendering where the ROPs become the bottleneck.

The AMD Radeon R7 M370 wins the memory bandwidth battle. Its 57.60 GB/s of bandwidth is double the 945M's 28.80 GB/s. For textures, frame buffers, or compute workloads that access large data sets repeatedly, the R7 M370 can keep its fewer cores busier for longer. This does not show up in the Geekbench OpenCL score, where the 945M still wins by a wide margin, but in real-world scenarios like high-resolution texture loading or multi-sample anti-aliasing, the AMD part has a structural edge. The R7 M370 also has a slightly higher DirectX feature level (12_1 versus 12_0), which could matter for specific legacy titles that use those features, though the practical impact is small given the age of both parts.

The R7 M370 also has a recorded Vulkan score of 6,465, which is lower than its OpenCL score. Without a comparable Vulkan score for the 945M, it is impossible to say which part wins under Vulkan, but the absolute number suggests the AMD part is not unusually strong in that API. The 945M, by contrast, supports Vulkan 1.4, which is a much later revision, so for any modern Vulkan title that still runs on these old parts, the NVIDIA card is likely to have better driver-level compatibility.

The Verdict

From the recorded data, the NVIDIA GeForce 945M is the better all-around mobile GPU for compute performance. It wins the only direct benchmark by 14.7%, and its architectural superiority in shading units, TMUs, ROPs, and FP32 throughput is overwhelming. The 42nd percentile ranking versus the R7 M370's 38th percentile confirms that the 945M sits in a higher performance tier across the entire database, not just in this head-to-head. The nearest rivals for the 945M are all within 0.4% of its score, meaning it is competitive with cards like the GTX 950M and R9 M360, while the R7 M370's rivals are a tier lower.

The R7 M370 should only be chosen if memory bandwidth is the absolute priority and compute throughput is secondary. Its GDDR5 memory at 57.60 GB/s is a real advantage, and for specific workloads that are purely bandwidth-bound, it could edge out the 945M. However, the database contains no benchmark that shows such a win. The only recorded head-to-head result favors NVIDIA. The 945M also has a lower TDP at 75 W, which is relevant for thin-and-light laptops, though the R7 M370's power draw is unlisted.

For a user deciding between these two in a used or refurbished laptop, the data says pick the NVIDIA part for general compute, gaming, and any OpenCL workload. Pick the AMD part only if the system configuration requires it, or if the specific application is known to be memory-bandwidth-limited and does not rely on shader throughput.

FAQ

Q: Which GPU wins in the Geekbench OpenCL benchmark?

A: The NVIDIA GeForce 945M scores 8,099 points, while the AMD Radeon R7 M370 scores 7,063 points. The 945M wins by a delta of 14.7%.

Q: How does the memory bandwidth compare between the two?

A: The AMD Radeon R7 M370 has 57.60 GB/s of bandwidth from its GDDR5 memory, while the NVIDIA GeForce 945M has 28.80 GB/s from its DDR3 memory. The AMD part has exactly double the bandwidth.

Q: What is the performance percentile of each GPU in the database?

A: The NVIDIA GeForce 945M sits at the 42nd percentile of all GPUs, while the AMD Radeon R7 M370 sits at the 38th percentile.

Q: Which GPU has more shading units?

A: The NVIDIA GeForce 945M has 640 shading units, while the AMD Radeon R7 M370 has 384 shading units. The NVIDIA part has 66% more.

Q: What are the closest rivals to the NVIDIA GeForce 945M?

A: The nearest rivals are the NVIDIA GRID K2 (8,080, delta 0.2%), the AMD Radeon R9 M360 (8,129, delta -0.4%), the NVIDIA GeForce GTX 650 Ti Boost (8,067, delta 0.4%), and the NVIDIA GeForce GTX 950M (8,135, delta -0.4%). All are within 0.4% of the 945M's score.

Q: Does the AMD Radeon R7 M370 have any benchmark results not recorded for the NVIDIA part?

A: Yes, the R7 M370 has a Geekbench Vulkan score of 6,465. The 945M has no Vulkan score in the database, so a direct comparison in that API is not possible.

Specification Differences

| Specification | NVIDIA GeForce 945M | AMD Radeon R7 M370 |

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

| Chip | GM107 | Litho |

| Architecture | Maxwell | GCN 1.0 |

| Process Node | 28 nm | 28 nm |

| Transistors | 1,870 million | 950 million |

| Die Size | 148 mm² | 77 mm² |

| Transistor Density | 12.6M / mm² | 12.3M / mm² |

| Base Clock | 928 MHz | 875 MHz |

| Boost Clock | 1020 MHz | 960 MHz |

| Memory Clock | 900 MHz, 1800 Mbps effective | 900 MHz, 3.6 Gbps effective |

| Memory Type | DDR3 | GDDR5 |

| Memory Bandwidth | 28.80 GB/s | 57.60 GB/s |

| Shading Units | 640 | 384 |

| TMUs | 40 | 24 |

| ROPs | 16 | 8 |

| Pixel Rate | 16.32 GPixel/s | 7.680 GPixel/s |

| Texture Rate | 40.80 GTexel/s | 23.04 GTexel/s |

| FP32 | 1,305.6 GFLOPS | 737.3 GFLOPS |

| TDP | 75 W | Not listed |

| Bus Interface | MXM-B (3.0) | PCIe 3.0 x8 |

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

| Vulkan | 1.4 | 1.2.170 |

| Release Date | 2015-10-26 | 2015-05-04 |