AMD Radeon R7 M360 vs NVIDIA GeForce GTX 670MX Comparison

The Verdict

The data separates these two mobile GPUs clearly. The NVIDIA GeForce GTX 670MX is the stronger part overall, with an average benchmark score of 5721 compared to 4931 for the AMD Radeon R7 M360, a gap of roughly 16%. The GTX 670MX sits at the 33rd percentile of all GPUs in the database, while the R7 M360 lands at the 29th percentile. If you want the faster mobile graphics solution, the GTX 670MX is the pick.

However, the R7 M360 is not without its own advantages. It is a newer design from 2015, whereas the GTX 670MX dates back to late 2012. It uses a newer architecture, GCN 3.0 versus Kepler, and it supports DirectX 12 (12_0) at a higher feature level than the GTX 670MX's DirectX 12 (11_0) implementation. For software that leverages newer API features, the R7 M360 has a specification-level edge, even though it loses in raw compute benchmarks.

For a practical recommendation: choose the GTX 670MX if you prioritize raw performance in general-purpose compute and Vulkan workloads. The benchmark data shows it wins both recorded head-to-head tests. Choose the R7 M360 if you need a more modern API feature set, lower memory footprint, or if the workload is specifically tuned for GCN 3.0 features. The R7 M360 also offers 1:1 FP16 to FP32 throughput, which the GTX 670MX does not list, making it potentially more relevant for certain compute tasks.

Architecture Differences

The two GPUs come from different architectural lineages. The NVIDIA GeForce GTX 670MX is built on Kepler, using the GK104 chip manufactured by TSMC on a 28 nm process. It packs 3,540 million transistors on a 294 mm² die, with a transistor density of 12.0 million per mm². The AMD Radeon R7 M360 uses GCN 3.0, with the Meso chip, also on TSMC's 28 nm node. Its die is much smaller at 125 mm², housing 1,550 million transistors, giving it a slightly higher density of 12.4 million per mm².

The compute configurations differ substantially. The GTX 670MX has 960 shading units, 80 texture mapping units, and 24 raster operation pipelines. The R7 M360 has 384 shading units, 24 TMUs, and 8 ROPs. This means the NVIDIA part has more than double the shading units and more than triple the TMUs, which explains its higher theoretical fill rates. The GTX 670MX posts 12.02 GPixel/s pixel rate and 48.08 GTexel/s texture rate, while the R7 M360 manages 9.00 GPixel/s and 27.00 GTexel/s respectively.

Memory architecture is also a key differentiator. The GTX 670MX uses 3 GB of GDDR5 on a 192-bit bus, delivering 67.20 GB/s of bandwidth. The R7 M360 comes with 2 GB of DDR3 on a much narrower 64-bit bus, yielding only 14.40 GB/s. That is a massive bandwidth advantage for the NVIDIA part, nearly 4.7 times higher. The R7 M360 compensates with a higher base clock of 1100 MHz (boost 1125 MHz) versus the GTX 670MX's fixed 601 MHz, but the memory bandwidth deficit is too large to overcome in most scenarios.

API support differs in one notable way. The R7 M360 supports DirectX 12 (12_0), while the GTX 670MX only reaches DirectX 12 (11_0). Both support OpenGL 4.6, but Vulkan versions differ slightly: the GTX 670MX lists Vulkan 1.2.175, and the R7 M360 lists Vulkan 1.2.170. The R7 M360 also explicitly lists FP16 performance at 864.0 GFLOPS (1:1 ratio), while the GTX 670MX does not list FP16 capabilities. The bus interface differs as well: PCIe 3.0 x16 for NVIDIA versus PCIe 3.0 x8 for AMD.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 670MX is ahead, with an average benchmark score of 5721 versus 4931 for the AMD Radeon R7 M360. That is a lead of about 16% for the NVIDIA part.

Q: How do they compare in memory bandwidth?

A: The GTX 670MX has 67.20 GB/s of bandwidth from 3 GB of GDDR5 on a 192-bit bus. The R7 M360 has only 14.40 GB/s from 2 GB of DDR3 on a 64-bit bus. The NVIDIA part offers roughly 4.7 times the bandwidth.

Q: Does the newer AMD GPU support a higher DirectX feature level?

A: Yes. The R7 M360 supports DirectX 12 (12_0), while the GTX 670MX supports DirectX 12 (11_0). This means the AMD part can handle newer API feature levels despite its lower raw compute performance.

Q: Which GPU has more shading units?

A: The NVIDIA GeForce GTX 670MX has 960 shading units. The AMD Radeon R7 M360 has 384 shading units. The NVIDIA part has 2.5 times as many.

Q: What is the transistor count difference?

A: The GTX 670MX contains 3,540 million transistors on a 294 mm² die. The R7 M360 contains 1,550 million transistors on a 125 mm² die. The NVIDIA chip is larger and has more than double the transistors.

Q: Which GPU wins in Vulkan performance?

A: The GTX 670MX wins the Vulkan benchmark with a score of 5316, compared to 5223 for the R7 M360. The margin is narrow at 1.8%, but the NVIDIA part still comes out ahead.

Specification Differences

The two GPUs differ in nearly every key specification category. The GTX 670MX uses the GK104 chip with Kepler architecture, while the R7 M360 uses the Meso chip with GCN 3.0. The GTX 670MX belongs to the GeForce 600M generation; the R7 M360 belongs to the Gem System (R7 M300) generation. Both are 28 nm parts from TSMC, but the GTX 670MX has a larger die at 294 mm² versus 125 mm², and more transistors at 3,540 million versus 1,550 million.

Clock speeds favor the AMD part. The R7 M360 runs at 1100 MHz base and 1125 MHz boost, while the GTX 670MX is locked at 601 MHz for both base and boost. Memory clocks also differ: the GTX 670MX runs at 700 MHz with 2.8 Gbps effective, while the R7 M360 runs at 900 MHz with 1800 Mbps effective. However, the memory bus width and type reverse the practical outcome: the GTX 670MX has a 192-bit GDDR5 interface, while the R7 M360 has a 64-bit DDR3 interface.

The compute unit counts are markedly different. The GTX 670MX has 960 shading units, 80 TMUs, and 24 ROPs. The R7 M360 has 384 shading units, 24 TMUs, and 8 ROPs. Pixel rate is 12.02 GPixel/s for NVIDIA versus 9.00 GPixel/s for AMD. Texture rate is 48.08 GTexel/s versus 27.00 GTexel/s. FP32 performance is 1,153.9 GFLOPS for the GTX 670MX versus 864.0 GFLOPS for the R7 M360. The AMD part lists FP16 at 864.0 GFLOPS with a 1:1 ratio, which the NVIDIA part does not list.

Power and interface details also differ. The GTX 670MX has a TDP of 75 W and uses no power connectors. The R7 M360 does not list a TDP or power connector requirement. The GTX 670MX uses PCIe 3.0 x16, while the R7 M360 uses PCIe 3.0 x8. Display outputs are listed as portable device dependent for the GTX 670MX, while the R7 M360 does not list display outputs. The GTX 670MX was released on 2012-09-30, and the R7 M360 on 2015-05-04. Both are end-of-life products.

Head-to-Head Benchmarks

The database records two head-to-head benchmark comparisons between these GPUs, and the GTX 670MX wins both. In the Geekbench OpenCL test, the GTX 670MX scores 6125, while the R7 M360 scores 4638. That is a 32.1% advantage for the NVIDIA part. This is the largest margin between the two and reflects the GTX 670MX's superior shading unit count, texture rate, and memory bandwidth. For compute-heavy workloads such as OpenCL acceleration, the data shows a decisive win for the older NVIDIA chip.

In the Geekbench Vulkan test, the margin narrows dramatically. The GTX 670MX scores 5316, and the R7 M360 scores 5223, a delta of only 1.8%. This suggests that the R7 M360's newer GCN 3.0 architecture and higher clock speeds close much of the gap in Vulkan workloads. The AMD part's 1:1 FP16 capability may also contribute to its relatively stronger showing here. Still, the GTX 670MX holds the lead in this test as well.

Overall, the GTX 670MX records 2 wins and the R7 M360 records 0 wins in head-to-head benchmarks. The average benchmark score of 5721 for the GTX 670MX versus 4931 for the R7 M360 reinforces that pattern. The nearest rivals for the GTX 670MX include the Intel Iris Pro Graphics P6300 at 5712 (0.2% delta), the NVIDIA GeForce GTX 550 Ti at 5731 (-0.2% delta), and the AMD Radeon HD 8790M at 5691 (0.5% delta). The R7 M360's nearest rivals include the AMD Radeon R7 M265 at 4929 (0% delta) and the AMD FirePro W5130M at 4904 (0.6% delta). This places the two GPUs in different performance tiers, with the GTX 670MX firmly in the higher one.

Where Each One Wins

The GTX 670MX wins in raw compute performance. Its OpenCL score of 6125 is 32.1% higher than the R7 M360's 4638. This makes it the clear choice for applications that offload heavy math to the GPU, such as rendering, video encoding, or scientific compute. Its 1,153.9 GFLOPS FP32 throughput versus 864.0 GFLOPS gives it a 33.5% theoretical advantage in single-precision compute. The 67.20 GB/s memory bandwidth versus 14.40 GB/s also makes it far better suited for tasks that stream large datasets, such as texture-heavy workloads or large frame buffers.

The R7 M360 wins in API modernity. It supports DirectX 12 (12_0), which the GTX 670MX cannot match with its 11_0 feature level. This means games or applications that require 12_0 features will only run properly on the AMD part. The R7 M360 also lists FP16 performance at 864.0 GFLOPS with a 1:1 ratio, which the GTX 670MX does not list at all. For workloads that use FP16 arithmetic, such as certain machine learning inference or image processing pipelines, the R7 M360 may be the better option despite its lower FP32 output.

The R7 M360 also holds an advantage in clock speed, with a base of 1100 MHz and boost of 1125 MHz versus the GTX 670MX's fixed 601 MHz. This does not translate into overall performance wins in the recorded benchmarks, but it does suggest the AMD part may scale better in clock-sensitive workloads. For casual mobile use, the R7 M360's smaller die (125 mm²) and lower transistor count (1,550 million) could imply lower power draw, though the database does not list a TDP for this part. The GTX 670MX lists 75 W TDP, so the R7 M360 may be a more efficient choice for thin-and-light laptops, but that is not confirmed by the data.

In practical terms: choose the GTX 670MX for compute-heavy tasks, high-resolution textures, or any workload that benefits from memory bandwidth and shading unit count. Choose the R7 M360 for DirectX 12 (12_0) compatibility, FP16 compute, or if you need a more modern architecture with a smaller die footprint. The benchmark data favors the NVIDIA part in every recorded test, but the AMD part wins on specification features that matter for specific software environments.