AMD Radeon R9 M390X vs AMD Radeon RX 460 Comparison

Head-to-Head Benchmarks

The benchmark data presents a split decision between these two AMD mobile and desktop GPUs. In the Geekbench Metal test, the AMD Radeon R9 M390X delivers a score of 24,373, while the AMD Radeon RX 460 scores 17,065. This gives the R9 M390X a commanding 42.8% lead in this specific API workload. That margin is substantial, indicating a clear advantage for the older, larger chip when running Metal-based applications.

However, the situation reverses in the Geekbench Vulkan test. Here, the AMD Radeon RX 460 scores 20,198, outperforming the R9 M390X's score of 16,951. The delta of -16.1% means the RX 460 is 16.1% ahead of the R9 M390X in this test. This is a significant swing, showing that the newer architecture has a distinct edge in Vulkan performance despite its smaller size and lower raw specifications.

Looking at the average benchmark score across all recorded tests, the R9 M390X averages 20,662 points, placing it in the 66th percentile of all GPUs in the database. The RX 460 averages 18,373 points, which lands it in the 62nd percentile. The R9 M390X's nearest rival in the database is the AMD Radeon RX 5600 XT with an average score of 20,713, a delta of only -0.2%, meaning the R9 M390X is essentially neck-and-neck with that much newer card. On the other side, the RX 460 sits close to the Intel Arc A770M, which averages 18,383 points, a delta of -0.1%.

The head-to-head results show one win for each GPU. The R9 M390X wins the Metal test decisively, while the RX 460 wins the Vulkan test by a solid margin. For users prioritizing Metal performance, the data clearly favors the R9 M390X. For those working in Vulkan environments, the RX 460 is the stronger option. The average scores suggest the R9 M390X holds a slight overall edge, but the specific API used matters greatly.

Architecture Differences

The two GPUs come from different generations of AMD's graphics technology. The R9 M390X uses the Amethyst chip built on the GCN 3.0 architecture, manufactured on a 28 nm process at TSMC. It packs 5,000 million transistors onto a die size of 366 mm², giving it a transistor density of 13.7 million transistors per square millimeter. The RX 460 uses the Baffin chip based on the newer GCN 4.0 architecture, built on a 14 nm process at GlobalFoundries. It contains 3,000 million transistors on a much smaller 123 mm² die, achieving a higher transistor density of 24.4 million transistors per square millimeter.

These architectural differences directly influence performance characteristics. The R9 M390X has 2,048 shading units, 128 texture mapping units, and 32 render output units. The RX 460 has 896 shading units, 56 texture mapping units, and 16 render output units. In terms of raw throughput, the R9 M390X reaches 2.961 TFLOPS for FP32 and FP16 computations, while the RX 460 achieves 2.150 TFLOPS for both. Pixel rate for the R9 M390X is 23.14 GPixel/s, compared to 19.20 GPixel/s for the RX 460. Texture rate stands at 92.54 GTexel/s for the R9 M390X versus 67.20 GTexel/s for the RX 460.

The memory subsystems also differ significantly. The R9 M390X comes with 4 GB of GDDR5 memory on a 256-bit bus, delivering 160.0 GB/s of bandwidth. The RX 460 has 2 GB of GDDR5 memory on a 128-bit bus, offering 112.0 GB/s of bandwidth. The R9 M390X's memory clock is listed at 1250 MHz, which translates to 5 Gbps effective, while the RX 460 runs its memory at 1750 MHz, or 7 Gbps effective. Despite the faster per-pin memory speed on the RX 460, the wider bus on the R9 M390X provides higher total bandwidth.

In terms of API support, both GPUs support DirectX 12 (12_0) and OpenGL 4.6. The R9 M390X supports Vulkan 1.2.170, while the RX 460 supports the newer Vulkan 1.3 specification. This newer Vulkan version on the RX 460 likely contributes to its superior Vulkan benchmark score, as the architecture and driver stack are more optimized for the modern API. The RX 460 also has a base clock of 1090 MHz and a boost clock of 1200 MHz, whereas the R9 M390X's base and boost clocks are not recorded in the database.

The Verdict

The data paints a clear picture for different use cases. The AMD Radeon R9 M390X is the choice for Metal-centric workloads, where it outperforms the RX 460 by 42.8%. Its larger memory footprint of 4 GB and higher bandwidth of 160.0 GB/s also make it better suited for handling larger textures and higher resolutions. The R9 M390X's average benchmark score of 20,662 places it at the 66th percentile of all GPUs, slightly ahead of the RX 460's 62nd percentile ranking.

Conversely, the AMD Radeon RX 460 is the better option for Vulkan-based applications, leading the R9 M390X by 16.1% in that specific test. Its newer GCN 4.0 architecture and support for Vulkan 1.3 give it a modern advantage in newer game engines and compute workloads that leverage the Vulkan API. The RX 460 also has a higher boost clock of 1200 MHz compared to the R9 M390X's unlisted clock speeds.

For users who need maximum performance in Metal environments, such as those in the Apple ecosystem or using Metal-based creative software, the R9 M390X is the data-supported choice. For gamers or professionals working with Vulkan-based titles or applications, the RX 460 offers better measured performance. The overall average scores are close, with the R9 M390X holding a 2,289-point advantage, but the API-specific results show that neither card dominates universally.

Specification Differences

The two GPUs differ across several key specification fields. The R9 M390X uses the Amethyst chip on GCN 3.0 architecture, while the RX 460 uses the Baffin chip on GCN 4.0. The process node differs: 28 nm at TSMC for the R9 M390X versus 14 nm at GlobalFoundries for the RX 460. Transistor count is 5,000 million for the R9 M390X against 3,000 million for the RX 460. Die size is 366 mm² for the R9 M390X and 123 mm² for the RX 460, with transistor densities of 13.7M per mm² and 24.4M per mm² respectively.

Memory specifications vary considerably. The R9 M390X has 4 GB of GDDR5 on a 256-bit bus with 160.0 GB/s bandwidth and a memory clock of 1250 MHz (5 Gbps effective). The RX 460 has 2 GB of GDDR5 on a 128-bit bus with 112.0 GB/s bandwidth and a memory clock of 1750 MHz (7 Gbps effective). Shading units number 2,048 for the R9 M390X versus 896 for the RX 460. Texture mapping units are 128 against 56, and render output units are 32 versus 16.

Performance rates differ as well. Pixel rate is 23.14 GPixel/s for the R9 M390X and 19.20 GPixel/s for the RX 460. Texture rate is 92.54 GTexel/s for the R9 M390X and 67.20 GTexel/s for the RX 460. FP32 and FP16 compute are 2.961 TFLOPS for the R9 M390X and 2.150 TFLOPS for the RX 460. Both have a TDP of 75 W, but the R9 M390X uses an MXM Module slot width while the RX 460 is a dual-slot card measuring 170 mm (6.7 inches) in length. The R9 M390X has no power connectors and a PCIe 3.0 x16 interface, while the RX 460 also has no power connectors but uses PCIe 3.0 x8 and lists a suggested PSU of 250 W.

Display outputs differ: the R9 M390X is portable device dependent, while the RX 460 offers 1x DVI, 1x HDMI 2.0b, and 1x DisplayPort 1.4a. Vulkan API versions differ, with the R9 M390X at 1.2.170 and the RX 460 at 1.3. Release dates also diverge, with the R9 M390X released on 2015-05-04 and the RX 460 on 2016-08-07.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon R9 M390X has an average benchmark score of 20,662, which is higher than the AMD Radeon RX 460's average of 18,373.

Q: How much faster is the R9 M390X in the Metal benchmark?

A: The R9 M390X scores 24,373 in Geekbench Metal, which is 42.8% higher than the RX 460's score of 17,065.

Q: Does the RX 460 win any benchmark against the R9 M390X?

A: Yes, the RX 460 wins the Geekbench Vulkan test with a score of 20,198, which is 16.1% higher than the R9 M390X's score of 16,951.

Q: What is the memory bandwidth difference between these two GPUs?

A: The R9 M390X has 160.0 GB/s of memory bandwidth from its 256-bit bus, while the RX 460 has 112.0 GB/s from its 128-bit bus.

Q: Which GPU supports a newer Vulkan version?

A: The AMD Radeon RX 460 supports Vulkan 1.3, while the AMD Radeon R9 M390X supports Vulkan 1.2.170.

Q: What are the shading unit counts for each GPU?

A: The R9 M390X has 2,048 shading units, while the RX 460 has 896 shading units.

Where Each One Wins

The AMD Radeon R9 M390X wins in scenarios that rely on Metal API performance. Its 42.8% advantage in the Geekbench Metal test makes it the superior choice for applications that use Apple's Metal graphics framework. The larger 4 GB memory capacity and 160.0 GB/s bandwidth also give it an edge in workload-heavy tasks such as high-resolution texture rendering or multi-monitor setups. Its higher pixel rate of 23.14 GPixel/s and texture rate of 92.54 GTexel/s provide more raw throughput for fill-rate-bound operations. The R9 M390X also leads in compute performance with 2.961 TFLOPS of FP32 and FP16, making it stronger for general-purpose GPU compute tasks that are not tied to a specific API.

The AMD Radeon RX 460 wins in Vulkan-based environments. Its 16.1% lead in the Geekbench Vulkan test, combined with support for Vulkan 1.3, makes it the better choice for modern games and applications that use this API. The newer GCN 4.0 architecture and higher boost clock of 1200 MHz contribute to its efficiency in these workloads. The RX 460 also has a higher memory clock of 1750 MHz (7 Gbps effective), which helps in bandwidth-sensitive tasks despite its narrower 128-bit bus. Its smaller die size of 123 mm² and higher transistor density of 24.4M per mm² indicate a more modern, efficient design. The RX 460 also benefits from standard display outputs including DVI, HDMI 2.0b, and DisplayPort 1.4a, making it more versatile for desktop users. For Vulkan gaming or compute tasks, the RX 460 is the data-supported winner.