AMD Radeon R9 M390X vs NVIDIA GeForce RTX 2070 Comparison

The Verdict

The data presents a stark generational and architectural divide. The NVIDIA GeForce RTX 2070 is the clear performance victor, decisively winning the only shared benchmark. The AMD Radeon R9 M390X, while competitive in its own era, is outclassed in raw compute and modern feature support. The RTX 2070 is for users who need maximum performance and access to contemporary rendering features. The R9 M390X, with its single benchmark win and lower scores, is a legacy mobile part that should only be considered for specific, older system compatibility. The RTX 2070 sits at the 63rd percentile of all GPUs, while the R9 M390X is at the 66th, but this is misleading; the newer NVIDIA card's average score is dragged down by the inclusion of many lower-end synthetic tests, while the AMD part's average is based on only two modern API tests.

Benchmark results indicate that for any workload leveraging Vulkan, the RTX 2070 is in a different league. The R9 M390X's 16,951 in Geekbench Vulkan is dwarfed by the RTX 2070's 82,521, a massive 79.5% deficit for the AMD card. This is not a close contest; it is a generational leap. The RTX 2070 is the only choice for anyone prioritizing current application performance or future-proofing with hardware-accelerated ray tracing and DLSS, as implied by its Turing architecture's dedicated cores. The R9 M390X, conversely, holds no advantage in any measured metric. Its architecture is end-of-life, its process node is older, and its feature set is limited to DirectX 12_0. For a user with a legacy laptop that requires an MXM module and who is locked into older software, the R9 M390X might be a functional, if unremarkable, option.

Architecture Differences

The two GPUs are separated by three years of silicon evolution and radically different design philosophies. The AMD Radeon R9 M390X is built on the Graphics Core Next (GCN) 3.0 architecture, fabricated on a 28 nm process at TSMC. Its chip, codenamed Amethyst, contains 5,000 million transistors on a 366 mm² die. This results in a transistor density of 13.7 million per square millimeter. In contrast, the NVIDIA GeForce RTX 2070 uses the Turing architecture, built on a more advanced 12 nm process, also at TSMC. Its TU106 chip packs 10,800 million transistors—more than double the AMD part—onto a larger 445 mm² die, achieving a density of 24.3 million per square millimeter. This process advantage is a primary driver of the NVIDIA card's superior performance and efficiency.

The core configurations differ substantially. The R9 M390X has 2,048 shading units, 128 texture mapping units (TMUs), and 32 render output units (ROPs). The RTX 2070 increases these counts to 2,304 shading units, 144 TMUs, and 64 ROPs. More critically, the RTX 2070 introduces dedicated hardware that the GCN part lacks entirely: 36 RT cores for ray tracing and 288 tensor cores for AI-accelerated tasks. This is the fundamental architectural shift. The R9 M390X is a pure rasterization engine, while the RTX 2070 is a heterogeneous processor designed for a wider range of workloads.

Memory subsystems also diverge. The R9 M390X uses 4 GB of GDDR5 on a 256-bit bus, yielding 160.0 GB/s of bandwidth. The RTX 2070 doubles capacity to 8 GB of faster GDDR6 on the same 256-bit bus, more than doubling bandwidth to 448.0 GB/s. The NVIDIA card supports DirectX 12 Ultimate (12_2), while the AMD card is limited to DirectX 12 (12_0). Both support OpenGL 4.6 and Vulkan, but the NVIDIA card's Vulkan implementation is version 1.4 versus 1.2.170 for the AMD part. The RTX 2070 also supports fp16 compute at a 2:1 rate (14.93 TFLOPS) versus the R9 M390X's 1:1 fp16 rate (2.961 TFLOPS), which is a significant advantage for AI and compute workloads.

Where Each One Wins

There is only one direct comparison point, and it is a landslide. The NVIDIA GeForce RTX 2070 wins the Geekbench Vulkan test by a 79.5% margin. This means that in any Vulkan-based game or compute application, the RTX 2070 will be dramatically faster. The data shows no scenario where the AMD Radeon R9 M390X wins. However, the context of their respective eras matters. The R9 M390X, released in May 2015, was designed for high-end gaming laptops of that period. Its 66th percentile rank, based on its two benchmark scores, suggests it was a capable mobile part in its time. The RTX 2070, released in October 2018, is a desktop card with a 63rd percentile rank, but its average score is a composite of many more tests, including older DirectX 9 and 10 benchmarks where it scores lower, which pulls its average down.

For a user with a 2015-era laptop, the R9 M390X's 4 GB of GDDR5 memory and 160.0 GB/s bandwidth were adequate for the games of that time. Its 2.961 TFLOPS of fp32 performance was respectable. In a modern context, however, the RTX 2070's advantages are overwhelming: 7.465 TFLOPS of fp32, 14.93 TFLOPS of fp16, 8 GB of VRAM, and the dedicated RT and tensor cores. The RTX 2070 is the winner for any current or future application. The R9 M390X's only potential "win" is in a legacy system where its MXM module form factor is required for a laptop upgrade and where the software does not benefit from Turing's features. The benchmark data, however, provides no evidence of any performance scenario where the AMD part is preferable.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon R9 M390X has a higher average benchmark score of 20,662 compared to the NVIDIA GeForce RTX 2070's 18,789. This is misleading, as the RTX 2070's average includes many older DirectX 9/10/11 tests where it scores low, while the R9 M390X's average is based only on Geekbench Metal and Vulkan scores.

Q: How much faster is the RTX 2070 in the shared Vulkan benchmark?

A: The RTX 2070 scores 82,521 in Geekbench Vulkan, which is 79.5% higher than the R9 M390X's score of 16,951. This represents a dominant performance advantage for the NVIDIA card.

Q: What are the key architectural differences?

A: The R9 M390X uses the older GCN 3.0 architecture on a 28 nm process, while the RTX 2070 uses Turing on a 12 nm process. The RTX 2070 also has dedicated RT cores for ray tracing and tensor cores for AI, which the AMD card lacks entirely.

Q: Which GPU has more memory bandwidth?

A: The NVIDIA GeForce RTX 2070 has significantly more, with 448.0 GB/s of bandwidth from its GDDR6 memory, compared to 160.0 GB/s from the R9 M390X's GDDR5 memory.

Q: Does the R9 M390X support hardware ray tracing?

A: No. The R9 M390X has no RT cores, and its DirectX 12 support is limited to the 12_0 feature level. The RTX 2070 has 36 RT cores and supports DirectX 12 Ultimate (12_2), which is required for hardware-accelerated ray tracing in DirectX titles.

Q: What is the form factor difference?

A: The R9 M390X is an MXM Module, designed for laptops. The RTX 2070 is a Dual-slot desktop card measuring 229 mm in length, 113 mm in height, and 35 mm in width, requiring a 450 W suggested PSU and a single 8-pin power connector.

Head-to-Head Benchmarks

The single head-to-head benchmark result is Geekbench Vulkan, and it tells a story of absolute dominance. The NVIDIA GeForce RTX 2070 scores 82,521, while the AMD Radeon R9 M390X scores 16,951. The deltaPct is -79.5% from the NVIDIA card's perspective, meaning the R9 M390X is nearly five times slower in this test. This is not a marginal win; it is a complete rout. The RTX 2070's 82,521 score reflects its superior shading units, higher clock speeds, massive memory bandwidth advantage (448.0 GB/s vs 160.0 GB/s), and the efficiency of the Turing architecture in modern APIs.

The data shows no other direct comparisons, so we must rely on the broader benchmark suites. The RTX 2070's Passmark G3D score is 16,094, while its Geekbench OpenCL score is 79,966. The R9 M390X's only other benchmark is Geekbench Metal, where it scores 24,373. While these are different tests, the pattern is clear: in modern compute-heavy APIs, the RTX 2070 is overwhelmingly faster. The R9 M390X's 2.961 TFLOPS of fp32 performance is less than half of the RTX 2070's 7.465 TFLOPS. In terms of texture and pixel rates, the RTX 2070 achieves 233.3 GTexel/s and 103.7 GPixel/s, respectively, versus the R9 M390X's 92.54 GTexel/s and 23.14 GPixel/s. Every measurable performance metric in the FACT PACK favors the RTX 2070 by a wide margin.

Specification Differences

The two GPUs differ across nearly every core specification. The most significant differences are in the compute and memory subsystems.

• Process Node: The R9 M390X is on a 28 nm process, while the RTX 2070 is on a 12 nm process.
• Transistors: The R9 M390X has 5,000 million, while the RTX 2070 has 10,800 million.
• Die Size: The R9 M390X is 366 mm², while the RTX 2070 is 445 mm².
• Transistor Density: The R9 M390X is 13.7M / mm², while the RTX 2070 is 24.3M / mm².
• Base Clock: The R9 M390X has none listed, while the RTX 2070 has 1410 MHz.
• Boost Clock: The R9 M390X has none listed, while the RTX 2070 has 1620 MHz.
• Memory Size: The R9 M390X has 4 GB, while the RTX 2070 has 8 GB.
• Memory Type: The R9 M390X uses GDDR5, while the RTX 2070 uses GDDR6.
• Memory Clock: The R9 M390X is 1250 MHz (5 Gbps effective), while the RTX 2070 is 1750 MHz (14 Gbps effective).
• Memory Bandwidth: The R9 M390X is 160.0 GB/s, while the RTX 2070 is 448.0 GB/s.
• Shading Units: The R9 M390X has 2,048, while the RTX 2070 has 2,304.
• TMUs: The R9 M390X has 128, while the RTX 2070 has 144.
• ROPs: The R9 M390X has 32, while the RTX 2070 has 64.
• RT Cores: The R9 M390X has none, while the RTX 2070 has 36.
• Tensor Cores: The R9 M390X has none, while the RTX 2070 has 288.
• Pixel Rate: The R9 M390X is 23.14 GPixel/s, while the RTX 2070 is 103.7 GPixel/s.
• Texture Rate: The R9 M390X is 92.54 GTexel/s, while the RTX 2070 is 233.3 GTexel/s.
• FP32 Performance: The R9 M390X is 2.961 TFLOPS, while the RTX 2070 is 7.465 TFLOPS.
• FP16 Performance: The R9 M390X is 2.961 TFLOPS (1:1), while the RTX 2070 is 14.93 TFLOPS (2:1).
• TDP: The R9 M390X is 75 W, while the RTX 2070 is 175 W.
• Slot Width: The R9 M390X is an MXM Module, while the RTX 2070 is Dual-slot.
• Power Connectors: The R9 M390X has none, while the RTX 2070 has 1x 8-pin.
• Suggested PSU: The R9 M390X has none listed, while the RTX 2070 is 450 W.
• DirectX Support: The R9 M390X supports 12 (12_0), while the RTX 2070 supports 12 Ultimate (12_2).
• Vulkan Support: The R9 M390X is 1.2.170, while the RTX 2070 is 1.4.
• Dimensions: The R9 M390X has no listed dimensions, while the RTX 2070 is 229 mm x 113 mm x 35 mm.
• Display Outputs: The R9 M390X is portable device dependent, while the RTX 2070 has 1x DVI, 1x HDMI 2.0, 2x DisplayPort 1.4a, and 1x USB Type-C.
• Release Date: The R9 M390X was released on May 4, 2015, while the RTX 2070 was released on October 16, 2018.