AMD Radeon R9 M390X vs NVIDIA GeForce RTX 3070 Mobile Comparison

The AMD Radeon R9 M390X and the NVIDIA GeForce RTX 3070 Mobile represent two distinct eras of mobile graphics, separated by nearly six years of architectural evolution. The data shows a clear generational gap: the RTX 3070 Mobile holds a single head-to-head benchmark win, while the R9 M390X, despite its age, remains competitive in aggregate scoring. This analysis breaks down their respective strengths, architectural differences, and specification gaps based solely on the provided benchmark data.

Where Each One Wins

The benchmark results indicate a stark division of labor. The NVIDIA GeForce RTX 3070 Mobile is the decisive winner in the only directly comparable compute test, the geekbench_vulkan, where it scores 86,768 against the R9 M390X’s 16,951. This represents an 80.5% advantage for the NVIDIA part, demonstrating overwhelming superiority in modern, low-level graphics API workloads. The RTX 3070 Mobile also benefits from a much broader benchmark suite, including a strong geekbench_opencl score of 92,939 and a passmark_g3d score of 15,309, suggesting robust performance across diverse compute and DirectX legacy scenarios.

Conversely, the AMD Radeon R9 M390X does not win any head-to-head benchmark in the provided data. However, its aggregate average benchmark score of 20,662 slightly edges out the RTX 3070 Mobile’s 20,534, a 0.6% difference. This is notable because the R9 M390X achieves this with only two benchmark results (Geekbench Metal and Vulkan), while the RTX 3070 Mobile’s average is pulled from ten tests. The R9 M390X’s geekbench_metal score of 24,373 is its strongest showing, indicating that on Apple’s Metal API, the older GCN architecture remains unexpectedly potent. In terms of general percentile ranking, the R9 M390X sits at the 66th percentile of all GPUs, one point higher than the RTX 3070 Mobile’s 65th percentile, reinforcing that its average performance is not to be dismissed despite its age.

Architecture Differences

The architectural gap between these two mobile GPUs is immense. The R9 M390X is built on AMD’s GCN 3.0 architecture, using the Amethyst chip, and is fabricated on a 28 nm process at TSMC. It packs 5,000 million transistors into a 366 mm² die, yielding a transistor density of 13.7M per mm². In contrast, the RTX 3070 Mobile uses NVIDIA’s Ampere architecture with the GA104 chip, manufactured by Samsung on an 8 nm node. This newer process allows for 17,400 million transistors in a slightly larger 392 mm² die, achieving a much higher density of 44.4M per mm². This density advantage is a primary driver of the performance disparity.

The compute core composition also diverges sharply. The R9 M390X features 2,048 shading units, 128 texture mapping units (TMUs), and 32 raster output units (ROPs). The RTX 3070 Mobile more than doubles the shader count to 5,120, alongside 160 TMUs and 80 ROPs. Critically, the RTX 3070 Mobile introduces dedicated hardware absent from the older AMD part: 40 RT cores for ray tracing and 160 tensor cores for AI-accelerated workloads. These features enable DirectX 12 Ultimate (12_2) support, whereas the R9 M390X is limited to DirectX 12 (12_0). The NVIDIA GPU also supports Vulkan 1.4, compared to the R9 M390X’s Vulkan 1.2.170. The R9 M390X’s GCN 3.0 architecture lacks any equivalent hardware acceleration for ray tracing or tensor operations, making the RTX 3070 Mobile fundamentally more capable for future-proofed gaming and compute tasks.

Head-to-Head Benchmarks

The only direct head-to-head benchmark available is geekbench_vulkan, and the results are decisive. The RTX 3070 Mobile scores 86,768, while the R9 M390X scores 16,951. The delta percentage of -80.5% indicates that the AMD part trails by more than four-fifths. This massive gap is not merely a clock speed difference; it reflects the architectural chasm between GCN 3.0 and Ampere, particularly in Vulkan’s low-overhead, multi-threaded nature which heavily favors newer hardware with higher shader counts and better driver optimization.

Beyond this single comparison, the broader benchmark suites tell a complementary story. The RTX 3070 Mobile’s geekbench_opencl score of 92,939 is over five times higher than its Vulkan score, suggesting exceptional general-purpose compute performance. Its passmark_g3d score of 15,309 indicates strong DirectX 11 gaming performance, while its passmark_gpu_compute score of 6,827 shows robust non-graphics throughput. The R9 M390X’s geekbench_metal score of 24,373 is respectable, but it is the only high point in its sparse benchmark profile. When comparing average scores, the R9 M390X leads by a hair (20,662 vs 20,534), but this is misleading given the RTX 3070 Mobile’s superior performance in every modern cross-platform workload where they overlap.

Specification Differences

The specification sheets reveal profound differences in nearly every measurable category. The RTX 3070 Mobile operates with a base clock of 1110 MHz and a boost clock of 1560 MHz, while the R9 M390X has no listed base or boost clocks. Memory configurations differ significantly: the RTX 3070 Mobile offers 8 GB of GDDR6 on a 256-bit bus, delivering 448.0 GB/s of bandwidth, whereas the R9 M390X has 4 GB of GDDR5 on a 256-bit bus, yielding only 160.0 GB/s. The memory clock is also higher on the NVIDIA part (1750 MHz vs 1250 MHz), translating to 14 Gbps effective versus 5 Gbps effective.

Compute throughput is where the gap widens most. The RTX 3070 Mobile delivers 15.97 TFLOPS of FP32 and FP16 performance (1:1 ratio), while the R9 M390X manages just 2.961 TFLOPS in both. Pixel and texture rates follow suit: the NVIDIA GPU achieves 124.8 GPixel/s and 249.6 GTexel/s, compared to the AMD’s 23.14 GPixel/s and 92.54 GTexel/s. The RTX 3070 Mobile’s TDP is 115 W versus the R9 M390X’s 75 W, reflecting the higher power envelope required for its performance. The bus interface also advances from PCIe 3.0 x16 to PCIe 4.0 x16. The R9 M390X is specified as a MXM Module with no power connectors, while the RTX 3070 Mobile lists no slot width or power connector details.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon R9 M390X has a slightly higher average benchmark score of 20,662, compared to the NVIDIA GeForce RTX 3070 Mobile’s 20,534. This puts the AMD part 0.6% ahead in this specific metric.

Q: What is the most significant performance difference in the head-to-head tests?

A: In the geekbench_vulkan test, the NVIDIA GeForce RTX 3070 Mobile scores 86,768, which is 80.5% higher than the AMD Radeon R9 M390X’s 16,951. This is the only direct head-to-head benchmark provided.

Q: Does the RTX 3070 Mobile support hardware ray tracing?

A: Yes, the NVIDIA GeForce RTX 3070 Mobile includes 40 dedicated RT cores for ray tracing acceleration. The AMD Radeon R9 M390X does not have any RT cores.

Q: How do the memory bandwidth figures compare?

A: The RTX 3070 Mobile has a memory bandwidth of 448.0 GB/s, while the R9 M390X offers 160.0 GB/s. The NVIDIA GPU also has double the memory capacity (8 GB vs 4 GB) and uses faster GDDR6 memory.

Q: What is the DirectX feature level difference?

A: The NVIDIA GeForce RTX 3070 Mobile supports DirectX 12 Ultimate (12_2), while the AMD Radeon R9 M390X is limited to DirectX 12 (12_0). This means the RTX 3070 Mobile can leverage newer rendering features like ray tracing and mesh shaders.

Q: Which GPU has a higher transistor density?

A: The NVIDIA GeForce RTX 3070 Mobile has a transistor density of 44.4M per mm², significantly higher than the AMD Radeon R9 M390X’s 13.7M per mm². This is due to the newer 8 nm process node versus the older 28 nm node.