AMD Radeon R9 M390X vs AMD Radeon RX 6700S Comparison
AMD Radeon R9 M390X
Radeon RX 6700S
PERFORMANCE BENCHMARKS
Analysis: AMD Radeon R9 M390X vs AMD Radeon RX 6700S
# AMD Radeon RX 6700S vs AMD Radeon R9 M390X
The AMD Radeon RX 6700S and AMD Radeon R9 M390X represent two distinct eras of mobile graphics, separated by nearly seven years of architectural evolution. The RX 6700S, built on RDNA 2.0 at 7 nm, delivers benchmark results that dwarf the older GCN 3.0-based R9 M390X. In the two shared tests, the RX 6700S wins decisively, with a 199% lead in Geekbench Metal and a 349.2% lead in Geekbench Vulkan. The data shows a generational leap in compute throughput, memory bandwidth, and feature support, though the R9 M390X retains a higher shading unit count and a wider memory bus. This analysis breaks down the architectural, benchmark, and specification differences between the two mobile GPUs.
FAQ
Q: Which GPU has the higher average benchmark score?
A: The AMD Radeon RX 6700S has an average benchmark score of 22154, which is 7.2% higher than the R9 M390X’s 20662. The RX 6700S sits at the 67th percentile of all GPUs, while the R9 M390X sits at the 66th percentile.
Q: How much faster is the RX 6700S in Geekbench Vulkan?
A: The RX 6700S scores 76150 in Geekbench Vulkan, while the R9 M390X scores 16951. This represents a 349.2% advantage for the RX 6700S, the largest delta in any shared benchmark.
Q: What is the memory configuration difference?
A: The RX 6700S features 8 GB of GDDR6 memory on a 128-bit bus with 224.0 GB/s bandwidth. The R9 M390X has 4 GB of GDDR5 memory on a 256-bit bus with 160.0 GB/s bandwidth. Despite the wider bus, the R9 M390X’s bandwidth is 28.6% lower.
Q: Which GPU has more shading units?
A: The R9 M390X has 2048 shading units, which is 256 more than the RX 6700S’s 1792. However, the RX 6700S achieves much higher FP32 throughput at 7.168 TFLOPS versus 2.961 TFLOPS for the R9 M390X.
Q: What is the process node difference?
A: The RX 6700S is fabricated on TSMC’s 7 nm process, while the R9 M390X uses TSMC’s 28 nm process. The RX 6700S packs 11,060 million transistors on a 237 mm² die, whereas the R9 M390X has 5,000 million transistors on a 366 mm² die.
Q: Which GPU supports newer API features?
A: The RX 6700S supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the R9 M390X supports DirectX 12 (12_0) and Vulkan 1.2.170. The RX 6700S also includes 28 ray tracing cores, which the R9 M390X lacks entirely.
Architecture Differences
The architectural gulf between these two GPUs is stark. The RX 6700S is built on RDNA 2.0, a modern architecture designed for efficiency and scalability, while the R9 M390X uses GCN 3.0, a design that dates back to the mid-2010s. This difference manifests in nearly every metric.
The process node is the most fundamental split. The RX 6700S uses TSMC’s 7 nm process, allowing 11,060 million transistors on a 237 mm² die. This yields a transistor density of 46.7 million per mm². The R9 M390X, by contrast, uses TSMC’s 28 nm process, cramming only 5,000 million transistors onto a much larger 366 mm² die, for a density of just 13.7 million per mm². The RX 6700S achieves 3.4 times the transistor density, a direct result of the process advantage.
The compute layout differs significantly. The R9 M390X has 2048 shading units and 128 texture mapping units (TMUs), but only 32 ROPs. The RX 6700S has fewer shading units (1792) and TMUs (112), but doubles the ROP count to 64. This configuration gives the RX 6700S a pixel rate of 128.0 GPixel/s versus 23.14 GPixel/s for the R9 M390X, a 5.5-fold advantage. Texture rate follows a similar pattern: 224.0 GTexel/s for the RX 6700S versus 92.54 GTexel/s for the R9 M390X.
Memory architecture is another major divergence. The RX 6700S uses 8 GB of GDDR6 on a 128-bit bus, running at 1750 MHz with 14 Gbps effective speed. The R9 M390X uses 4 GB of GDDR5 on a 256-bit bus, running at 1250 MHz with 5 Gbps effective speed. Although the R9 M390X’s bus is twice as wide, its slower memory clock results in 160.0 GB/s bandwidth, far below the RX 6700S’s 224.0 GB/s. The RX 6700S also has a smaller memory footprint per bandwidth unit, which indicates better memory efficiency.
Feature support is where the RX 6700S pulls decisively ahead. It includes 28 ray tracing cores, enabling hardware-accelerated ray tracing, a feature completely absent from the R9 M390X. The RX 6700S supports DirectX 12 Ultimate (12_2), while the R9 M390X is limited to DirectX 12 (12_0). Vulkan support also differs, with the RX 6700S on version 1.4 versus 1.2.170 for the R9 M390X. Both support OpenGL 4.6.
The RX 6700S’s FP16 throughput is 14.34 TFLOPS (2:1 ratio), while the R9 M390X’s FP16 is 2.961 TFLOPS (1:1 ratio). This means the RX 6700S can process half-precision workloads at twice its FP32 rate, a feature absent on the older GPU. The R9 M390X’s FP32 and FP16 rates are identical, reflecting its older compute architecture.
Physical and interface differences also matter. The RX 6700S is an IGP (integrated graphics processor) with no power connectors and a PCIe 4.0 x8 interface. The R9 M390X is an MXM module with no power connectors and a PCIe 3.0 x16 interface. The RX 6700S’s TDP is 80 W, slightly higher than the R9 M390X’s 75 W, but that margin is negligible given the performance gap.
Head-to-Head Benchmarks
The head-to-head data is limited to two tests, but the results are unambiguous. In Geekbench Metal, the RX 6700S scores 72864, while the R9 M390X scores 24373. This gives the RX 6700S a 199% advantage. In Geekbench Vulkan, the RX 6700S scores 76150, versus 16951 for the R9 M390X, a 349.2% lead. The RX 6700S wins both tests, with the Vulkan delta being particularly lopsided.
These deltas are consistent with the architectural differences. The RX 6700S’s FP32 throughput (7.168 TFLOPS) is 2.4 times higher than the R9 M390X’s (2.961 TFLOPS). Memory bandwidth is 1.4 times higher. The ROP count is double. These raw numbers translate directly into the benchmark outcomes, with the Vulkan test showing an even larger gap, likely due to the RX 6700S’s newer API support and more efficient compute pipeline.
The average benchmark score further contextualizes these results. The RX 6700S averages 22154 across all its benchmark runs, while the R9 M390X averages 20662. The RX 6700S’s nearest rival is the AMD Radeon RX 5700, with an average score of 22170 and a delta of -0.1%, meaning the RX 6700S is practically tied with that desktop GPU. The R9 M390X’s nearest rival is the AMD Radeon RX 5600 XT, with an average score of 20713 and a delta of -0.2%, placing it in a similar performance tier.
The RX 6700S also has more benchmark data points, including 3DMark Steel Nomad (1451), Passmark G3D (15066), and Passmark GPU Compute (5990). The R9 M390X has no such results, making direct comparisons impossible beyond the two Geekbench tests. The RX 6700S’s Passmark G2D score of 739 and DirectX 11 score of 134 indicate solid legacy performance, while its DirectX 9 score of 195 suggests it handles older workloads well.
The Verdict
The data makes the choice clear for most use cases. The RX 6700S outperforms the R9 M390X by 199% in Metal and 349.2% in Vulkan, with a 7.2% higher average benchmark score. It offers twice the memory capacity (8 GB vs 4 GB), higher bandwidth (224.0 GB/s vs 160.0 GB/s), and modern features like ray tracing and DirectX 12 Ultimate. The RX 6700S is the superior GPU for any workload that leverages newer APIs, higher resolutions, or ray-traced content.
The R9 M390X, despite its lower performance, has a few structural advantages. It has more shading units (2048 vs 1792), more TMUs (128 vs 112), and a wider 256-bit memory bus. Its 75 W TDP is slightly lower than the RX 6700S’s 80 W. However, these advantages do not translate into benchmark wins, as the R9 M390X’s older GCN 3.0 architecture and 28 nm process limit its efficiency and throughput.
For users considering a laptop with the R9 M390X, the data suggests it is a legacy part, end-of-life since its 2015 release. The RX 6700S, released in 2022, is also end-of-life, but it represents a far more capable modern solution. The RX 6700S’s 67th percentile ranking versus the R9 M390X’s 66th percentile shows both sit near the middle of the GPU distribution, but the RX 6700S does so with substantially higher absolute performance.
The verdict is straightforward: the RX 6700S is the clear winner in every shared benchmark and most specification categories. The R9 M390X is only preferable in scenarios where its MXM form factor or lower TDP is a hard requirement, and even then, the performance penalty is severe.
Specification Differences
The following table highlights only the fields where the two GPUs differ:
| Field | AMD Radeon RX 6700S | AMD Radeon R9 M390X |
|-------|---------------------|----------------------|
| Architecture | RDNA 2.0 | GCN 3.0 |
| Generation | Navi Mobile (RX 6000M) | Gem System (R9 M300) |
| Process Node | 7 nm | 28 nm |
| Transistors | 11,060 million | 5,000 million |
| Die Size | 237 mm² | 366 mm² |
| Transistor Density | 46.7M / mm² | 13.7M / mm² |
| Memory Clock | 1750 MHz (14 Gbps effective) | 1250 MHz (5 Gbps effective) |
| Memory Size | 8 GB | 4 GB |
| Memory Type | GDDR6 | GDDR5 |
| Memory Bus Width | 128 bit | 256 bit |
| Memory Bandwidth | 224.0 GB/s | 160.0 GB/s |
| Shading Units | 1792 | 2048 |
| TMUs | 112 | 128 |
| ROPs | 64 | 32 |
| RT Cores | 28 | None |
| Pixel Rate | 128.0 GPixel/s | 23.14 GPixel/s |
| Texture Rate | 224.0 GTexel/s | 92.54 GTexel/s |
| FP32 | 7.168 TFLOPS | 2.961 TFLOPS |
| FP16 | 14.34 TFLOPS (2:1) | 2.961 TFLOPS (1:1) |
| TDP | 80 W | 75 W |
| Slot Width | IGP | MXM Module |
| Bus Interface | PCIe 4.0 x8 | PCIe 3.0 x16 |
| DirectX | 12 Ultimate (12_2) | 12 (12_0) |
| Vulkan | 1.4 | 1.2.170 |
| Release Date | 2022-01-03 | 2015-05-04 |
| Predecessor | Polaris Mobile | Solar System |
| Successor | None | Polaris Mobile |
| Geekbench Metal | 72864 | 24373 |
| Geekbench Vulkan | 76150 | 16951 |
| Avg Benchmark Score | 22154 | 20662 |
| Percentile | 67 | 66 |