AMD Radeon R9 M380 vs NVIDIA P106-090 Comparison

The AMD Radeon R9 M380 and the NVIDIA P106-090 occupy very different corners of the hardware landscape, and the recorded benchmark data reflects that divergence. One is a mobile-oriented GPU from AMD’s GCN 2.0 era, while the other is a mining-focused NVIDIA board with no display outputs. Their head-to-head comparison is sparse, but the numbers available reveal a decisive gap in raw compute performance, while the surrounding specification sheet shows two designs built for entirely separate purposes.

Head-to-Head Benchmarks

The only direct benchmark comparison in the database is the Geekbench OpenCL test. Here, the NVIDIA P106-090 scores 21,304, while the AMD Radeon R9 M380 scores 12,565. That difference translates to the P106-090 leading by 41% in this workload. The delta is substantial, not marginal. In practical terms, the NVIDIA part delivers roughly seven-tenths more compute throughput in OpenCL, which often serves as a proxy for general-purpose GPU workloads, including rendering, physics simulation, and some machine learning tasks.

What makes this gap more striking is the context of their respective average scores. The R9 M380 holds an average benchmark score of 15,521, while the P106-090 averages 13,470. That is a curious inversion: the R9 M380 has a higher overall average across all tested workloads, yet it loses decisively in the one test where they directly compete. The explanation lies in the benchmark mix. The R9 M380’s average is buoyed by a Geekbench Metal score of 18,476, a test the P106-090 does not appear in. The P106-090, meanwhile, posts a Geekbench Vulkan score of 18,596, which is also excluded from the head-to-head table. So the OpenCL result is the only common ground, and on that ground, NVIDIA wins by a wide margin.

The nearest rivals for each product further illuminate their standing. The R9 M380’s average score of 15,521 places it within 0.2% of the NVIDIA GeForce GTX 1080 Ti (15,548), and 1% behind the AMD Radeon Pro W5500 (15,679). It also edges out the RTX 2060 (15,290) by 1.5% and the GTX 580 (15,283) by 1.6%. That is an unusual neighborhood for a mobile GPU, but the data suggests the R9 M380 punches well above its apparent class in aggregate benchmarks. The P106-090’s average of 13,470 sits nearly level with the GTX 570 (13,515, a 0.3% deficit) and the Radeon HD 7770M (13,536, a 0.5% deficit). It also trails the Radeon RX 9070 XT by 0.5%, a name that implies far higher performance, yet the recorded numbers show near parity in average score.

The percentile rankings tell a similar story. The R9 M380 sits at the 58th percentile among all GPUs in the database, while the P106-090 sits at the 54th percentile. That four-point gap aligns with their average scores, but the OpenCL head-to-head flips the narrative. The database records zero wins for the R9 M380 and one win for the P106-090, so if the question is which GPU wins the only direct contest, the answer is unambiguous.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon R9 M380, with an average score of 15,521, compared to 13,470 for the NVIDIA P106-090. That puts the AMD part 15% higher in aggregate performance across all recorded tests.

Q: Why does the P106-090 win the OpenCL test by such a large margin?

A: The P106-090 scores 21,304 in Geekbench OpenCL, while the R9 M380 scores 12,565. That is a 41% difference in favor of the NVIDIA board. The P106-090’s higher clock speeds and newer architecture likely contribute, though the database only records the scores, not the causal factors.

Q: Does the R9 M380 have any benchmark win over the P106-090?

A: No. In the head-to-head benchmark table, the R9 M380 records zero wins, while the P106-090 records one win. The only shared test is Geekbench OpenCL, and the NVIDIA part wins it outright.

Q: How do these GPUs compare to their nearest rivals in average score?

A: The R9 M380 is within 0.2% of the GTX 1080 Ti, 1% behind the Radeon Pro W5500, 1.5% ahead of the RTX 2060, and 1.6% ahead of the GTX 580. The P106-090 is 0.3% behind the GTX 570, 0.5% ahead of the Radeon Pro 555, and 0.5% behind both the HD 7770M and RX 9070 XT.

Q: What is the percentile ranking of each GPU?

A: The R9 M380 sits at the 58th percentile of all GPUs, while the P106-090 sits at the 54th percentile. This suggests the AMD part outperforms a larger fraction of the database’s GPU population in aggregate terms.

Q: Which GPU has more shading units?

A: Both have 768 shading units, so they are equal in that regard. The R9 M380 also has 48 texture mapping units, and the P106-090 has 48 as well, though their render output units differ significantly.

The Verdict

The data points to a clear split: the NVIDIA P106-090 is the better compute performer in the one test where they meet, but the AMD Radeon R9 M380 holds a higher aggregate standing across the broader benchmark suite. If the task is OpenCL-heavy, such as certain rendering pipelines or data-parallel workloads, the P106-090 is the stronger choice, and not by a small amount. The 41% lead is decisive. Conversely, if the workload relies on Metal or Vulkan, the R9 M380’s recorded scores suggest it has strengths in those areas, though direct comparison is impossible because the P106-090 lacks Metal support in the database and the R9 M380 lacks a Vulkan score.

For a user who needs a GPU with display outputs, the P106-090 is immediately disqualified, since the database lists its display outputs as “No outputs.” The R9 M380, by contrast, is a conventional GPU that can drive displays. For a mining or compute-only setup where display output is irrelevant, the P106-090’s higher OpenCL score and lower power footprint make it the more compelling option from the recorded data alone. The R9 M380’s higher average score and better percentile placement suggest it is a more balanced performer, but the head-to-head result tips the scale toward NVIDIA for raw compute.

Specification Differences

The two GPUs diverge sharply across nearly every hardware specification. The R9 M380 uses a 28 nm process, while the P106-090 uses 16 nm. Both are manufactured by TSMC, but the transistor counts differ dramatically: the R9 M380 has 2,080 million transistors on a 160 mm² die, while the P106-090 has 4,400 million transistors on a 200 mm² die. That translates to a transistor density of 13.0 million per square millimeter for AMD versus 22.0 million per square millimeter for NVIDIA.

Clock speeds also favor NVIDIA. The R9 M380 runs at a 900 MHz base and 1000 MHz boost, while the P106-090 runs at 1354 MHz base and 1531 MHz boost. Memory clocks show a similar pattern: the R9 M380 uses 1500 MHz (6 Gbps effective) GDDR5, while the P106-090 uses 2002 MHz (8 Gbps effective) GDDR5. Memory capacity and bus width differ as well: the R9 M380 has 4 GB on a 128-bit bus, yielding 96.00 GB/s of bandwidth, while the P106-090 has 3 GB on a 192-bit bus, yielding 192.2 GB/s. That is exactly double the bandwidth in raw terms, a massive advantage for NVIDIA.

The render output units present the largest structural difference. The R9 M380 has 16 ROPs, while the P106-090 has 48 ROPs. This affects pixel fill rate: the R9 M380 achieves 16.00 GPixel/s, while the P106-090 achieves 73.49 GPixel/s. Texture rate also favors NVIDIA, with 73.49 GTexel/s versus 48.00 GTexel/s. Floating-point performance follows: the R9 M380 delivers 1.536 TFLOPS, while the P106-090 delivers 2.352 TFLOPS. The P106-090 also lists a half-precision figure of 36.74 GFLOPS at a 1:64 ratio, while the R9 M380 has no recorded FP16 data.

The power and physical profile differ entirely. The R9 M380 has no listed TDP, slot width, or power connectors, whereas the P106-090 has a 75 W TDP, a dual-slot form factor, a single 6-pin power connector, and a suggested power supply of 250 W. The P106-090 is 250 mm long (9.8 inches), while the R9 M380 has no recorded dimensions. Bus interfaces also diverge: the R9 M380 uses PCIe 3.0 x16, while the P106-090 uses PCIe 1.0 x1, a severe limitation for data transfer that does not appear in the compute benchmarks but would matter in real-world use.

Architecture Differences

The R9 M380 is built on GCN 2.0, using the Strato chip, and belongs to the Gem System generation under the R9 M300 series. It is a 2015 design, released on May 4, 2015, and is marked as end-of-life. Its predecessor is listed as Solar System, and its successor is Polaris Mobile. The P106-090, meanwhile, uses the Pascal architecture with the GP106 chip and falls under the Mining GPUs generation. It was released on July 30, 2017, and is also end-of-life. No predecessor or successor is recorded for the NVIDIA part.

The API support differs meaningfully. The R9 M380 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.2.170. The P106-090 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The higher DirectX feature level and newer Vulkan version on the P106-090 suggest more modern API capabilities, which could influence performance in newer titles or compute frameworks. Both GPUs lack ray tracing and tensor cores, as those fields are null for both.

The transistor and die differences point to architectural maturity. The P106-090 packs over twice the transistor count into a die that is only 25% larger, thanks to the 16 nm process. That density allows for higher clock speeds and more ROPs while keeping power at 75 W. The R9 M380, on a larger 28 nm node, cannot match that efficiency. The P106-090’s lack of display outputs is also an architectural choice: it is a mining SKU, stripped of video connectivity to reduce cost and complexity. The R9 M380, as a mobile GPU, retains full display capabilities.

Where Each One Wins

The NVIDIA P106-090 wins in raw compute throughput, as shown by the OpenCL score. It also wins on memory bandwidth (192.2 GB/s), pixel fill rate (73.49 GPixel/s), texture rate (73.49 GTexel/s), FP32 performance (2.352 TFLOPS), and clock speeds. Its API support is newer, with Vulkan 1.4 and DirectX 12_1. For any workload that stresses these metrics, such as compute shaders, high-resolution rendering, or memory-bandwidth-intensive tasks, the P106-090 is the better card.

The AMD Radeon R9 M380 wins on aggregate benchmark average, with 15,521 versus 13,470. It also holds a higher percentile ranking (58th versus 54th) and has a higher individual Metal score of 18,476, a test the P106-090 does not appear in. Its 4 GB memory capacity is larger than the P106-090’s 3 GB, which could matter for workloads that exceed 3 GB of VRAM. It uses a standard PCIe 3.0 x16 interface, whereas the P106-090 is limited to PCIe 1.0 x1, a bottleneck that could hurt in any data-transfer-heavy scenario. And critically, the R9 M380 has display outputs, making it a usable GPU for normal computing, while the P106-090 has none.

The database records the R9 M380 as having zero head-to-head wins and the P106-090 as having one. That single win is the OpenCL test, and it is a large one. But the broader picture shows two GPUs with different strengths: one is a compute-focused mining board with high throughput and no video output, the other is a more balanced mobile GPU with better aggregate scores and full display functionality. The choice depends entirely on whether the use case is headless compute or general-purpose graphics.