AMD Radeon Pro 5700 vs AMD Radeon R9 M380 Comparison

The AMD Radeon Pro 5700 and the AMD Radeon R9 M380 occupy very different points on the GPU timeline, and the recorded data reflects that distance plainly. The Pro 5700, built on the Navi 10 chip with RDNA 1.0 architecture, holds an average benchmark score of 18189 against the R9 M380's 15521, and it wins both head-to-head tests available in the database. The older card, a Strato-based GCN 2.0 part from the R9 M300 generation, sits in the 58th percentile of all GPUs, while the Pro 5700 sits in the 62nd. Those percentiles are closer than the benchmark gaps suggest, which raises an interesting question about how each card's score profile is distributed, and one worth digging into test by test.

Head-to-Head Benchmarks

Only two tests exist for both cards in the database, and the Pro 5700 dominates both decisively.

In Geekbench Metal, the Pro 5700 scored 60126 while the R9 M380 managed 18476, a margin of 225.4 percent in the Pro card's favor. Geekbench OpenCL shows an even wider chasm: 52787 versus 12565, a 320.1 percent delta. These are not close contests. They are generational blowouts, and the underlying hardware explains why: the Pro 5700 delivers 6.221 TFLOPS of FP32 compute against the R9 M380's 1.536 TFLOPS, a theoretical compute advantage of roughly four times that shows up clearly, though not linearly, in the measured results.

What the OpenCL numbers imply is worth pausing on. A 320.1 percent gap exceeds even the raw compute ratio, suggesting the R9 M380 is further handicapped by memory bandwidth. The Pro 5700 has 384.0 GB/s of bandwidth on a 256-bit GDDR6 bus; the R9 M380 has 96.00 GB/s on a 128-bit GDDR5 bus, exactly one quarter. In bandwidth-hungry compute workloads, that constraint compounds the shader deficit, and the OpenCL result appears to reflect exactly that compounding.

For context within the database's ranking system, the Pro 5700's average score of 18189 places it in a cluster with the NVIDIA GeForce RTX 3060 Mobile (avgScore 18159, just 0.2 percent apart) and the GeForce RTX 2060 SUPER (18093, 0.5 percent apart). The R9 M380's 15521 average puts it within 0.2 percent of the GeForce GTX 1080 Ti's recorded 15548 and 1 percent of the Radeon Pro W5500's 15679. Curiously, its rival list also includes cards like the RTX 2060 (15290) and the GTX 580 (15283), a remarkably spread-out neighborhood that hints at uneven score distributions across test types. That spread deserves scrutiny before drawing conclusions from the average alone.

The Pro 5700's fuller benchmark record reinforces the picture. It posted 11583 in PassMark G3D, 4961 in PassMark GPU Compute, 805 in PassMark G2D, and PassMark DirectX results of 143 (DX9), 78 (DX11), 71 (DX10), and 48 (DX12). The R9 M380 has no recorded results in any of those tests, so cross-comparison there is not possible; the head-to-head record stands on the two Geekbench entries alone.

FAQ

Q: Which card is faster overall?

A: The Radeon Pro 5700. It won both head-to-head benchmarks, by 225.4 percent in Geekbench Metal and 320.1 percent in Geekbench OpenCL, and its average benchmark score of 18189 exceeds the R9 M380's 15521.

Q: How do the two cards compare in raw compute throughput?

A: The Pro 5700 delivers 6.221 TFLOPS FP32 and 12.44 TFLOPS FP16 (at a 2:1 ratio). The R9 M380 delivers 1.536 TFLOPS FP32, and no FP16 figure is recorded for it.

Q: What memory does each card have?

A: The Pro 5700 has 8 GB of GDDR6 on a 256-bit bus with 384.0 GB/s of bandwidth. The R9 M380 has 4 GB of GDDR5 on a 128-bit bus with 96.00 GB/s of bandwidth.

Q: Which GPUs does the database list as nearest rivals?

A: For the Pro 5700: the RTX 3060 Mobile (0.2 percent), the RTX 2060 SUPER (0.5 percent), the Radeon RX 460 (-1 percent), and the Arc A770M (-1.1 percent). For the R9 M380: the GTX 1080 Ti (-0.2 percent), the Radeon Pro W5500 (-1 percent), the RTX 2060 (1.5 percent), and the GTX 580 (1.6 percent).

Q: Are either of these cards still in production?

A: No. Both are listed as end-of-life. The R9 M380 was released on 2015-05-04, and the Pro 5700 on 2020-08-03.

Q: How do their percentile rankings compare?

A: The Pro 5700 sits in the 62nd percentile versus all GPUs in the database; the R9 M380 sits in the 58th. The four-point gap is modest relative to the much larger benchmark deltas, which suggests the percentile calculation weights the score profiles differently than the two head-to-head tests do.

Architecture Differences

The architectural gulf between these two is wide. The Pro 5700 uses the Navi 10 chip on AMD's RDNA 1.0 architecture, fabricated at TSMC on a 7 nm process. The R9 M380 uses the Strato chip on GCN 2.0, fabricated at TSMC on a 28 nm process. Four process-node generations separate them in the data, and the density numbers make the consequence explicit: the Pro 5700 packs 41.0M transistors per square millimeter against 13.0M per square millimeter for the R9 M380.

That density advantage lets the Pro 5700 fit 10,300 million transistors into a 251 mm² die. The R9 M380 carries 2,080 million transistors on a 160 mm² die, meaning the newer chip has roughly five times the transistor count on a die only about half again as large. Nearly all of that extra silicon goes to execution resources: 2304 shading units, 144 TMUs, and 64 ROPs on the Pro 5700 versus 768 shading units, 48 TMUs, and 16 ROPs on the R9 M380. In every case the Pro 5700 has three or four times the resources.

Throughput rates follow. The Pro 5700 achieves 86.40 GPixel/s pixel rate and 194.4 GTexel/s texture rate; the R9 M380 achieves 16.00 GPixel/s and 48.00 GTexel/s respectively. Neither card has RT cores or tensor cores recorded, so hardware ray tracing and dedicated AI acceleration are absent from both feature sets.

API support differs in minor ways. Both support DirectX 12 and OpenGL 4.6, but the Pro 5700 is recorded at feature level 12_1 versus the R9 M380's 12_0, and its Vulkan support is listed at 1.4 versus 1.2.170. The Pro 5700 also connects over PCIe 4.0 x16, while the R9 M380 is limited to PCIe 3.0 x16.

Lineage is recorded only for the older card: the R9 M380's predecessor is listed as Solar System and its successor as Polaris Mobile. No predecessor or successor is recorded for the Pro 5700, and both belong to different product generations in the database, Radeon Pro Mac (Navi Series) versus Gem System (R9 M300).

Specification Differences

The spec sheet divergences, where the two cards differ:

• Clocks: Pro 5700 at 1243 MHz base and 1350 MHz boost; R9 M380 at 900 MHz base and 1000 MHz boost. Memory runs at the same 1500 MHz, but effective rates differ: 12 Gbps versus 6 Gbps.
• Memory: 8 GB GDDR6 on a 256-bit bus (384.0 GB/s) versus 4 GB GDDR5 on a 128-bit bus (96.00 GB/s).
• Power: The Pro 5700 records a 130 W TDP, a 300 W suggested PSU, an IGP slot width, and no power connectors. None of these fields are recorded for the R9 M380.
• Outputs: The Pro 5700 lists no display outputs; the R9 M380's output field is unrecorded.
• FP16: 12.44 TFLOPS (2:1) on the Pro 5700; not recorded on the R9 M380.

Fields where the two agree or where data is missing for both include transistor foundry (TSMC), DirectX and OpenGL support levels, production status (both end-of-life), and the absence of recorded dimensions or launch MSRP for either card.

The Verdict

Strictly from the recorded data, the Pro 5700 is the stronger graphics card in every measurable dimension. It wins both head-to-head benchmarks by margins exceeding 225 percent, it has three to four times the execution resources, four times the memory bandwidth, twice the memory capacity, higher clocks on every recorded metric, and a more modern API and bus feature set. Its average score places it alongside the RTX 3060 Mobile and RTX 2060 SUPER, a demonstrably higher tier of performance than the R9 M380's neighborhood.

The R9 M380's case is thinner but not empty. It belongs to a mobile-focused generation, carries recorded lineage from Solar System to Polaris Mobile, and its percentile placement, 58th versus 62nd, indicates the database's overall ranking sees the two as closer than the Geekbench results alone would suggest. Whether that closeness survives in workloads the older card was never tested on is an open question the data cannot answer, since its record contains only the two Geekbench entries. What the data does say, unambiguously, is that in every test both cards have run, the Pro 5700 wins, and by a wide margin.

Where Each One Wins

The Pro 5700 wins everywhere the database has measurements. In Geekbench Metal, its 60126 to 18476 victory makes it the clear choice for Metal-based graphics work on Apple platforms. In Geekbench OpenCL, its 52787 to 12565 result, a 320.1 percent margin, marks it as the pick for general GPU compute. Its additional recorded results in PassMark G3D, GPU Compute, G2D, and DirectX 9 through 12 tests, all without counterparts for the R9 M380, mean the modern card is the only one with verifiable results in rasterization and legacy DirectX workloads at all.

The R9 M380 wins nothing in the head-to-head record, with zero wins to the Pro 5700's two. Its relevance in the database is historical rather than competitive: it documents an era of 28 nm GCN mobile hardware, and its rival cluster, which spans from the GTX 580 to the GTX 1080 Ti and RTX 2060, suggests its recorded score profile mixes eras in ways that make its average an unreliable single number. For anyone choosing between these two on performance grounds, the recorded data leaves no ambiguity: the Pro 5700 is the card that wins, and it wins big.