AMD Radeon R5 M320 vs AMD Radeon R7 M340 Comparison

When two mobile GPUs from the same generation carry nearly identical specifications on paper, the recorded benchmark data becomes the only reliable way to separate them. The AMD Radeon R7 M340 and AMD Radeon R5 M320 share the same release date, the same process node, and even the same shading unit count, yet the database shows them splitting results in opposite directions across the two recorded compute tests. That split is the story here, and it raises a genuinely interesting question about how much a newer architecture generation matters when raw hardware counts are equal.

Head-to-Head Benchmarks

The head-to-head record is a clean one-for-one split, but the margins are anything but symmetric.

In the Geekbench OpenCL test, the Radeon R5 M320 comes out ahead with a score of 5051 against the R7 M340's 4932, a delta of 2.4 percent in the M320's favor. That is a narrow win, close enough that it sits within the range where run-to-run variance could plausibly blur the picture. The M320's OpenCL result is also nearly identical to its overall average score of 4657-adjusted profile, suggesting the OpenCL path is where this chip punches hardest.

The Geekbench Vulkan test flips the script emphatically. The R7 M340 scores 5193 while the M320 manages only 4262, an advantage of 21.8 percent for the M340. This is not a marginal difference; it is a gap large enough to define the practical difference between these two parts. A 21.8 percent lead in a modern graphics API test points toward something architectural rather than incidental, especially given that both chips have the same number of shading units, texture units, and render outputs.

Averaging the recorded benchmarks gives the R7 M340 an average score of 5063 against the M320's 4657, which tracks with their percentile placements: the M340 sits at the 30th percentile versus all GPUs in the database, while the M320 sits at the 27th. Both are entry-level placements, but the M340 holds a consistent edge in the aggregate view. Each chip wins exactly one benchmark, so the raw tally is tied at one apiece, yet the magnitude of the Vulkan result tilts the overall profile firmly toward the M340.

Where Each One Wins

The R5 M320's single win comes in OpenCL compute. Its score of 5051 there actually exceeds the M340's 4932 despite the M320 being the weaker part on paper by clock and compute throughput. What could explain that? The database offers one conspicuous clue: memory capacity. The M320 carries 4 GB of DDR3, double the M340's 2 GB, on an identical 64-bit bus with identical 16.00 GB/s of bandwidth. For OpenCL workloads that stream larger working sets, the extra capacity may keep data resident where the M340's smaller allocation would not. The data cannot prove causation, but the correlation is worth noting for anyone running capacity-sensitive compute tasks.

The R7 M340's win is in Vulkan, and by a wide margin. Vulkan 1.2.170 support is listed for both chips, but the M340's GCN 3.0 architecture, with its DirectX 12_0 feature level, is the newer design of the pair. The M320's GCN 1.0 roots cap it at DirectX 11_1 feature level. That gap in feature level support lines up neatly with the 21.8 percent Vulkan delta, implying the M340 is better positioned for modern API-driven rendering workloads, while the M320's strength is confined to legacy compute paths.

In short: for graphics work through current APIs, the M340 is the clear pick of the data. For OpenCL compute with larger memory footprints, the M320 posts the better number, narrowly.

Architecture Differences

Both GPUs were released on the same date, 2015-05-04, manufactured by TSMC on a 28 nm process, and both belong to the Gem System (R7 M300 and R5 M300) generation with the Solar System predecessor and Polaris Mobile successor. From there, the paths diverge.

The R7 M340 uses the Meso chip built on GCN 3.0. The R5 M320 uses the Jet chip built on the considerably older GCN 1.0. The silicon scale differs dramatically: the Meso die measures 125 mm² with 1,550 million transistors, while the Jet die is a mere 56 mm² with 690 million transistors. Interestingly, transistor density is almost identical between them, 12.4M per mm² for the M340 versus 12.3M per mm² for the M320, which is consistent with both sharing the same 28 nm node. The M340's die is more than twice the area, yet the functional counts are the same: 320 shading units, 20 TMUs, and 8 ROPs on both.

Clocks are where the functional parity breaks. The M340 runs a 943 MHz base and 1021 MHz boost, against the M320's 780 MHz base and 855 MHz boost. Those higher clocks translate directly into the throughput figures: the M340 delivers 653.4 GFLOPS FP32, 8.168 GPixel/s, and 20.42 GTexel/s, while the M320 delivers 547.2 GFLOPS FP32, 6.840 GPixel/s, and 17.10 GTexel/s. The M340 also lists FP16 throughput of 653.4 GFLOPS at a 1:1 ratio with FP32; the M320 has no FP16 figure recorded.

Memory is inverted relative to compute strength. The M320 carries 4 GB of DDR3 on a 64-bit bus at 16.00 GB/s; the M340 carries 2 GB of DDR3 on the same bus width and bandwidth. Both use a PCIe 3.0 x8 interface, and both memory setups run at 1000 MHz, 2 Gbps effective.

Neither chip has ray tracing cores, tensor cores, or recorded TDP figures. Both are end-of-life. The M320 is classified as an IGP with portable-device-dependent display outputs; no display output data is recorded for the M340. API support is identical on OpenGL 4.6 and Vulkan 1.2.170, but differs on DirectX: 12_0 for the M340 versus 11_1 for the M320, a meaningful distinction for feature-level compatibility.

FAQ

Q: Which GPU is faster overall according to the database? A: The Radeon R7 M340, with an average benchmark score of 5063 versus 4657 for the R5 M320, and a higher percentile placement, 30th versus 27th against all GPUs in the database.

Q: Did the R5 M320 win any benchmark? A: Yes. It won the Geekbench OpenCL test, 5051 to 4932, a 2.4 percent margin over the R7 M340.

Q: How large was the R7 M340's Vulkan win? A: The M340 scored 5193 in Geekbench Vulkan against the M320's 4262, a 21.8 percent advantage, the biggest gap in the head-to-head record.

Q: Do the two GPUs have the same number of shading units? A: Yes. Both have 320 shading units, 20 TMUs, and 8 ROPs. The difference lies in clocks, architecture generation, and die size.

Q: Which GPU has more memory? A: The R5 M320, with 4 GB of DDR3 compared to 2 GB on the R7 M340. Both use a 64-bit bus with 16.00 GB/s of bandwidth.

Q: Which rivals score closest to these GPUs? A: The R7 M340's nearest rivals include the AMD Radeon R7 240 at an identical average score, the AMD FirePro W4170M within 0.6 percent, and the AMD Radeon R5 M430 within 0.9 percent. The R5 M320's rivals include the NVIDIA Quadro P400, the NVIDIA GeForce GTX 970M, and the NVIDIA Quadro M3000M, all within roughly one percent of its average.

Specification Differences

| Field | AMD Radeon R7 M340 | AMD Radeon R5 M320 |

|---|---|---|

| Chip | Meso | Jet |

| Architecture | GCN 3.0 | GCN 1.0 |

| Transistors | 1,550 million | 690 million |

| Die Size | 125 mm² | 56 mm² |

| Transistor Density | 12.4M / mm² | 12.3M / mm² |

| Base Clock | 943 MHz | 780 MHz |

| Boost Clock | 1021 MHz | 855 MHz |

| Memory Size | 2 GB | 4 GB |

| FP32 | 653.4 GFLOPS | 547.2 GFLOPS |

| FP16 | 653.4 GFLOPS (1:1) | Not recorded |

| Pixel Rate | 8.168 GPixel/s | 6.840 GPixel/s |

| Texture Rate | 20.42 GTexel/s | 17.10 GTexel/s |

| DirectX | 12 (12_0) | 12 (11_1) |

| Slot Width | Not recorded | IGP |

| Display Outputs | Not recorded | Portable Device Dependent |

Both share the 28 nm TSMC node, 64-bit DDR3 bus at 16.00 GB/s, PCIe 3.0 x8, OpenGL 4.6, Vulkan 1.2.170, the same release date, and end-of-life status.

The Verdict

The data favors the Radeon R7 M340 for nearly everyone. It wins the modern API benchmark by 21.8 percent, posts the higher average score, holds the higher percentile ranking, delivers more compute throughput across FP32, pixel, and texture rates, and supports the newer DirectX 12_0 feature level on a newer GCN 3.0 architecture. Its rivals cluster tightly around it, including an exact score match with the desktop Radeon R7 240, indicating its performance profile is well characterized and stable in the database.

The R5 M320 earns its niche only under specific conditions. Its OpenCL win, though narrow at 2.4 percent, combined with double the memory capacity, makes it the recorded choice for capacity-bound compute workloads on legacy APIs. Its smaller 56 mm² die and IGP classification also mark it as the more integrated, space-constrained solution of the pair.

For graphics workloads, the answer is unambiguous: the M340's Vulkan margin is decisive. For OpenCL with large working sets, the M320's 4 GB allocation and slight score edge give it the only win it needs. Both are end-of-life parts now, but the benchmark record preserves a clear hierarchy between them.