AMD Radeon R5 M320 vs NVIDIA GeForce GT 645M Comparison

The AMD Radeon R5 M320 and NVIDIA GeForce GT 645M are two end-of-life mobile graphics solutions from different architectural generations, with the AMD part arriving roughly three years later. Benchmark data shows a split decision: the AMD R5 M320 dominates in OpenCL compute workloads with an 88.5% lead, while the NVIDIA GT 645M counters with a 12.6% advantage in Vulkan performance, resulting in a 1-1 tie in head-to-head tests.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon R5 M320 holds a higher average benchmark score of 4657, compared to 4411 for the NVIDIA GeForce GT 645M. This represents a difference of 246 points, or roughly 5.6% in favor of the AMD part.

Q: How do the two compare in OpenCL performance?

A: The AMD Radeon R5 M320 is dramatically faster in OpenCL, scoring 5051 versus 2680 for the NVIDIA GeForce GT 645M. This is an 88.5% improvement for the AMD GPU, making it the clear winner in this compute API test.

Q: Which GPU wins in Vulkan benchmarks?

A: The NVIDIA GeForce GT 645M takes the Vulkan test with a score of 4875, while the AMD Radeon R5 M320 scores 4262. The NVIDIA part leads by 12.6% in this particular workload.

Q: What are the memory specifications of each GPU?

A: The AMD Radeon R5 M320 features 4 GB of DDR3 memory on a 64-bit bus, delivering 16.00 GB/s bandwidth. The NVIDIA GeForce GT 645M has 2 GB of DDR3 memory on a wider 128-bit bus, providing 28.80 GB/s bandwidth.

Q: Which GPU has a higher transistor count?

A: The NVIDIA GeForce GT 645M uses 1,270 million transistors on its GK107 chip, nearly double the 690 million transistors found on the AMD Radeon R5 M320's Jet chip. The NVIDIA die is also larger at 118 mm² versus 56 mm².

Q: How does the R5 M320 rank among all GPUs?

A: The AMD Radeon R5 M320 sits at the 27th percentile among all GPUs, with its nearest rival being the AMD Radeon RX 9060 XT 16 GB at a 0% delta, and the NVIDIA Quadro P400 just 0.6% ahead. The GT 645M is at the 26th percentile, with the AMD Radeon R7 M260 1.9% ahead.

Architecture Differences

The two GPUs represent fundamentally different design philosophies from their respective manufacturers. The AMD Radeon R5 M320 is built on the Graphics Core Next 1.0 architecture, using the Jet chip manufactured by TSMC on a 28 nm process. This chip contains 690 million transistors packed into a 56 mm² die, resulting in a transistor density of 12.3 million transistors per square millimeter. The architecture supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170.

The NVIDIA GeForce GT 645M, by contrast, employs the Kepler architecture with the GK107 chip, also built on TSMC's 28 nm process. This is a substantially larger chip at 118 mm², housing 1,270 million transistors, though the density is slightly lower at 10.8 million per square millimeter. Kepler supports DirectX 12 (11_0), OpenGL 4.6, and a newer Vulkan version at 1.2.175.

In terms of compute resources, the NVIDIA GPU has more of nearly everything: 384 shading units versus 320, 32 texture mapping units versus 20, and 16 raster operation pipelines versus 8. The AMD part compensates with higher clock speeds, running at 780 MHz base and 855 MHz boost, compared to the NVIDIA's 709 MHz base and 780 MHz boost. Memory configurations also differ significantly, with the AMD card offering 4 GB but on a narrow 64-bit bus, while the NVIDIA card provides 2 GB on a 128-bit bus.

The bus interface shows another key difference: the AMD R5 M320 connects via PCIe 3.0 x8, while the NVIDIA GT 645M uses the full PCIe 3.0 x16 path. Power characteristics also diverge, with the NVIDIA part having a rated TDP of 32 W and no power connectors, while the AMD card has no listed TDP in the data. Both are integrated-class mobile parts with portable device-dependent display outputs.

Head-to-Head Benchmarks

The benchmark results reveal a stark contrast in workload preferences between these two architectures. In the Geekbench OpenCL test, the AMD Radeon R5 M320 delivers a score of 5051, absolutely crushing the NVIDIA GeForce GT 645M's 2680. This 88.5% advantage is the single largest margin in any comparison between these two parts. The result is particularly striking given that the NVIDIA GPU has more shading units, more TMUs, more ROPs, and nearly double the memory bandwidth. The AMD architecture's compute-oriented design, with its GCN 1.0 foundation, clearly excels in this general-purpose compute workload.

The Vulkan test tells a different story. Here, the NVIDIA GeForce GT 645M rebounds with a score of 4875, surpassing the AMD R5 M320's 4262 by 12.6%. While this is a solid win for NVIDIA, it is importantly the margin is far smaller than AMD's OpenCL victory. The Vulkan results suggest that the Kepler architecture handles modern graphics APIs reasonably well despite its age, though it cannot match the sheer compute throughput of the GCN design in OpenCL.

Looking at average scores across all benchmarks, the AMD R5 M320's 4657 average is 5.6% higher than the NVIDIA's 4411. This aggregate figure masks the extreme divergence in individual tests, but it does indicate that the AMD part is generally the stronger performer in compute-oriented tasks. The R5 M320's percentile ranking of 27 versus 26 for the GT 645M reflects this modest overall advantage, with both parts sitting in the lower quartile of all GPUs.

The nearest rival data provides additional context. The AMD R5 M320 is statistically tied with the AMD Radeon RX 9060 XT 16 GB (0% delta), while the NVIDIA Quadro P400 is 0.6% faster and the GTX 970M is 0.6% slower. For the GT 645M, the NVIDIA GeForce 930M is just 0.5% ahead, and the Intel Iris Pro Graphics 5200 is 1.2% ahead, meaning the GT 645M is competitive within a tight cluster of similar-performance mobile GPUs.

The Verdict

The data paints a clear picture for different use cases. The AMD Radeon R5 M320 is the choice for OpenCL compute workloads, where its 88.5% advantage over the GT 645M is decisive. Its higher average benchmark score of 4657 versus 4411 also makes it the better all-round performer in the tested metrics. Gamers or users relying on Vulkan-based applications, however, would find the NVIDIA GeForce GT 645M more capable, given its 12.6% lead in that specific API test.

The AMD part's strengths lie in its compute architecture and higher clock speeds. Despite having fewer shading units (320 versus 384), fewer TMUs (20 versus 32), and fewer ROPs (8 versus 16), the R5 M320 leverages its GCN 1.0 design to achieve 547.2 GFLOPS of FP32 performance, compared to 599.0 GFLOPS for the NVIDIA part. The AMD card also offers double the memory capacity at 4 GB, though on a narrower bus that limits bandwidth to 16.00 GB/s versus the GT 645M's 28.80 GB/s.

For users prioritizing raw compute in OpenCL applications, the R5 M320 is the superior option. For those needing Vulkan performance or wider memory bandwidth, the GT 645M is preferable. The 1-1 split in head-to-head wins accurately reflects this duality. Neither card is competitive in the modern GPU landscape, as both sit at the 26th-27th percentile, but between the two, the AMD R5 M320 offers the better overall benchmark profile. The NVIDIA part's lower TDP of 32 W, however, may make it more suitable for thermally constrained systems, though the AMD card has no TDP listed for comparison.

Specification Differences

| Specification | AMD Radeon R5 M320 | NVIDIA GeForce GT 645M |

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

| Architecture | GCN 1.0 | Kepler |

| Chip | Jet | GK107 |

| Process Node | 28 nm | 28 nm |

| Transistors | 690 million | 1,270 million |

| Die Size | 56 mm² | 118 mm² |

| Transistor Density | 12.3M / mm² | 10.8M / mm² |

| Base Clock | 780 MHz | 709 MHz |

| Boost Clock | 855 MHz | 780 MHz |

| Memory Clock | 1000 MHz (2 Gbps effective) | 900 MHz (1800 Mbps effective) |

| Memory Size | 4 GB | 2 GB |

| Memory Type | DDR3 | DDR3 |

| Memory Bus | 64 bit | 128 bit |

| Memory Bandwidth | 16.00 GB/s | 28.80 GB/s |

| Shading Units | 320 | 384 |

| TMUs | 20 | 32 |

| ROPs | 8 | 16 |

| Pixel Rate | 6.840 GPixel/s | 6.240 GPixel/s |

| Texture Rate | 17.10 GTexel/s | 24.96 GTexel/s |

| FP32 Performance | 547.2 GFLOPS | 599.0 GFLOPS |

| TDP | Not listed | 32 W |

| Power Connectors | Not listed | None |

| Bus Interface | PCIe 3.0 x8 | PCIe 3.0 x16 |

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

| OpenGL Support | 4.6 | 4.6 |

| Vulkan Support | 1.2.170 | 1.2.175 |

| Release Date | 2015-05-04 | 2012-09-30 |

| Predecessor | Solar System | GeForce 500M |

| Successor | Polaris Mobile | GeForce 700M |

| Production Status | End-of-life | End-of-life |