AMD Radeon R5 Graphics vs NVIDIA GeForce 920M Comparison

AMD Radeon R5 Graphics and NVIDIA GeForce 920M are both end-of-life mobile graphics solutions from the 28 nm era, but they approach performance from very different design philosophies. The AMD part is an integrated graphics processor (IGP) built into a Kaveri-generation APU, while the NVIDIA part is a discrete-class chip designed for portable devices. Benchmark data from the database shows a split decision: AMD wins the OpenCL compute test by a wide margin, while NVIDIA wins the Vulkan graphics test by a smaller margin. Understanding where each part excels requires looking at their architectural priorities, memory configurations, and the specific workloads behind each benchmark.

Where Each One Wins

The clearest victory for the AMD Radeon R5 Graphics comes in the Geekbench OpenCL test. The database records a score of 5183 for the AMD part, against 3725 for the NVIDIA GeForce 920M. That is a 39.1% advantage for the AMD chip. OpenCL workloads tend to stress raw compute throughput, memory bandwidth, and the ability to keep many parallel lanes busy. The AMD Radeon R5 Graphics, with its GCN 2.0 architecture, is built around a design that scales well in such compute-heavy scenarios. The data indicates that for general-purpose GPU compute tasks, the AMD part is substantially ahead.

The NVIDIA GeForce 920M takes the Geekbench Vulkan test with a score of 2849, compared to 2582 for the AMD Radeon R5 Graphics. The delta is 9.4% in NVIDIA's favor. Vulkan is a low-level graphics API that places a premium on driver efficiency, draw-call handling, and the ability to manage command buffers with minimal overhead. The NVIDIA part, based on Kepler 2.0, shows a clear advantage in this environment. While the margin is smaller than AMD's OpenCL win, it is consistent and meaningful for applications that rely on Vulkan.

Looking at the broader database context, the AMD Radeon R5 Graphics sits at the 23rd percentile among all GPUs, with an average benchmark score of 3883. Its nearest rivals in the database include the NVIDIA Quadro 2000 at 3898 (0.4% lower), the NVIDIA Quadro K2000D at 3919 (0.9% lower), and the NVIDIA Quadro 2000D at 3930 (1.2% lower). Interestingly, the AMD part is also 1.3% ahead of the NVIDIA GeForce MX110, which scores 3834. This places the AMD IGP in a competitive position against older discrete workstation parts.

The NVIDIA GeForce 920M, by comparison, holds the 20th percentile among all GPUs, with an average benchmark score of 3287. Its nearest rivals include the NVIDIA GeForce GT 730M at 3316 (0.9% higher), the Intel HD Graphics 530 at 3332 (1.4% higher), and the Intel HD Graphics P4600 at 3389 (3% higher). The NVIDIA part is 2.4% ahead of the NVIDIA GeForce GT 640, which scores 3210. The data shows the 920M is positioned slightly below the AMD part in average score, but its Vulkan performance is superior.

Architecture Differences

The two chips come from different manufacturers and use different process nodes and foundries. The AMD Radeon R5 Graphics is built on a 28 nm process at GlobalFoundries, using the GCN 2.0 architecture. Its chip is labeled Spectre SL, and it belongs to the GCN 2.0 IGP generation for Kaveri. The die is 245 mm² with 2,410 million transistors, giving a transistor density of 9.8 million per square millimeter. The NVIDIA GeForce 920M uses a 28 nm process at TSMC, with the Kepler 2.0 architecture. Its chip is GK208B, part of the GeForce 900M generation. The die is much smaller at 87 mm², with 1,020 million transistors, yielding a higher transistor density of 11.7 million per square millimeter.

Shader resources differ significantly. The AMD part has 256 shading units, 16 texture mapping units (TMUs), and 4 raster operation units (ROPs). The NVIDIA part has 384 shading units, 32 TMUs, and 8 ROPs. In terms of pure pixel and texture throughput, the NVIDIA part is much higher: it delivers 7.632 GPixel/s and 30.53 GTexel/s, while the AMD part delivers 3.032 GPixel/s and 12.13 GTexel/s. The NVIDIA part also has a significantly higher FP32 compute rating at 732.7 GFLOPS, versus 388.1 GFLOPS for the AMD part. This is a striking contrast: the NVIDIA chip has nearly double the raw floating-point throughput of the AMD chip, yet loses badly in the OpenCL benchmark. This suggests that the OpenCL test is not purely a measure of peak FLOPs, but also depends on memory behavior, driver scheduling, and workload characteristics.

Memory configuration is a major differentiator. The AMD Radeon R5 Graphics uses system shared memory for both capacity and bandwidth. Its memory size, type, bus width, and bandwidth are all listed as system shared. The memory clock is listed as system shared. This means the AMD part is entirely dependent on the host system's memory speed and architecture. The NVIDIA GeForce 920M has a dedicated 2 GB DDR3 memory pool on a 64-bit bus, with a memory clock of 900 MHz (1800 Mbps effective) and a bandwidth of 14.40 GB/s. The presence of dedicated memory gives the NVIDIA part a predictable memory subsystem, whereas the AMD part's performance will vary with the platform.

Clock behavior also differs. The NVIDIA GeForce 920M has a base clock of 954 MHz and a boost clock of 954 MHz, meaning it runs at a fixed frequency. The AMD Radeon R5 Graphics has no listed base or boost clock in the database, so its operating frequency is not recorded. The NVIDIA part uses a PCIe 3.0 x8 bus interface, while the AMD part is an IGP with a bus interface listed as IGP. Display outputs are motherboard dependent for the AMD part and portable device dependent for the NVIDIA part.

API support shows a notable difference. The AMD Radeon R5 Graphics supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.2.170. The NVIDIA GeForce 920M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The AMD part has a higher DirectX feature level (12_0 versus 11_0), while the NVIDIA part has a slightly newer Vulkan version. Power consumption is also different: the AMD part has a TDP of 15 W, while the NVIDIA part has a TDP of 33 W. The NVIDIA part uses no power connectors, and the database lists its slot width as IGP, similar to the AMD part.

The Verdict

The data supports a clear split recommendation. For compute-oriented workloads that stress OpenCL, the AMD Radeon R5 Graphics is the stronger choice. Its 39.1% lead in the OpenCL benchmark is decisive, and its 23rd percentile average score of 3883 places it above the 20th percentile of the NVIDIA GeForce 920M. The AMD part also benefits from a much lower TDP of 15 W versus 33 W, which is important for thermally constrained portable systems. Its DirectX 12 (12_0) support is also more modern than the NVIDIA part's DirectX 12 (11_0).

For users who prioritize Vulkan performance, the NVIDIA GeForce 920M is the better option. It wins the Vulkan benchmark by 9.4%, and its average score of 3287, while lower, is still competitive with its nearest rivals. The dedicated 2 GB DDR3 memory with a fixed 14.40 GB/s bandwidth ensures consistent performance that does not depend on system memory. The NVIDIA part also has higher pixel rate, texture rate, and FP32 throughput, which could benefit certain workloads beyond the recorded benchmarks.

The database shows that the AMD part wins one benchmark and the NVIDIA part wins one benchmark. There is no overall winner in a strict sense. The choice depends entirely on the target application. If the system will run software that uses OpenCL for acceleration, the AMD Radeon R5 Graphics is the clear pick. If the system will use Vulkan for gaming or rendering, the NVIDIA GeForce 920M is preferable. The AMD part is better for compute per watt, while the NVIDIA part is better for raw graphics throughput and memory consistency.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon R5 Graphics has an average benchmark score of 3883, while the NVIDIA GeForce 920M has an average score of 3287. The AMD part also sits at the 23rd percentile among all GPUs, versus the 20th percentile for the NVIDIA part.

Q: How much faster is the AMD Radeon R5 Graphics in OpenCL?

A: The AMD part scores 5183 in Geekbench OpenCL, compared to 3725 for the NVIDIA GeForce 920M. This is a 39.1% advantage for the AMD Radeon R5 Graphics.

Q: Does the NVIDIA GeForce 920M win any benchmark?

A: Yes. The NVIDIA GeForce 920M scores 2849 in Geekbench Vulkan, while the AMD Radeon R5 Graphics scores 2582. The NVIDIA part leads by 9.4% in that test.

Q: What memory configurations do these GPUs use?

A: The AMD Radeon R5 Graphics uses system shared memory, with its size, type, bus width, and bandwidth all dependent on the host system. The NVIDIA GeForce 920M has 2 GB of dedicated DDR3 memory on a 64-bit bus, with a bandwidth of 14.40 GB/s.

Q: How do their power requirements compare?

A: The AMD Radeon R5 Graphics has a TDP of 15 W, while the NVIDIA GeForce 920M has a TDP of 33 W. The NVIDIA part uses no power connectors.

Q: Which GPU has better DirectX support?

A: The AMD Radeon R5 Graphics supports DirectX 12 (12_0), while the NVIDIA GeForce 920M supports DirectX 12 (11_0). The AMD part has a higher DirectX feature level.

Head-to-Head Benchmarks

The most significant result in the head-to-head comparison is the Geekbench OpenCL test. The AMD Radeon R5 Graphics scores 5183, and the NVIDIA GeForce 920M scores 3725. The delta is 39.1%, which is the largest performance gap between the two parts in any recorded test. This outcome is notable because the NVIDIA GeForce 920M has considerably higher peak specifications: 384 shading units versus 256, 32 TMUs versus 16, 8 ROPs versus 4, and an FP32 rating of 732.7 GFLOPS versus 388.1 GFLOPS. Despite having nearly double the raw compute throughput, the NVIDIA part is far behind in OpenCL. The likely explanation lies in the memory subsystem: the AMD part uses system shared memory, which can be faster in certain APU configurations when paired with dual-channel high-speed DDR3, while the NVIDIA part is limited to its 64-bit DDR3 interface at 14.40 GB/s. The OpenCL test may also be sensitive to driver scheduling and the way work items are dispatched, where GCN 2.0 has a proven track record.

The second head-to-head result is the Geekbench Vulkan test. Here, the NVIDIA GeForce 920M scores 2849, and the AMD Radeon R5 Graphics scores 2582. The delta is 9.4% in favor of NVIDIA. This is a much smaller margin than AMD's OpenCL win, but it is still a clear victory. The Vulkan API benefits from low-level control over the GPU, and NVIDIA's Kepler 2.0 architecture, with its higher texture rate (30.53 GTexel/s versus 12.13 GTexel/s) and pixel rate (7.632 GPixel/s versus 3.032 GPixel/s), appears to translate those throughput advantages into a measurable performance lead. The fixed 954 MHz clock of the NVIDIA part also ensures consistent operation, while the AMD part has no recorded clock frequency.

Looking at the nearest rivals in the database, the AMD Radeon R5 Graphics is closely matched with several NVIDIA Quadro workstation cards. The Quadro 2000 has an average score of 3898, which is 0.4% lower than the AMD part's 3883. The Quadro K2000D scores 3919, 0.9% higher than the AMD part, and the Quadro 2000D scores 3930, 1.2% higher. The GeForce MX110 scores 3834, 1.3% lower. This places the AMD IGP in a tight cluster of mid-range GPUs from a previous generation. The NVIDIA GeForce 920M, meanwhile, sits near the GeForce GT 730M (3316, 0.9% higher), the Intel HD Graphics 530 (3332, 1.4% higher), and the Intel HD Graphics P4600 (3389, 3% higher). The GeForce GT 640 is 2.4% lower at 3210. These comparisons show that both parts are in the lower quartile of GPU performance, but the AMD part has a slight edge in average score.

The production status for both parts is end-of-life. The AMD Radeon R5 Graphics was released on 2014-09-16, and its predecessor is TeraScale 3 IGP, with a successor of GCN 3.0 IGP. The NVIDIA GeForce 920M was released on 2015-03-12, with a predecessor of GeForce 800M and a successor of GeForce 10 Mobile. Neither part has a launch MSRP in the database, so no pricing information is available. The AMD part uses a 28 nm GlobalFoundries process, and the NVIDIA part uses a 28 nm TSMC process. The AMD die is 245 mm² with 2,410 million transistors, while the NVIDIA die is 87 mm² with 1,020 million transistors. The NVIDIA part has a higher transistor density at 11.7M per mm², versus 9.8M per mm² for the AMD part.

In practical terms, the head-to-head data tells a story of specialization. The AMD Radeon R5 Graphics is a compute-oriented IGP that can outperform a much larger discrete GPU in OpenCL workloads, despite having fewer shaders, fewer TMUs, fewer ROPs, and lower peak FP32. The NVIDIA GeForce 920M is a graphics-oriented chip that wins in Vulkan, likely due to its dedicated memory and higher fill rates. The database records one win for each part, and the final analysis must respect that split. For a system where OpenCL compute is the primary workload, the AMD part is the better choice. For a system where Vulkan rendering is the priority, the NVIDIA part is the better choice. There is no single GPU that dominates the other across all recorded tests.