AMD Radeon HD 8790M vs NVIDIA GeForce 940MX Comparison

Head-to-Head Benchmarks

The recorded data shows a clear, if uneven, victory for the AMD Radeon HD 8790M across the two benchmark tests in the database. In the Geekbench OpenCL test, the AMD part scores 5017 against 4939 for the NVIDIA GeForce 940MX, a delta of -1.6% from the NVIDIA card's perspective. That is a narrow margin, roughly within the noise of a single test run, and it indicates the two mobile GPUs are effectively peers when the workload is a generic compute task. The NVIDIA card's 512 shading units and higher FP32 rating of 881.7 GFLOPS do not translate into a win here, as the AMD card's 691.2 GFLOPS still manages to edge it out.

The Vulkan test tells a very different story. The AMD Radeon HD 8790M scores 6365, while the NVIDIA GeForce 940MX scores 4749. That is a delta of -25.4% for the NVIDIA part, a substantial gap that dwarfs the OpenCL result. The AMD card is more than a third faster in this specific API workload. Because Vulkan is a low-overhead, CPU-efficient graphics API, the result suggests the AMD architecture handles draw calls and command buffers more efficiently in this scenario, or that its driver implementation for this test is better optimized. Either way, the benchmark data is unambiguous: in Vulkan, the AMD part is the superior mobile GPU.

The average benchmark score reinforces the AMD card's position. The Radeon HD 8790M posts an average of 5691, while the GeForce 940MX averages 4844. That is a difference of roughly 17.5% in favor of the AMD card across the two tests. The percentile ranking also favors AMD: the HD 8790M sits at the 33rd percentile among all GPUs in the database, versus the 28th percentile for the 940MX. The NVIDIA card's nearest rivals include the GeForce GTX 560M (avg 4855, -0.2%), the Radeon R6 M255DX (avg 4867, -0.5%), and the GeForce GTS 450 (avg 4893, -1%). For the AMD card, the nearest rivals are the Intel Iris Pro Graphics P6300 (avg 5712, -0.4%), the GeForce GTX 670MX (avg 5721, -0.5%), and the GeForce GTX 550 Ti (avg 5731, -0.7%). The AMD card also sits 1.6% ahead of the Quadro M500M, which averages 5604. The data shows the HD 8790M competes in a slightly higher performance tier than the 940MX, whose closest neighbors are all older or lower-tier parts.

Where Each One Wins

Looking strictly at the benchmark results, the AMD Radeon HD 8790M wins both recorded tests. It takes the Geekbench OpenCL test by a small margin, 5017 to 4939, and it dominates the Geekbench Vulkan test, 6365 to 4749. There is no test in the database where the NVIDIA GeForce 940MX comes out ahead. The wins total is 0 for the NVIDIA card and 2 for the AMD card.

For use-case planning, the OpenCL result suggests that general-purpose compute workloads, such as OpenCL-accelerated image processing or physics calculations, will perform nearly identically on either GPU. The 1.6% delta is small enough that other system factors, memory bandwidth, driver version, or CPU speed, could easily flip the result in a real laptop. The Vulkan result, however, is decisive. Any application or game that uses the Vulkan API will run noticeably better on the Radeon HD 8790M. The 25.4% gap is large enough to affect playable frame rates at the same settings.

The NVIDIA part does have theoretical advantages in raw compute throughput. Its FP32 rating of 881.7 GFLOPS is higher than the AMD card's 691.2 GFLOPS, and its texture rate of 27.55 GTexel/s beats the AMD card's 21.60 GTexel/s. The pixel rates are close: 6.888 GPixel/s for the NVIDIA card versus 7.200 GPixel/s for the AMD card. Yet none of these theoretical numbers translate into a benchmark win in the database. The AMD card's wider 128-bit memory bus and 64.00 GB/s of bandwidth, versus the NVIDIA card's 64-bit bus and 40.10 GB/s, likely help it in the Vulkan test, where memory pressure from command buffers and scene data can be significant.

Architecture Differences

The two GPUs come from different design philosophies. The NVIDIA GeForce 940MX uses the GM107 chip built on the Maxwell architecture, a 28 nm design from TSMC with 1,870 million transistors on a 148 mm² die. The transistor density is 12.6 million per square millimeter. The AMD Radeon HD 8790M uses the Mars chip based on GCN 1.0, also on TSMC's 28 nm process, but with only 950 million transistors on a 77 mm² die, for a density of 12.3 million per square millimeter. Both are old, end-of-life mobile parts, but the NVIDIA chip is physically larger and packs nearly twice the transistor count.

The compute resources differ significantly. The NVIDIA card has 512 shading units, 32 texture mapping units, and 8 ROPs. The AMD card has 384 shading units, 24 TMUs, and 8 ROPs. Despite having more shading units and TMUs, the NVIDIA card loses the Vulkan test by a wide margin, which points to architectural efficiency differences rather than raw unit counts. The Maxwell architecture was designed for high efficiency per clock, but the GCN 1.0 architecture in the AMD card appears to handle the Vulkan workload more effectively in this comparison.

Memory configuration is a major differentiator. The NVIDIA card uses a 64-bit memory bus with 2 GB of GDDR5 running at 1253 MHz, effective 5 Gbps, yielding 40.10 GB/s of bandwidth. The AMD card also has 2 GB of GDDR5, but on a 128-bit bus running at 1000 MHz, effective 4 Gbps, for 64.00 GB/s. That is roughly 60% more memory bandwidth for the AMD card, a decisive advantage in bandwidth-sensitive workloads.

The clock speeds tell a mixed story. The NVIDIA card has a base clock of 795 MHz and a boost of 861 MHz. The AMD card runs higher, with a base of 850 MHz and a boost of 900 MHz. Memory clocks differ as well: the NVIDIA memory runs at 1253 MHz while the AMD memory runs at 1000 MHz, though the AMD part's wider bus more than compensates for the lower memory clock.

Both cards support DirectX 12, but with different feature levels. The NVIDIA card supports DirectX 12 (11_0), while the AMD card supports DirectX 12 (11_1). OpenGL support is identical at 4.6. Vulkan support differs: the NVIDIA card supports Vulkan 1.4, while the AMD card supports Vulkan 1.2.170. The NVIDIA card also has a higher TDP rating of 23 W, while the AMD card has no TDP listed in the database. Both use MXM modules, require no power connectors, and have display outputs that are portable-device dependent.

The NVIDIA card is from the GeForce 900M generation, released in 2016, with the GeForce 800M as its predecessor and GeForce 10 Mobile as its successor. The AMD card is from the Solar System (HD 8700M) generation, released in 2013, with London as its predecessor and Gem System as its successor. The NVIDIA card uses a PCIe 3.0 x8 bus interface, while the AMD card uses MXM-A (3.0).

FAQ

Q: Which GPU has higher memory bandwidth?

A: The AMD Radeon HD 8790M has 64.00 GB/s of bandwidth on a 128-bit bus, versus 40.10 GB/s on a 64-bit bus for the NVIDIA GeForce 940MX.

Q: Does the NVIDIA card have more shading units?

A: Yes, the NVIDIA GeForce 940MX has 512 shading units, while the AMD Radeon HD 8790M has 384 shading units.

Q: Which card performs better in the Vulkan benchmark?

A: The AMD Radeon HD 8790M scores 6365 in Geekbench Vulkan, which is 25.4% ahead of the NVIDIA GeForce 940MX's 4749.

Q: Are both cards end-of-life products?

A: Yes, both the NVIDIA GeForce 940MX and the AMD Radeon HD 8790M are marked as end-of-life in the database.

Q: What process node are both GPUs built on?

A: Both are built on TSMC's 28 nm process. The NVIDIA chip has 1,870 million transistors on a 148 mm² die, and the AMD chip has 950 million transistors on a 77 mm² die.

Q: Do both cards support the same DirectX version?

A: No. The AMD Radeon HD 8790M supports DirectX 12 (11_1), while the NVIDIA GeForce 940MX supports DirectX 12 (11_0).

Specification Differences

| Specification | NVIDIA GeForce 940MX | AMD Radeon HD 8790M |

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

| Architecture | Maxwell | GCN 1.0 |

| Chip | GM107 | Mars |

| Generation | GeForce 900M | Solar System (HD 8700M) |

| Transistors | 1,870 million | 950 million |

| Die Size | 148 mm² | 77 mm² |

| Transistor Density | 12.6M / mm² | 12.3M / mm² |

| Base Clock | 795 MHz | 850 MHz |

| Boost Clock | 861 MHz | 900 MHz |

| Memory Clock | 1253 MHz, 5 Gbps effective | 1000 MHz, 4 Gbps effective |

| Memory Bus Width | 64 bit | 128 bit |

| Memory Bandwidth | 40.10 GB/s | 64.00 GB/s |

| Shading Units | 512 | 384 |

| TMUs | 32 | 24 |

| ROPs | 8 | 8 |

| Pixel Rate | 6.888 GPixel/s | 7.200 GPixel/s |

| Texture Rate | 27.55 GTexel/s | 21.60 GTexel/s |

| FP32 | 881.7 GFLOPS | 691.2 GFLOPS |

| TDP | 23 W | Not listed |

| Bus Interface | PCIe 3.0 x8 | MXM-A (3.0) |

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

| Vulkan Support | 1.4 | 1.2.170 |

| Release Date | 2016-06-27 | 2013-03-31 |

| Predecessor | GeForce 800M | London |

| Successor | GeForce 10 Mobile | Gem System |

| Percentile vs All GPUs | 28 | 33 |

| Average Benchmark Score | 4844 | 5691 |

The Verdict

The benchmark data is clear: the AMD Radeon HD 8790M is the faster GPU in this head-to-head. It wins both recorded tests, with an average score of 5691 against the NVIDIA GeForce 940MX's 4844, and it holds a 33rd percentile ranking versus the NVIDIA card's 28th percentile. The Vulkan result is especially lopsided, with the AMD card leading by 25.4% in that test. The OpenCL result is close enough to call it a tie for practical purposes, but the overall trend favors the AMD card.

Buyers who prioritize Vulkan-based gaming or modern low-overhead API performance should choose the AMD Radeon HD 8790M without hesitation. The 64.00 GB/s of memory bandwidth on the 128-bit bus gives it a structural advantage that the NVIDIA card's higher shading unit count and FP32 rating cannot overcome in the recorded benchmarks.

The NVIDIA GeForce 940MX is not without merit. Its higher transistor count of 1,870 million on the GM107 chip and higher texture rate of 27.55 GTexel/s suggest it could hold its own in texture-bound scenarios, and its 23 W TDP indicates low power draw. But the database contains no test where it beats the AMD card. For anyone sourcing an MXM module for a laptop, the data points to the Radeon HD 8790M as the better performer, particularly in Vulkan workloads. The NVIDIA card is a reasonable fallback for legacy DirectX 11_0 environments, but the AMD part's DirectX 12 (11_1) support and superior bandwidth make it the stronger overall choice from the recorded measurements.