AMD Radeon HD 8690M vs NVIDIA GeForce 940M Comparison

Head-to-Head Benchmarks

The only direct benchmark comparison recorded in the database is the Geekbench OpenCL test. In that test, the AMD Radeon HD 8690M scores 6137, while the NVIDIA GeForce 940M scores 6018. The AMD part wins by a 2% margin. This is a narrow victory, and it places the two GPUs in essentially the same performance tier for OpenCL compute workloads.

The AMD Radeon HD 8690M posts an average benchmark score of 6137, which places it at the 36th percentile among all GPUs in the database. Its nearest rivals tell a consistent story. The Intel Iris Pro Graphics 6200 scores 6117, a 0.3% deficit to the AMD part. The NVIDIA RTX A400 scores 6078, a 1% deficit. The NVIDIA GeForce MX230 also scores 6077, again a 1% deficit. The only rival ahead of it in the recorded set is the AMD Radeon HD 6950, which scores 6210, a 1.2% margin over the HD 8690M. The HD 8690M is therefore clustered tightly among several mid-range parts, with no rival in its nearest group separating itself by more than 1.2%.

The NVIDIA GeForce 940M presents a more complicated picture because the database records two benchmark results for it. Its Geekbench OpenCL score is 6018, and its Geekbench Vulkan score is 4549. Its average benchmark score across both tests is 5284, which places it at the 31st percentile among all GPUs. That is 5 percentile points lower than the AMD part. The average is pulled down substantially by the Vulkan result, which is far below the OpenCL result. The nearest rivals for the 940M, based on its average score, include the NVIDIA GeForce GTX 980M at 5308, a 0.4% deficit to the 940M; the NVIDIA GeForce 930A at 5317, a 0.6% deficit; the NVIDIA GeForce 840M at 5322, a 0.7% deficit; and the NVIDIA GeForce GTX 760M at 5236, a 0.9% margin for the 940M. These margins are all under 1%, meaning the 940M sits in a very dense performance cluster when averaged across its two recorded tests.

The single head-to-head OpenCL result favors the AMD part, but the margin is small enough that it falls within typical run-to-run variation. The data shows a 2% delta, which is real but not transformative. For compute workloads that stress OpenCL performance specifically, the HD 8690M holds a measurable edge. For Vulkan workloads, only the 940M has a recorded score, and that score is 4549, which is 24% below its own OpenCL result. The AMD part has no recorded Vulkan score in the database, so no direct comparison is possible for that API.

The bigger story in the numbers is the divergence between the two GPUs' average scores and their percentile rankings. The HD 8690M, with a single benchmark score of 6137, sits at the 36th percentile. The 940M, with an average of 5284, sits at the 31st percentile. If the 940M's OpenCL score of 6018 were its only recorded result, its percentile placement would likely be much closer to the AMD part. The Vulkan score is what drags its average down. This suggests that the 940M's standing in the database is heavily influenced by API-specific performance, not by raw compute capability alone.

The Verdict

From the recorded data, the AMD Radeon HD 8690M is the stronger choice for OpenCL compute workloads. It wins the only direct head-to-head benchmark, and its single-score average of 6137 is 2% higher than the 940M's OpenCL result. It also ranks higher in the overall database, at the 36th percentile versus the 31st percentile. For users whose primary workload is OpenCL-based, the HD 8690M is the data-backed pick.

The NVIDIA GeForce 940M, however, offers something the AMD part does not: a recorded Vulkan score. The 940M's Vulkan result of 4549 demonstrates that it can execute Vulkan workloads, while the database contains no Vulkan score for the HD 8690M. For software that relies on Vulkan, the 940M is the only one of the two with verified performance data. The HD 8690M does support Vulkan 1.2.170 according to its API list, but no benchmark score confirms its actual Vulkan performance.

The 940M also has a much higher average benchmark score if one ignores the Vulkan result and looks only at OpenCL, but that is not how the database records its average. The average of 5284 reflects both tests, and it is 853 points lower than the HD 8690M's average of 6137. That is a 14% gap in average scores. The HD 8690M is therefore the safer choice for general compute performance as measured by the database.

For gaming or graphics workloads, the data does not provide direct frame-rate comparisons. The database records only OpenCL and Vulkan compute results. The pixel rate and texture rate figures favor the 940M heavily, but those are theoretical specifications, not measured game performance. The 940M's pixel rate of 17.57 GPixel/s is more than double the HD 8690M's 7.800 GPixel/s, and its texture rate of 35.14 GTexel/s is 80% higher than the HD 8690M's 19.50 GTexel/s. Those figures suggest the 940M has more raw rasterization and texturing throughput, which could translate to better graphics performance in fill-rate-bound scenarios, but the database does not record a benchmark that directly confirms this.

The verdict splits cleanly. Choose the AMD Radeon HD 8690M for OpenCL compute and for a higher overall database percentile. Choose the NVIDIA GeForce 940M for Vulkan compute workloads and for the higher theoretical pixel and texture throughput.

FAQ

Q: Which GPU has the higher Geekbench OpenCL score?

A: The AMD Radeon HD 8690M scores 6137, while the NVIDIA GeForce 940M scores 6018. The AMD part wins by 2%.

Q: Does the NVIDIA GeForce 940M have any benchmark advantage over the AMD Radeon HD 8690M?

A: The 940M has a recorded Geekbench Vulkan score of 4549. The HD 8690M has no Vulkan score in the database, so the 940M is the only one of the two with verified Vulkan performance data.

Q: How do the two GPUs compare in overall database percentile ranking?

A: The AMD Radeon HD 8690M sits at the 36th percentile among all GPUs, while the NVIDIA GeForce 940M sits at the 31st percentile.

Q: What is the average benchmark score for each GPU?

A: The AMD Radeon HD 8690M has an average benchmark score of 6137, based on a single OpenCL test. The NVIDIA GeForce 940M has an average benchmark score of 5284, based on OpenCL and Vulkan tests.

Q: Which GPU has higher pixel rate and texture rate specifications?

A: The NVIDIA GeForce 940M has a pixel rate of 17.57 GPixel/s and a texture rate of 35.14 GTexel/s. The AMD Radeon HD 8690M has a pixel rate of 7.800 GPixel/s and a texture rate of 19.50 GTexel/s.

Q: How do the nearest rivals compare for each GPU?

A: The AMD Radeon HD 8690M's closest rival is the Intel Iris Pro Graphics 6200, which scores 6117, a 0.3% deficit. The NVIDIA GeForce 940M's closest rival is the NVIDIA GeForce GTX 980M, which scores 5308, a 0.4% deficit to the 940M.

Specification Differences

The two GPUs differ across nearly every recorded specification. The AMD Radeon HD 8690M has a base clock of 925 MHz and a boost clock of 975 MHz. The NVIDIA GeForce 940M has a base clock of 1020 MHz and a boost clock of 1098 MHz. The NVIDIA part runs 95 MHz higher at base and 123 MHz higher at boost.

Memory configuration differs sharply. The AMD part uses 2 GB of GDDR5 memory on a 64-bit bus, with a memory clock of 1000 MHz and 4 Gbps effective, yielding a bandwidth of 32.00 GB/s. The NVIDIA part uses 2 GB of DDR3 memory on a 64-bit bus, with a memory clock of 900 MHz and 1800 Mbps effective, yielding a bandwidth of 14.40 GB/s. The AMD part has more than double the memory bandwidth.

Compute unit counts also differ. The AMD Radeon HD 8690M has 320 shading units, 20 texture mapping units, and 8 raster operation units. The NVIDIA GeForce 940M has 512 shading units, 32 TMUs, and 16 ROPs. The NVIDIA part has 60% more shading units, 60% more TMUs, and double the ROPs. Its FP32 throughput is 1,124.4 GFLOPS, versus 624.0 GFLOPS for the AMD part. Its pixel rate is 17.57 GPixel/s versus 7.800 GPixel/s, and its texture rate is 35.14 GTexel/s versus 19.50 GTexel/s.

The bus interface differs. The AMD part uses PCIe 3.0 x8. The NVIDIA part uses MXM-B (3.0). The NVIDIA part has a recorded TDP of 75 W, a slot width of MXM Module, and no power connectors. The AMD part has no recorded TDP, slot width, or power connector data. The NVIDIA part lists its display outputs as "Portable Device Dependent", while the AMD part has no recorded display output data.

The API support differs slightly. The AMD part supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The NVIDIA part supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. The NVIDIA part has a newer Vulkan version, while the AMD part has a newer DirectX feature level.

Architecture Differences

The two GPUs come from different manufacturers and different architectural generations. The AMD Radeon HD 8690M is built on the GCN 1.0 architecture, using the Sun chip. It belongs to the Solar System generation (HD 8600M series). The NVIDIA GeForce 940M is built on the Maxwell architecture, using the GM107 chip. It belongs to the GeForce 900M generation.

Both are fabricated by TSMC on a 28 nm process node. The transistor counts differ enormously. The AMD chip contains 690 million transistors on a die size of 56 mm², giving a transistor density of 12.3 million transistors per mm². The NVIDIA chip contains 1,870 million transistors on a die size of 148 mm², giving a transistor density of 12.6 million transistors per mm². The NVIDIA chip has 2.7 times as many transistors and a die more than 2.6 times larger, but the density is nearly identical.

The release dates are nearly two years apart. The AMD Radeon HD 8690M was released on 2013-02-28, while the NVIDIA GeForce 940M was released on 2015-03-12. Both are now end-of-life products. The AMD part's predecessor is listed as London and its successor as Gem System. The NVIDIA part's predecessor is GeForce 800M and its successor is GeForce 10 Mobile.

Neither GPU has ray tracing cores or tensor cores. Neither has a recorded FP16 performance figure. The architectural differences are therefore concentrated in the compute layout: the NVIDIA Maxwell part has more shading units, TMUs, and ROPs, while the AMD GCN part has a narrower memory bus but faster memory technology. The Maxwell architecture's higher clock speeds and wider compute configuration give it the theoretical throughput advantage, while the GCN part's GDDR5 memory gives it the bandwidth advantage. The benchmark data shows that, at least in OpenCL compute, the bandwidth advantage translates into a small but measurable win for the AMD part.