AMD Radeon HD 8690M vs NVIDIA GeForce 840M Comparison

Head-to-Head Benchmarks

The recorded data includes a single direct comparison between the AMD Radeon HD 8690M and the NVIDIA GeForce 840M, and it is a close contest. In the Geekbench OpenCL test, the AMD Radeon HD 8690M scores 6137, while the NVIDIA GeForce 840M scores 5764. This gives the AMD part a 6.5% advantage, a meaningful but not overwhelming lead in raw compute throughput.

Looking at the broader database context, the AMD Radeon HD 8690M sits at the 36th percentile among all GPUs, while the NVIDIA GeForce 840M sits at the 31st percentile. That percentile gap reflects the OpenCL result, but the NVIDIA part has an additional data point: a Geekbench Vulkan score of 4880, which the AMD part does not have recorded. That Vulkan result is not directly comparable to the AMD part, but it does show the 840M has a functional Vulkan path that the database has measured.

The AMD Radeon HD 8690M's nearest rivals in the database include the Intel Iris Pro Graphics 6200 (average score 6117, 0.3% behind), the NVIDIA RTX A400 (6078, 1% behind), and the NVIDIA GeForce MX230 (6077, 1% behind). The only rival ahead of it is the AMD Radeon HD 6950 with an average score of 6210, which is 1.2% faster. This places the 8690M in a tight cluster where a few percentage points separate it from several other mobile and entry-level parts.

The NVIDIA GeForce 840M's nearest rivals tell a similar story of close competition. The NVIDIA GeForce 930A averages 5317 (0.1% behind), the NVIDIA GeForce GTX 980M averages 5308 (0.3% behind), and the NVIDIA GeForce 940M averages 5284 (0.7% behind). The AMD Radeon R7 M445 leads this group with an average of 5358, which is 0.7% ahead of the 840M. Notably, the GTX 980M appears in this rival list with an average score close to the 840M, which is surprising given the positioning of those two products, but the database records what it records.

The head-to-head result is the only direct benchmark comparison available, and it shows the AMD Radeon HD 8690M winning by 6.5%. That is a solid margin for a single test, but the absence of a Vulkan result for the AMD part means the full picture is incomplete. The 840M's Vulkan score of 4880 suggests it can handle that API workload, but without a comparable AMD figure, no direct verdict can be drawn.

Where Each One Wins

The AMD Radeon HD 8690M wins the only recorded head-to-head benchmark, the Geekbench OpenCL test, by 6.5%. This indicates that in OpenCL compute workloads, the 8690M holds an edge over the 840M. For users relying on OpenCL acceleration in applications such as video encoding, image processing, or general GPU compute tasks, the data favors the AMD part.

The NVIDIA GeForce 840M does not win any recorded head-to-head benchmark, because only one such benchmark exists. However, the 840M has a Vulkan score of 4880 in the database, which suggests it can handle Vulkan-based workloads. The AMD Radeon HD 8690M has no Vulkan score recorded, so for Vulkan-specific applications, the database shows the NVIDIA part has verified performance while the AMD part has none. That is not a direct comparison, but it is a meaningful distinction: one product has a measured Vulkan result, the other does not.

The average benchmark score also separates the two. The AMD Radeon HD 8690M has an average benchmark score of 6137, while the NVIDIA GeForce 840M averages 5322. That is a difference of roughly 15%, which is larger than the 6.5% head-to-head delta because the 840M's Vulkan score drags its average down. The 840M's OpenCL score is 5764, but its Vulkan score is 4880, and the average blends both.

For use cases, the AMD part is the stronger choice for OpenCL compute. The NVIDIA part is the only one with a recorded Vulkan result, so for Vulkan rendering or compute, the 840M has demonstrated capability. The 840M also carries a 33 W TDP, while the AMD part has no TDP listed in the database, so power consumption comparisons cannot be made from the available data.

FAQ

Q: Which GPU wins the Geekbench OpenCL benchmark?

A: The AMD Radeon HD 8690M scores 6137, while the NVIDIA GeForce 840M scores 5764, giving the AMD part a 6.5% lead.

Q: Does the NVIDIA GeForce 840M have any benchmark results that the AMD Radeon HD 8690M lacks?

A: Yes, the 840M has a Geekbench Vulkan score of 4880. The AMD Radeon HD 8690M has no Vulkan score recorded in the database.

Q: How do the two GPUs compare in average benchmark score?

A: The AMD Radeon HD 8690M averages 6137, while the NVIDIA GeForce 840M averages 5322. The AMD part is higher by a significant margin.

Q: What are the nearest rivals to the AMD Radeon HD 8690M?

A: The nearest rivals are the Intel Iris Pro Graphics 6200 (6117, 0.3% behind), the NVIDIA RTX A400 (6078, 1% behind), the NVIDIA GeForce MX230 (6077, 1% behind), and the AMD Radeon HD 6950 (6210, 1.2% ahead).

Q: What are the nearest rivals to the NVIDIA GeForce 840M?

A: The nearest rivals are the NVIDIA GeForce 930A (5317, 0.1% behind), the NVIDIA GeForce GTX 980M (5308, 0.3% behind), the AMD Radeon R7 M445 (5358, 0.7% ahead), and the NVIDIA GeForce 940M (5284, 0.7% behind).

Q: What is the percentile ranking of each GPU?

A: The AMD Radeon HD 8690M is at the 36th percentile among all GPUs, while the NVIDIA GeForce 840M is at the 31st percentile.

Specification Differences

The two GPUs differ in several core specifications. The AMD Radeon HD 8690M uses a Sun chip with the GCN 1.0 architecture and belongs to the Solar System generation (HD 8600M). The NVIDIA GeForce 840M uses a GM108S chip with the Maxwell architecture and belongs to the GeForce 800M generation.

Clock speeds differ notably. The AMD part has a base clock of 925 MHz and a boost clock of 975 MHz. The NVIDIA part has a base clock of 1029 MHz and a boost clock of 1124 MHz, making it faster in both metrics.

Memory configuration also differs. Both have 2 GB of memory and a 64 bit bus width, but the AMD Radeon HD 8690M uses GDDR5 memory at 1000 MHz with 4 Gbps effective speed, delivering 32.00 GB/s of bandwidth. The NVIDIA GeForce 840M uses DDR3 memory at 1001 MHz with 2 Gbps effective speed, delivering only 16.02 GB/s of bandwidth. The AMD part has exactly twice the memory bandwidth.

The compute units differ as well. The AMD Radeon HD 8690M has 320 shading units, 20 texture mapping units, and 8 ROPs. The NVIDIA GeForce 840M has 384 shading units, 16 texture mapping units, and 8 ROPs. The NVIDIA part has more shading units but fewer TMUs.

Pixel and texture rates reflect these differences. The AMD part delivers 7.800 GPixel/s and 19.50 GTexel/s. The NVIDIA part delivers 8.992 GPixel/s and 17.98 GTexel/s. The NVIDIA part has a higher pixel rate, while the AMD part has a higher texture rate.

FP32 performance also differs. The AMD Radeon HD 8690M delivers 624.0 GFLOPS, while the NVIDIA GeForce 840M delivers 863.2 GFLOPS, a substantial advantage for the NVIDIA part despite its lower benchmark score.

The NVIDIA part has a TDP of 33 W, while the AMD part has no TDP listed. The NVIDIA part is described as an IGP with no power connectors and portable device dependent display outputs. The AMD part has no such fields recorded.

Both use a PCIe 3.0 x8 bus interface. 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 Vulkan version difference is notable: the NVIDIA part supports a newer Vulkan revision.

Release dates differ. The AMD Radeon HD 8690M was released on 2013-02-28, while the NVIDIA GeForce 840M was released on 2014-03-11. Both are end-of-life products. The AMD part's predecessor is London and its successor is Gem System. The NVIDIA part's predecessor is the GeForce 700M and its successor is the GeForce 900M.

Architecture Differences

The architectural split is clear. The AMD Radeon HD 8690M uses GCN 1.0, AMD's Graphics Core Next architecture, built on a Sun chip. The NVIDIA GeForce 840M uses Maxwell, built on a GM108S chip. Both are manufactured on a 28 nm process at TSMC, but the transistor counts and die sizes differ.

The AMD part has 690 million transistors on a 56 mm² die, giving a transistor density of 12.3M per mm². The NVIDIA part has 1,020 million transistors on a 77 mm² die, giving a transistor density of 13.2M per mm². The NVIDIA chip is larger in both transistor count and die size, and it packs transistors more densely.

Memory architecture differs significantly. The AMD Radeon HD 8690M uses GDDR5 memory with a 64 bit bus and 32.00 GB/s bandwidth. The NVIDIA GeForce 840M uses DDR3 memory with the same 64 bit bus but only 16.02 GB/s bandwidth. This is a 2x difference in memory bandwidth, which likely affects memory-bound workloads despite the NVIDIA part's higher raw compute rates.

The shading and texture configuration reflects different design priorities. The AMD part has 320 shading units and 20 TMUs. The NVIDIA part has 384 shading units and 16 TMUs. The NVIDIA part has 20% more shading units, while the AMD part has 25% more texture units. Both have 8 ROPs.

Clock behavior also differs. The AMD part runs at 925 MHz base and 975 MHz boost. The NVIDIA part runs at 1029 MHz base and 1124 MHz boost. The NVIDIA part's higher clocks contribute to its higher FP32 throughput of 863.2 GFLOPS versus 624.0 GFLOPS for the AMD part.

The API support has a notable difference in Vulkan versions. The AMD part supports Vulkan 1.2.170, while the NVIDIA part supports Vulkan 1.4. Both support DirectX 12, but with different feature levels: the AMD part lists 12 (11_1) while the NVIDIA part lists 12 (11_0). OpenGL support is identical at 4.6 for both.

The NVIDIA part includes a TDP of 33 W and is classified as an IGP with no power connectors. The AMD part has no TDP, slot width, or power connector data recorded. This makes thermal and power comparisons impossible from the available facts, but the NVIDIA part's listed 33 W gives it a documented power envelope.

In summary, the architecture story is one of competing designs: AMD's GCN 1.0 with faster memory and more texture units versus NVIDIA's Maxwell with more shading units, higher clocks, and a newer Vulkan implementation. The benchmark data favors the AMD part in OpenCL, but the NVIDIA part has a recorded Vulkan path and a much higher FP32 rating.