AMD Radeon HD 8690M vs AMD Radeon RX 6400 Comparison

The Verdict

The benchmark data presents an unambiguous outcome: the AMD Radeon RX 6400 is the superior GPU in every measurable category. The HD 8690M, a 2013-era mobile part, simply cannot compete with the architectural and manufacturing advancements of the RX 6400. The data shows the RX 6400 scoring 32,011 in Geekbench OpenCL, a figure that is 80.8% higher than the HD 8690M's 6,137. This is not a close contest; it is a generational chasm.

For anyone choosing between these two, the RX 6400 is the logical pick for virtually any task involving 3D graphics, compute, or modern API support. The HD 8690M, with its 36th percentile ranking among all GPUs, is strictly a legacy part. Its only conceivable advantage lies in its lower power draw, though the FACT PACK does not list a TDP for it, making a direct comparison impossible. The RX 6400, with a 53 W TDP and a single-slot design, is already extremely efficient. The verdict from the data is simple: do not choose the HD 8690M for any new or demanding workload. It is a relic. The RX 6400, despite its own end-of-life status, remains a functional entry-level card that can handle modern APIs and deliver playable performance in less demanding titles.

Architecture Differences

The two GPUs represent vastly different eras of AMD design. The HD 8690M is built on the Graphics Core Next (GCN) 1.0 architecture, using the "Sun" chip, and was manufactured on a 28 nm process at TSMC. It packs 690 million transistors onto a 56 mm² die, yielding a transistor density of 12.3 million per square millimeter. In contrast, the RX 6400 is based on the RDNA 2.0 architecture, using the "Navi 24" chip, and is fabricated on a much more advanced 6 nm process, also at TSMC. This newer process allows for 5,400 million transistors on a 107 mm² die, achieving a density of 50.5 million per square millimeter. This is a 4.1x increase in density, which is the primary driver of the performance difference.

The core configurations diverge significantly. The HD 8690M has 320 shading units, 20 texture mapping units (TMUs), and 8 raster operation units (ROPs). The RX 6400, on the other hand, features 768 shading units, 48 TMUs, and 32 ROPs. This represents a 2.4x increase in shading units, a 2.4x increase in TMUs, and a 4x increase in ROPs. Furthermore, the RX 6400 is equipped with 12 dedicated ray tracing cores, a feature entirely absent from the older GCN part. The RX 6400 also supports half-precision FP16 compute at a 2:1 ratio, with a rating of 7.130 TFLOPS, while the HD 8690M's FP16 capability is not listed.

The memory subsystems are also from different generations. The HD 8690M uses 2 GB of GDDR5 memory on a 64-bit bus, providing 32.00 GB/s bandwidth. The RX 6400 uses 4 GB of GDDR6 memory, also on a 64-bit bus, but with a much higher effective speed of 16 Gbps, resulting in 128.0 GB/s bandwidth. This is a 4x increase in raw bandwidth. The RX 6400 also supports newer PCIe 4.0 x4 interface, while the HD 8690M is limited to PCIe 3.0 x8. Finally, the API support tells the story of the generational leap: the HD 8690M supports DirectX 12 (11_1), while the RX 6400 supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, versus the older card's Vulkan 1.2.170.

Head-to-Head Benchmarks

The head-to-head benchmark data is sparse but decisive. The only shared test is Geekbench OpenCL, where the RX 6400 scores 32,011 against the HD 8690M's 6,137. This is a delta of -80.8% for the HD 8690M, meaning the RX 6400 is approximately 5.2 times faster in this compute workload. This single result encapsulates the entire performance relationship between the two cards. The RX 6400's score is not just higher; it is in a completely different performance class.

While the HD 8690M wins zero benchmarks in the head-to-head comparison, its individual benchmark score reveals its relative standing. Its Geekbench OpenCL score of 6,137 places it in the 36th percentile among all GPUs. The nearest rivals to the HD 8690M are the Intel Iris Pro Graphics 6200 (6,117, delta 0.3%), NVIDIA RTX A400 (6,078, delta 1%), and NVIDIA GeForce MX230 (6,077, delta 1%). These are all integrated or low-end discrete parts. The RX 6400, with its average benchmark score of 6,001, sits in the 35th percentile, with nearest rivals like the NVIDIA GeForce GTX 770M (6,000, delta 0%) and AMD FirePro W4100 (5,987, delta 0.2%). This is a fascinating data point: the average benchmark scores across all tests are nearly identical between the two (6,137 vs. 6,001), but the RX 6400's OpenCL score is dramatically higher. This suggests the HD 8690M performs relatively better in legacy tests, while the RX 6400 excels in modern compute tasks.

Specification Differences

The specifications that differ between the two cards are extensive and highlight the technological gap. The HD 8690M uses a 28 nm process, while the RX 6400 uses a 6 nm process. The HD 8690M has 690 million transistors on a 56 mm² die, while the RX 6400 has 5,400 million on a 107 mm² die. Clock speeds are a major differentiator: the HD 8690M has a base clock of 925 MHz and a boost of 975 MHz, while the RX 6400 has a base of 1923 MHz and a boost of 2321 MHz. The RX 6400 also has a defined game clock of 2039 MHz.

Memory specifications are starkly different: 2 GB GDDR5 vs. 4 GB GDDR6. While both use a 64-bit bus, the effective memory speed of the RX 6400 is 16 Gbps versus the HD 8690M's 4 Gbps, resulting in 128.0 GB/s vs. 32.00 GB/s bandwidth. The compute units are also different: 320 shading units, 20 TMUs, and 8 ROPs versus 768, 48, and 32, respectively. The RX 6400 adds 12 ray tracing cores, a feature the HD 8690M lacks. Pixel rate is 7.800 GPixel/s for the old card and 74.27 GPixel/s for the new one. Texture rate is 19.50 GTexel/s vs. 111.4 GTexel/s. FP32 compute is 624.0 GFLOPS vs. 3.565 TFLOPS.

The RX 6400 has a TDP of 53 W, a single-slot form factor, no power connectors, and a suggested PSU of 250 W. The HD 8690M has no TDP listed. The bus interface is PCIe 3.0 x8 for the old card and PCIe 4.0 x4 for the new one. Display outputs are listed only for the RX 6400: 1x HDMI 2.1 and 1x DisplayPort 1.4a. The RX 6400 is part of the Radeon RX 6000 series, while the HD 8690M is from the Solar System (HD 8600M) generation. The RX 6400 has a launch MSRP of 159 USD.

FAQ

Q: Is the AMD Radeon RX 6400 faster than the HD 8690M in compute workloads?

A: Yes. In the Geekbench OpenCL benchmark, the RX 6400 scores 32,011, which is 80.8% higher than the HD 8690M's score of 6,137.

Q: Which GPU has more memory bandwidth?

A: The RX 6400 has significantly more bandwidth at 128.0 GB/s, compared to the HD 8690M's 32.00 GB/s. This is due to the RX 6400's faster GDDR6 memory.

Q: Does the HD 8690M support ray tracing?

A: No. The HD 8690M has no ray tracing cores. The RX 6400 has 12 dedicated RT cores.

Q: What are the architecture names for these two GPUs?

A: The HD 8690M is based on the GCN 1.0 architecture, while the RX 6400 is based on the RDNA 2.0 architecture.

Q: Which GPU has a higher boost clock?

A: The RX 6400 has a boost clock of 2321 MHz, which is significantly higher than the HD 8690M's boost clock of 975 MHz.

Q: Is the RX 6400's transistor density higher than the HD 8690M's?

A: Yes. The RX 6400 has a transistor density of 50.5 million per mm², while the HD 8690M has a density of 12.3 million per mm².

Where Each One Wins

The data shows a clear split, but it is not a balanced one. The RX 6400 wins in every possible category. It wins in raw compute, as evidenced by the Geekbench OpenCL score. It wins in memory bandwidth, with 128.0 GB/s versus 32.00 GB/s. It wins in pixel and texture fill rates, with 74.27 GPixel/s and 111.4 GTexel/s versus 7.800 GPixel/s and 19.50 GTexel/s, respectively. It wins in FP32 compute, offering 3.565 TFLOPS versus 624.0 GFLOPS. It wins in API support, offering DirectX 12 Ultimate and Vulkan 1.4. It wins in memory capacity, with 4 GB versus 2 GB. It even wins in clock speeds, with a boost of 2321 MHz versus 975 MHz.

The HD 8690M has no benchmark wins to its name. Its only potential "win" is in the field of legacy compatibility, as it is from an older generation and may be present in older systems. However, from a pure performance standpoint, the data does not provide a single category where it excels. The RX 6400 is the clear winner for gaming, compute, and any modern workload. The HD 8690M's 36th percentile ranking, coupled with its single benchmark score of 6,137, places it firmly in the bottom tier of GPUs. The RX 6400, despite its own low percentile ranking (35th), possesses the architectural features and raw performance to handle tasks that would be impossible on the HD 8690M. The use-case split is simple: choose the RX 6400 for everything. The HD 8690M is only relevant as a historical artifact or a functional part in a system that requires no graphical performance beyond basic display output.