AMD Radeon HD 8690M vs NVIDIA GeForce GTX 970 Comparison

The Verdict

The data in this comparison is unambiguous. The NVIDIA GeForce GTX 970 is in a completely different performance class than the AMD Radeon HD 8690M. Across the only shared benchmark in the database, Geekbench OpenCL, the GTX 970 scores 29982 against 6137 for the HD 8690M, a 388.5% advantage. The GTX 970 is the choice for any desktop user needing serious compute, modern API support, or high-resolution gaming. The HD 8690M, with its 36th percentile ranking versus the GTX 970's 39th, represents a much more modest capability tier, suitable only for lightweight tasks on an older mobile platform. The GTX 970's average benchmark score of 7157 places it within 0.9% of the GeForce GTX 750 and 0.6% behind the AMD Radeon Vega 8 Mobile, while the HD 8690M's average of 6137 puts it just 0.3% ahead of the Intel Iris Pro Graphics 6200 and 1% ahead of both the NVIDIA RTX A400 and GeForce MX230. These are entirely different product categories, and the verdict follows directly: choose the GTX 970 for any real workload, and only consider the HD 8690M if the platform constraint forces it.

FAQ

Q: How much faster is the NVIDIA GeForce GTX 970 than the AMD Radeon HD 8690M in the only directly comparable benchmark?

A: In Geekbench OpenCL, the GTX 970 scores 29982 versus 6137 for the HD 8690M, a 388.5% lead for the NVIDIA part.

Q: Which GPU has better API support for modern applications?

A: The GTX 970 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The HD 8690M supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The NVIDIA part has the more complete DirectX 12 feature level and a newer Vulkan version.

Q: What is the memory configuration difference between the two?

A: The GTX 970 has 4 GB of GDDR5 on a 256-bit bus with 224.4 GB/s bandwidth. The HD 8690M has 2 GB of GDDR5 on a 64-bit bus with 32.00 GB/s bandwidth.

Q: Are both GPUs based on the same manufacturing process?

A: Yes, both are fabricated on a 28 nm process at TSMC. However, the GTX 970 uses the GM204 chip with Maxwell 2.0 architecture, while the HD 8690M uses the Sun chip with GCN 1.0 architecture.

Q: Which GPU has the higher transistor count?

A: The GTX 970 has 5,200 million transistors on a 398 mm² die, while the HD 8690M has 690 million transistors on a 56 mm² die. The density is similar at 13.1M per mm² for the GTX 970 and 12.3M per mm² for the HD 8690M.

Q: What is the production status and release timing for these parts?

A: Both are listed as end-of-life products. The GTX 970 was released on 2014-09-18 and the HD 8690M was released on 2013-02-28.

Architecture Differences

The two GPUs come from different architectural generations and design philosophies. The NVIDIA GeForce GTX 970 is built on the Maxwell 2.0 architecture using the GM204 chip, while the AMD Radeon HD 8690M uses the GCN 1.0 architecture with the Sun chip. Both are manufactured at TSMC on a 28 nm process, but the similarity ends there. The GTX 970 packs 5,200 million transistors onto a 398 mm² die, achieving a transistor density of 13.1M per mm². The HD 8690M is a much smaller chip, with 690 million transistors on a 56 mm² die at a density of 12.3M per mm².

The compute resources differ by an order of magnitude. The GTX 970 contains 1664 shading units, 104 texture mapping units, and 56 raster operation pipelines. The HD 8690M has 320 shading units, 20 TMUs, and 8 ROPs. This fundamental imbalance explains the performance gap before any clock speed considerations are introduced. The GTX 970's pixel rate is 65.97 GPixel/s and its texture rate is 122.5 GTexel/s, while the HD 8690M manages 7.800 GPixel/s and 19.50 GTexel/s respectively. Floating point throughput follows the same pattern: the GTX 970 delivers 3.920 TFLOPS of FP32 compute versus 624.0 GFLOPS for the HD 8690M.

Neither GPU includes ray tracing cores or tensor cores, so the architectural comparison stays within the traditional rasterization and compute framework. The GTX 970 belongs to the GeForce 900 generation, succeeding the GeForce 700 and preceding the GeForce 10 series. The HD 8690M is part of the Solar System (HD 8600M) generation, with the London as its predecessor and the Gem System as its successor. The GTX 970 is a dual-slot desktop card measuring 267 mm in length, 111 mm in height, and 40 mm in width, while the HD 8690M is a mobile part with no dimensions recorded in the database.

Specification Differences

The clock speeds show a meaningful gap, though not as large as the compute resource disparity. The GTX 970 runs at a base clock of 1050 MHz with a boost clock of 1178 MHz, while the HD 8690M operates at 925 MHz base and 975 MHz boost. Memory clocks also favor the NVIDIA part: 1753 MHz (7 Gbps effective) for the GTX 970 versus 1000 MHz (4 Gbps effective) for the HD 8690M.

Memory capacity and bandwidth are heavily skewed toward the GTX 970. The NVIDIA card offers 4 GB of GDDR5 across a 256-bit bus, yielding 224.4 GB/s of bandwidth. The AMD part offers 2 GB of GDDR5 on a 64-bit bus, producing 32.00 GB/s. That is a sevenfold difference in memory bandwidth, which will matter in any bandwidth-sensitive workload.

The bus interface differs as well. The GTX 970 uses PCIe 3.0 x16, while the HD 8690M uses PCIe 3.0 x8. Power requirements are only recorded for the GTX 970: a 148 W TDP, dual-slot width, 2x 6-pin power connectors, and a suggested 300 W power supply. No TDP, slot width, power connector, or PSU recommendation is listed for the HD 8690M. The GTX 970 also has defined display outputs: 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2. The HD 8690M has no display output information in the database.

API support differs in two notable ways. The GTX 970 supports DirectX 12 (12_1) and Vulkan 1.4, while the HD 8690M supports DirectX 12 (11_1) and Vulkan 1.2.170. Both support OpenGL 4.6. The GTX 970 also has a launch MSRP of 329 USD. The HD 8690M has no launch MSRP recorded.

Head-to-Head Benchmarks

The database contains exactly one head-to-head benchmark result between these two GPUs: Geekbench OpenCL. The GTX 970 scores 29982, and the HD 8690M scores 6137. The delta is 388.5% in favor of the NVIDIA part. This is not a close contest by any measure. The GTX 970's OpenCL result is nearly five times higher than the AMD mobile part's result.

Looking at the broader benchmark profiles, the GTX 970 has a much richer set of recorded results. It has scores for 3DMark Steel Nomad DX12 at 369, Geekbench Metal at 13595, Geekbench Vulkan at 20005, and a suite of Passmark tests: DirectX 10 at 46, DirectX 11 at 71, DirectX 12 at 41, DirectX 9 at 143, G2D at 764, G3D at 9638, and GPU Compute at 4073. The HD 8690M only has the single Geekbench OpenCL result of 6137. This means the GTX 970's average benchmark score of 7157 is computed from a broad set of tests, while the HD 8690M's average of 6137 is identical to its only recorded score.

The percentile rankings are close in absolute terms but mask the performance gulf. The GTX 970 sits at the 39th percentile of all GPUs, and the HD 8690M sits at the 36th percentile. A three-percentile difference in a database that includes integrated and low-end parts compresses the spread, but the raw OpenCL numbers tell the real story. The GTX 970's nearest rivals by average score are the Intel Iris Pro Graphics P580 at 7170 (-0.2%), the NVIDIA GeForce GTX 560 SE at 7171 (-0.2%), the AMD Radeon Vega 8 Mobile at 7203 (-0.6%), and the NVIDIA GeForce GTX 750 at 7222 (-0.9%). The HD 8690M's nearest rivals include the Intel Iris Pro Graphics 6200 at 6117 (+0.3%), the NVIDIA RTX A400 at 6078 (+1%), the NVIDIA GeForce MX230 at 6077 (+1%), and the AMD Radeon HD 6950 at 6210 (-1.2%). The GTX 970 competes in the lower-midrange desktop space, while the HD 8690M sits among older mobile and entry-level parts.

Where Each One Wins

The GTX 970 wins every category where data exists. In raw compute, the 3.920 TFLOPS FP32 throughput against 624.0 GFLOPS makes it the clear choice for any OpenCL or compute-oriented workload. The Geekbench OpenCL result of 29982 versus 6137 confirms this in practice. Memory bandwidth of 224.4 GB/s versus 32.00 GB/s gives the GTX 970 a decisive edge in texture-heavy and large-dataset scenarios. The 4 GB frame buffer versus 2 GB means the GTX 970 can handle larger assets and higher resolutions without spilling to system memory.

For gaming, the GTX 970's DirectX 12 (12_1) support is a full feature level above the HD 8690M's DirectX 12 (11_1), which matters for modern titles that use advanced DX12 features. The Vulkan 1.4 support versus 1.2.170 also positions the GTX 970 better for current and future Vulkan-based games. The GTX 970's pixel rate of 65.97 GPixel/s and texture rate of 122.5 GTexel/s are roughly 8.5 times and 6.3 times the HD 8690M's respective rates, which translates directly to fill-rate-bound gaming performance.

The HD 8690M's only advantage is platform suitability. It is a mobile GPU, so it fits into laptops where a dual-slot, 267 mm, 148 W desktop card cannot be installed. The HD 8690M's smaller 56 mm² die and 690 million transistors indicate a low-power design, though no TDP is recorded in the database. Its 36th percentile ranking, while lower than the GTX 970's 39th, still places it above many integrated solutions, and its nearest rival comparison shows it outperforming the Intel Iris Pro Graphics 6200 by 0.3% and the NVIDIA MX230 by 1%. For a user constrained to a laptop from the HD 8600M generation, the HD 8690M is a serviceable part for light gaming and general compute. But in any direct comparison where both cards could physically be installed, the GTX 970 wins decisively on every measured metric.