AMD Radeon HD 6950 vs NVIDIA GeForce GTX 980M Comparison

Where Each One Wins

The recorded data shows a starkly one-sided comparison. The NVIDIA GeForce GTX 980M wins the only head-to-head benchmark available in the database, taking the Geekbench OpenCL test with a score of 23832 against the AMD Radeon HD 6950's 6210. That is a delta of 73.9% in favor of the NVIDIA part. The AMD Radeon HD 6950 has zero wins in the direct comparison, while the GTX 980M has one.

The AMD Radeon HD 6950's case rests entirely on its position among its nearest rivals. Its Geekbench OpenCL score of 6210 places it within 1.1% of the NVIDIA Quadro K620 (6282), 0.9% behind the NVIDIA GeForce RTX 5070 Ti SUPER (6270) and NVIDIA GeForce RTX 4070 Ti SUPER AD102 (6270), and 0.2% behind the AMD FirePro W600 (6223). These deltas are all within a narrow band, meaning the HD 6950 is effectively competitive with that cluster in this single test.

The GTX 980M, by contrast, sits in a different performance tier. Its average benchmark score across all recorded tests is 5308, which is lower than the HD 6950's average of 6210, but that average pulls in many older DirectX tests where the mobile part scores low. The GTX 980M's nearest rivals include the AMD Radeon R7 M445 (5358, 0.9% ahead), the NVIDIA GeForce 940M (5284, 0.5% behind), the NVIDIA GeForce 840M (5322, 0.3% behind), and the NVIDIA GeForce 930A (5317, 0.2% behind). This shows the GTX 980M's average is anchored by its weaker legacy DirectX results, not by its OpenCL or Vulkan performance.

For use-case selection, the data indicates the HD 6950 is a desktop card from an older era that holds its own in compute-oriented OpenCL workloads relative to its immediate peers. The GTX 980M is a mobile module that excels in modern compute APIs (OpenCL, Vulkan) but shows much lower scores in DirectX 9, 10, 11, and 12 legacy tests. If the workload is OpenCL compute, the GTX 980M is decisively faster. If the workload is older DirectX gaming, the HD 6950 may be the safer bet given its higher average and the GTX 980M's sub-60 scores in those categories.

Architecture Differences

The AMD Radeon HD 6950 uses the Cayman chip built on TeraScale 3 architecture, fabricated by TSMC on a 40 nm process. The die measures 389 mm² and packs 2,640 million transistors, yielding a density of 6.8M per mm². It belongs to the Northern Islands (HD 6900) generation and was released on 2010-12-13, making it the older part by almost four years.

The NVIDIA GeForce GTX 980M uses the GM204 chip on Maxwell 2.0 architecture, also fabricated by TSMC but on a 28 nm process. The die is slightly larger at 398 mm² but contains 5,200 million transistors, more than double the HD 6950's count, resulting in a density of 13.1M per mm². It belongs to the GeForce 900M generation, released on 2014-10-06.

The transistor density difference is significant. The GTX 980M packs 13.1M transistors per mm² versus the HD 6950's 6.8M, a direct consequence of the newer 28 nm node versus the older 40 nm node. This allows the Maxwell part to more than double the transistor budget within a similar die area.

In terms of compute resources, the HD 6950 has 1408 shading units, 88 texture mapping units, and 32 ROPs. The GTX 980M has 1536 shading units, 96 TMUs, and 64 ROPs. The NVIDIA part leads in every category, with the ROP count being double that of the AMD card.

Memory configurations differ substantially. The HD 6950 has 2 GB of GDDR5 on a 256-bit bus, delivering 160.0 GB/s bandwidth. The GTX 980M has 8 GB of GDDR5 on a 256-bit bus, delivering 160.4 GB/s. The bus width and effective memory speed (5 Gbps) are the same, but the GTX 980M quadruples the capacity while edging out the bandwidth by a negligible 0.4 GB/s.

The feature sets reflect their eras. The HD 6950 supports DirectX 11.2 (11_0), OpenGL 4.4, and has no Vulkan support. The GTX 980M supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The GTX 980M also has no explicit power connector, uses an MXM-B (3.0) interface, and is a portable-device-dependent display output, whereas the HD 6950 uses PCIe 2.0 x16, dual DVI plus HDMI and mini-DisplayPort outputs, and requires two 6-pin power connectors.

Head-to-Head Benchmarks

The only direct head-to-head test in the database is Geekbench OpenCL. The GTX 980M scores 23832, the HD 6950 scores 6210. The delta is 73.9% in favor of the GTX 980M, meaning the NVIDIA part is nearly four times faster in this compute workload. This is not a marginal victory; it is a generational leap in raw compute throughput.

The GTX 980M's OpenCL result is consistent with its other modern API scores. Its Geekbench Vulkan score is 17703, which is lower than OpenCL but still far above the HD 6950's entire recorded output. The HD 6950 has no Vulkan score at all, as its architecture predates that API.

Looking at the GTX 980M's full benchmark suite, the OpenCL score of 23832 is its highest recorded result. The Vulkan score of 17703 is second. The Passmark G3D score is 7338, and the Passmark G2D score is 490. The Passmark GPU compute score is 2816. The legacy DirectX scores are much lower: DirectX 9 at 125, DirectX 11 at 57, DirectX 10 at 35, and DirectX 12 at 31.

The HD 6950's sole benchmark, the Geekbench OpenCL score of 6210, places it at the 36th percentile of all GPUs in the database. The GTX 980M, despite its much higher OpenCL score, is at the 31st percentile overall because its average includes those weak legacy DirectX results. This creates an interesting dynamic: the GTX 980M wins the head-to-head decisively, yet its average benchmark score (5308) is lower than the HD 6950's (6210).

This discrepancy is explained by the test mix. The GTX 980M has 10 recorded benchmarks; the HD 6950 has 1. The GTX 980M's average is dragged down by scores like 31 in DirectX 12 and 35 in DirectX 10. The HD 6950's single OpenCL score is relatively high for its era. In a pure compute comparison, the GTX 980M is the clear winner. In a legacy gaming comparison, the HD 6950 would likely perform better, though no direct gaming benchmark is recorded.

Specification Differences

The two cards differ in nearly every measurable hardware field except memory bus width (both 256-bit) and effective memory speed (both 5 Gbps). Here is the full breakdown:

| Field | AMD Radeon HD 6950 | NVIDIA GeForce GTX 980M |

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

| Chip | Cayman | GM204 |

| Architecture | TeraScale 3 | Maxwell 2.0 |

| Generation | Northern Islands (HD 6900) | GeForce 900M |

| Process node | 40 nm | 28 nm |

| Transistors | 2,640 million | 5,200 million |

| Die size | 389 mm² | 398 mm² |

| Transistor density | 6.8M / mm² | 13.1M / mm² |

| Base clock | null | 1038 MHz |

| Boost clock | null | 1127 MHz |

| Memory clock | 1250 MHz, 5 Gbps effective | 1253 MHz, 5 Gbps effective |

| Memory size | 2 GB | 8 GB |

| Memory type | GDDR5 | GDDR5 |

| Bus width | 256 bit | 256 bit |

| Bandwidth | 160.0 GB/s | 160.4 GB/s |

| Shading units | 1408 | 1536 |

| TMUs | 88 | 96 |

| ROPs | 32 | 64 |

| Pixel rate | 25.60 GPixel/s | 72.13 GPixel/s |

| Texture rate | 70.40 GTexel/s | 108.2 GTexel/s |

| FP32 | 2.253 TFLOPS | 3.462 TFLOPS |

| TDP | 200 W | null |

| Slot width | Dual-slot | MXM Module |

| Power connectors | 2x 6-pin | None |

| Suggested PSU | 550 W | null |

| Bus interface | PCIe 2.0 x16 | MXM-B (3.0) |

| Display outputs | 2x DVI, 1x HDMI 1.4a, 2x mini-DisplayPort 1.2 | Portable Device Dependent |

| DirectX | 11.2 (11_0) | 12 (12_1) |

| OpenGL | 4.4 | 4.6 |

| Vulkan | null | 1.4 |

| Release date | 2010-12-13 | 2014-10-06 |

| Launch MSRP | 299 USD | null |

The GTX 980M leads in pixel rate (72.13 GPixel/s versus 25.60), texture rate (108.2 GTexel/s versus 70.40), and FP32 throughput (3.462 TFLOPS versus 2.253). It also has a higher memory clock at 1253 MHz versus 1250 MHz, though the effective speed is the same.

The HD 6950 is a dual-slot desktop card with a 200 W TDP and a 550 W suggested PSU. The GTX 980M is an MXM module with no listed power connector or PSU requirement, designed for laptops. The HD 6950 has a fixed set of display outputs; the GTX 980M's outputs are dependent on the host portable device.

FAQ

Q: Which card wins the only direct benchmark in the database?

A: The NVIDIA GeForce GTX 980M wins the Geekbench OpenCL test with a score of 23832 against the AMD Radeon HD 6950's 6210, a delta of 73.9%.

Q: How does the GTX 980M's average benchmark score compare to the HD 6950's?

A: The GTX 980M has an average benchmark score of 5308 across 10 tests, while the HD 6950 has an average of 6210 from its single OpenCL test. The GTX 980M's average is lower because its legacy DirectX scores are very low.

Q: What is the memory capacity difference?

A: The GTX 980M has 8 GB of GDDR5 memory, while the HD 6950 has 2 GB. Both use a 256-bit bus and have nearly identical bandwidth (160.4 GB/s versus 160.0 GB/s).

Q: Which card has a higher transistor count and density?

A: The GTX 980M has 5,200 million transistors on a 398 mm² die, yielding 13.1M per mm². The HD 6950 has 2,640 million transistors on a 389 mm² die, yielding 6.8M per mm².

Q: What API support does each card offer?

A: The HD 6950 supports DirectX 11.2 (11_0) and OpenGL 4.4, with no Vulkan. The GTX 980M supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4.

Q: What are the ROP and shading unit counts?

A: The GTX 980M has 64 ROPs and 1536 shading units. The HD 6950 has 32 ROPs and 1408 shading units. The GTX 980M's ROP count is exactly double the HD 6950's.