AMD Radeon HD 6950 vs NVIDIA GeForce GTX 560 SE Comparison

The Verdict

The recorded benchmark data presents a clear but narrow outcome: the NVIDIA GeForce GTX 560 SE wins the sole head-to-head test, the Geekbench OpenCL benchmark, by a margin of 15.5% over the AMD Radeon HD 6950. The GTX 560 SE scores 7171 points, while the HD 6950 scores 6210 points. This places the GTX 560 SE at the 39th percentile among all GPUs, whereas the HD 6950 sits at the 36th percentile. The delta between these percentiles is small, and the average scores place both cards in a similar performance tier, but the direct comparison favors NVIDIA.

For a user choosing between these two end-of-life cards strictly from the database, the GTX 560 SE is the pick for OpenCL compute workloads, as it demonstrates a measurable advantage in the only recorded benchmark. The HD 6950, however, offers a substantially larger memory pool (2 GB versus 1 GB) and a wider 256-bit memory bus, which may matter for certain workloads that are not captured in the Geekbench OpenCL test. The data does not include any gaming, rendering, or encoding benchmarks, so any conclusion beyond the OpenCL score must remain speculative and grounded only in the architectural and memory specifications listed.

The HD 6950 also carries a launch MSRP of 299 USD, which can be stated once as a historical reference. No such price is recorded for the GTX 560 SE. Neither card is in production; both are marked end-of-life. The GTX 560 SE launched later (February 2012) than the HD 6950 (December 2010), but release timing does not factor into the benchmark outcome.

Architecture Differences

The GTX 560 SE is built on NVIDIA’s Fermi 2.0 architecture, using the GF114 chip manufactured on a 40 nm process at TSMC. It packs 1,950 million transistors onto a 332 mm² die, yielding a transistor density of 5.9 million transistors per square millimeter. The HD 6950 uses AMD’s TeraScale 3 architecture, with the Cayman chip also on a 40 nm TSMC process, but it contains 2,640 million transistors on a 389 mm² die, for a density of 6.8 million per square millimeter. The HD 6950 is thus the larger and more transistor-rich chip.

The shading unit count differs dramatically. The GTX 560 SE has 288 shading units, 48 texture mapping units (TMUs), and 24 render output units (ROPs). The HD 6950 has 1,408 shading units, 88 TMUs, and 32 ROPs. This is a 4.9x difference in shading units, a 1.8x difference in TMUs, and a 1.3x difference in ROPs. The HD 6950’s raw throughput numbers reflect this: its pixel rate is 25.60 GPixel/s versus 8.832 GPixel/s for the GTX 560 SE, and its texture rate is 70.40 GTexel/s versus 35.33 GTexel/s. The HD 6950 also delivers 2.253 TFLOPS of FP32 compute, while the GTX 560 SE delivers 847.9 GFLOPS, a 2.7x difference.

Memory configurations diverge as well. The GTX 560 SE uses 1024 MB of GDDR5 on a 192-bit bus, with a memory clock of 957 MHz (3.8 Gbps effective) and bandwidth of 91.87 GB/s. The HD 6950 has 2 GB of GDDR5 on a 256-bit bus, with a memory clock of 1250 MHz (5 Gbps effective) and bandwidth of 160.0 GB/s. The HD 6950 offers 1.74x the bandwidth and double the capacity.

Power requirements differ, with the GTX 560 SE rated at a 150 W TDP and a suggested 450 W power supply, while the HD 6950 is rated at 200 W TDP and a suggested 550 W power supply. Both use two 6-pin power connectors and occupy a dual-slot form factor. The GTX 560 SE is 210 mm long (8.3 inches), while the HD 6950 is 286 mm long (11.3 inches), 126 mm high (5 inches), and 42 mm wide (1.7 inches). The HD 6950 is a physically larger card.

API support shows minor differences. The GTX 560 SE supports DirectX 12 (11_0) and OpenGL 4.6, while the HD 6950 supports DirectX 11.2 (11_0) and OpenGL 4.4. Neither card lists Vulkan support. Display outputs also differ: the GTX 560 SE has two DVI ports and one mini-HDMI 1.3a; the HD 6950 has two DVI, one HDMI 1.4a, and two mini-DisplayPort 1.2 outputs.

Where Each One Wins

Based solely on the recorded head-to-head benchmark, the GTX 560 SE wins the only comparison: Geekbench OpenCL, with a 15.5% higher score. This is a single data point, but it is the only direct measurement available in the database. The GTX 560 SE’s victory in OpenCL suggests its compute pipeline, despite fewer shading units, is more efficient in this particular workload, or the driver implementation for this test favors the Fermi 2.0 architecture.

The HD 6950 does not win any recorded benchmark; its win count is zero. However, the architectural data points to areas where it would likely excel in unmeasured scenarios. Its 2 GB frame buffer and 160.0 GB/s bandwidth provide a substantial memory advantage, which is typically beneficial in high-resolution textures or large dataset processing. Its higher pixel rate (25.60 GPixel/s) and texture rate (70.40 GTexel/s) indicate stronger fill-rate-bound performance, which could matter in traditional rasterization workloads. The FP32 compute of 2.253 TFLOPS is more than double the GTX 560 SE’s 847.9 GFLOPS, so in compute tasks that scale with raw shader count, the HD 6950 would likely pull ahead, even though it lost the specific OpenCL test recorded.

The GTX 560 SE’s win in OpenCL, despite lower raw compute, may be attributed to driver optimization or the specific kernel used in the test. The database does not specify the workload details. For users prioritizing the recorded metric, the GTX 560 SE is the winner. For users extrapolating from memory bandwidth, texture rate, and shader count, the HD 6950 has a stronger theoretical foundation.

FAQ

Q: Which card scored higher in the Geekbench OpenCL benchmark?

A: The NVIDIA GeForce GTX 560 SE scored 7171 points, which is 15.5% higher than the AMD Radeon HD 6950’s 6210 points.

Q: What are the memory capacities and bus widths of each card?

A: The GTX 560 SE has 1024 MB of GDDR5 on a 192-bit bus. The HD 6950 has 2 GB of GDDR5 on a 256-bit bus.

Q: Which card has more shading units?

A: The AMD Radeon HD 6950 has 1,408 shading units, compared to 288 on the NVIDIA GeForce GTX 560 SE.

Q: What is the power consumption difference?

A: The GTX 560 SE has a TDP of 150 W and suggests a 450 W power supply. The HD 6950 has a TDP of 200 W and suggests a 550 W power supply.

Q: Which card supports a newer version of DirectX?

A: The GTX 560 SE supports DirectX 12 (11_0), while the HD 6950 supports DirectX 11.2 (11_0). Both are effectively DirectX 11-class under the hood.

Q: What is the physical size difference between the two cards?

A: The GTX 560 SE is 210 mm long (8.3 inches). The HD 6950 is 286 mm long (11.3 inches), 126 mm high (5 inches), and 42 mm wide (1.7 inches).

Head-to-Head Benchmarks

The only recorded head-to-head benchmark is Geekbench OpenCL. The GTX 560 SE posted a score of 7171, while the HD 6950 posted 6210. The delta is 15.5% in favor of NVIDIA. The GTX 560 SE’s nearest rivals in the database include the Intel Iris Pro Graphics P580 at 7170 (0% delta), the NVIDIA GeForce GTX 970 at 7157 (0.2% ahead of the GTX 560 SE), the AMD Radeon Vega 8 Mobile at 7203 (0.4% behind the GTX 560 SE), and the NVIDIA GeForce GTX 750 at 7222 (0.7% behind the GTX 560 SE). This clustering indicates the GTX 560 SE sits in a tight performance band around 7171, with no rival more than 0.7% away in either direction.

The HD 6950’s nearest rivals are the AMD FirePro W600 at 6223 (0.2% ahead of the HD 6950), the NVIDIA GeForce RTX 4070 Ti SUPER AD102 at 6270 (0.9% ahead), the NVIDIA GeForce RTX 5070 Ti SUPER at 6270 (0.9% ahead), and the NVIDIA Quadro K620 at 6282 (1.1% ahead). Notably, the HD 6950 is the lowest-scoring among its nearest rivals, while the GTX 560 SE is near the middle of its cluster. This means the GTX 560 SE’s 7171 score is more competitive relative to its immediate peers than the HD 6950’s 6210 is relative to its own peers.

Synthesizing the benchmark result with the architectural specifications, the GTX 560 SE wins the OpenCL test by a significant margin, but the HD 6950 holds advantages in memory bandwidth, capacity, pixel throughput, texture throughput, and raw FP32 compute. The GTX 560 SE counters with a lower TDP, smaller physical footprint, and a more recent OpenGL version. The recorded data does not include any multi-card scaling, power efficiency measurements, or thermal tests, so those aspects remain outside the scope of this analysis.

The 15.5% delta in OpenCL is not trivial. It places the GTX 560 SE in a performance class that overlaps with mid-range GPUs from later generations, as evidenced by its proximity to the GTX 970 (0.2% difference) and GTX 750 (0.7% difference). The HD 6950’s score, by contrast, is only 0.2% to 1.1% away from a set of professional and newer GPUs, but all of those rivals score higher. The HD 6950 thus occupies a lower relative standing in the database’s ranking, at the 36th percentile versus the GTX 560 SE’s 39th percentile.

For a user who values the OpenCL compute score as the primary metric, the GTX 560 SE is the clear choice. For a user who values memory capacity and theoretical fill rates, the HD 6950’s specifications are more compelling, but those advantages are not reflected in the sole recorded benchmark. The database does not provide any evidence that the HD 6950 outperforms the GTX 560 SE in any tested scenario. The GTX 560 SE wins the head-to-head, and that is the only outcome the measurements support.