NVIDIA GeForce GTX 560

The NVIDIA GeForce GTX 560, built on the Fermi 2.0 architecture with the GF114 chip, occupies a specific niche in the benchmark hierarchy. Its average score of 8899 places it at the 43rd percentile of all GPUs, indicating a mid-pack position that is defined by narrow margins over its immediate competitors. The data reveals a card whose performance profile is tightly clustered, with deltas of only a few percentage points separating it from both newer and older silicon.

Memory Subsystem

The GTX 560 is equipped with 1024 MB of GDDR5 memory, a capacity that was standard for its generation but modest by modern standards. The memory operates at 1000 MHz with a 4 Gbps effective data rate, paired with a 256-bit bus width. This combination yields a peak bandwidth of 128.0 GB/s. For high-resolution workloads, this bandwidth figure is a limiting factor; the data suggests that at 1080p and above, texture streaming and frame buffer operations will contend for the available throughput. The 256-bit interface is a positive aspect, as it provides a balanced path for the memory controller to feed the 32 ROPs, but the absolute capacity of 1 GB will likely be exhausted by modern game assets, leading to potential stuttering or texture pop-in. In contrast, the pixel rate of 11.34 GPixel/s and texture rate of 45.36 GTexel/s indicate that the card's compute resources can outpace its memory subsystem's ability to supply data in fill-rate-bound scenes. The 128.0 GB/s bandwidth is a hard ceiling, and benchmark results indicate that this becomes the primary constraint when resolution scaling is applied, rather than the raw shading power.

How It Compares

The GTX 560's nearest rival is the AMD Radeon R9 M265X, which posts an average score of 8851. The delta of 0.5% places the GTX 560 marginally ahead, a difference so small that it falls within typical run-to-run variance for synthetic workloads. This effectively makes the two cards performance equivalents, though their architectural approaches differ significantly.

Against the NVIDIA GeForce GTX 660, the comparison is less favorable. The GTX 660 achieves an average score of 9000, representing a 1.1% advantage over the GTX 560. This is a narrow lead, but it is consistent with the GTX 660 being a later-generation part with architectural refinements. The data shows that the GTX 560 trails, but the gap is not a generational chasm; it is a subtle efficiency improvement rather than a raw compute leap.

The NVIDIA TITAN V CEO Edition, with an average score of 9037, sits 1.5% higher than the GTX 560. This is a surprising result given the TITAN V's vastly different positioning, but the benchmark score reflects a specific OpenCL workload where the GTX 560's Fermi compute units remain competitive. The delta is small enough to suggest that the benchmark does not heavily weight the TITAN V's specialized tensor cores or memory bandwidth advantages.

Finally, the AMD Radeon Pro WX 5100 trails the GTX 560 by 1.6%, with a score of 8761. This workstation-oriented card is designed for professional applications, yet in this particular OpenCL test, the GTX 560 manages a lead. The margin is the largest among the four rivals, but still modest, reinforcing the GTX 560's position in a tightly packed performance band around the 8800-9000 score range.

Benchmark Performance

The Geekbench OpenCL score of 8899 is the single data point for this GPU, and its position at the 43rd percentile means that 57% of all GPUs in the database outperform it. This is a decisive indicator that the GTX 560 is not a high-performance part by contemporary standards, but its closest rivals show how competitive the mid-range was during its era. The FP32 compute of 1,088.6 GFLOPS is the theoretical peak, and the OpenCL score of 8899 reflects a realization of that capability under a parallel workload.

Looking at the deltas, the GTX 560 is 0.5% faster than the R9 M265X, a lead that is effectively negligible. The practical takeaway is that these two cards will trade blows depending on driver optimization and specific kernel instructions. The 1.1% deficit to the GTX 660 is more meaningful; it indicates that the GTX 660's architectural tweaks, likely in memory scheduling or shader efficiency, provide a tangible, if small, advantage. The 1.5% gap to the TITAN V CEO Edition is anomalous—the TITAN V should theoretically dominate in most compute tasks, but this specific test does not leverage its unique features. The 1.6% lead over the Radeon Pro WX 5100 suggests that the GTX 560's shader-heavy design is well-suited to OpenCL's execution model, even against a newer professional part.

The texture rate of 45.36 GTexel/s and pixel rate of 11.34 GPixel/s are fixed hardware properties, and they contextualize the benchmark score. A score of 8899 with these rates implies that the test is not purely fill-rate limited; otherwise, the card would rank higher relative to parts with similar bandwidth. The 128.0 GB/s bandwidth, combined with 336 shading units, creates a balance where compute throughput is the primary driver of the OpenCL result. For users, this means that the GTX 560 will perform admirably in compute-heavy tasks that fit within its 1 GB frame buffer, but will degrade noticeably in memory-intensive scenarios.

FAQ

Q: How does the GTX 560 compare to the AMD Radeon R9 M265X?

A: The GTX 560 has an average benchmark score of 8899, which is 0.5% higher than the R9 M265X's 8851. This indicates a statistical tie, with neither card holding a meaningful performance advantage in the tested OpenCL workload.

Q: What is the memory bandwidth of the GTX 560?

A: The card features a 256-bit memory bus paired with GDDR5 memory running at 1000 MHz (4 Gbps effective), yielding a peak bandwidth of 128.0 GB/s. This is a fixed hardware specification that limits high-resolution performance.

Q: Is the GTX 560 faster than the GTX 660?

A: No. The GTX 660 scores 9000, which is 1.1% higher than the GTX 560's 8899. The GTX 560 trails its successor by a narrow margin, indicating a small but consistent performance deficit.

Q: What DirectX version does the GTX 560 support?

A: The GPU supports DirectX 12 (11_0), along with OpenGL 4.6. It does not list Vulkan support in the specification data. This limits its compatibility with some modern APIs.

Q: What is the thermal design power of the GTX 560?

A: The TDP is rated at 150 W, with a suggested power supply of 450 W and two 6-pin power connectors. The card occupies a dual-slot form factor and is 210 mm in length.

Q: What is the GTX 560's percentile ranking among all GPUs?

A: It sits at the 43rd percentile, meaning that 57% of all GPUs in the database have a higher average benchmark score. This places it in the lower half of the performance distribution.

Ray Tracing and Feature Set

The GTX 560 does not include dedicated ray tracing cores or tensor cores; these fields are null in the specification data. This is consistent with its Fermi 2.0 architecture, which predates hardware-accelerated ray tracing and AI-based tensor operations. The card's feature set is therefore limited to traditional rasterization and compute workloads. Its API support includes DirectX 12 (11_0) and OpenGL 4.6, but Vulkan is not listed, which restricts its compatibility with modern cross-platform graphics engines that rely on Vulkan's lower overhead. The lack of tensor cores also means that any deep learning or DLSS-style upscaling features are absent; the card cannot accelerate such workloads. For ray tracing, the absence of RT cores means that any ray-traced effects would need to be computed on the 336 shading units, which would be prohibitively slow for real-time use. The display outputs include 2x DVI and 1x mini-HDMI 1.3a, which support older display standards without modern high-bandwidth interfaces. The PCIe 2.0 x16 bus interface is another legacy feature, providing half the bandwidth of PCIe 3.0, though this is unlikely to bottleneck the card's 128.0 GB/s memory bandwidth in most scenarios. The 1,950 million transistors on a 332 mm² die, manufactured on TSMC's 40 nm process, define the hardware's capabilities; the compute power is fixed at 1,088.6 GFLOPS FP32, and the card's end-of-life production status reflects its age. The feature set is firmly rooted in the early 2010s, offering no modern acceleration pathways beyond basic DirectX 12 and OpenGL support.