NVIDIA GeForce GTX 560 vs NVIDIA GeForce GTX 650 Ti Comparison

Head-to-Head Benchmarks

The head-to-head benchmark data records a single OpenCL test, and the GTX 560 takes the win with a score of 9058 against the GTX 650 Ti's 7877. That is a 15% delta in favor of the older card. The GTX 560 also holds a 45th percentile ranking across all GPUs, compared to the GTX 650 Ti's 42nd percentile, indicating the GTX 560 sits slightly higher in the overall performance distribution.

The GTX 560's nearest rivals in the database include the NVIDIA TITAN V CEO Edition, which scores 9037 (a 0.2% delta), and the GeForce GTX 660 at 9022 (0.4% delta). The AMD Radeon 550X trails at 8918 (1.6% behind), while the AMD Radeon 890M leads the GTX 560 by 1.7% with a score of 9210. This places the GTX 560 in a tight cluster where small percentage differences separate it from both older and newer parts.

For the GTX 650 Ti, the nearest rivals are tightly grouped: the GTX 650 Ti Boost scores 8067 (0.2% ahead), the GTX 880M scores 8040 (0.2% behind), the Quadro P5000 scores 8039 (0.2% behind), and the GRID K2 scores 8080 (0.3% ahead). The GTX 650 Ti's average benchmark score is 8053, which includes both its OpenCL result of 7877 and its Vulkan result of 8229. That Vulkan score is notably higher than its OpenCL score, suggesting the card's compute performance varies significantly by API.

The 15% delta in OpenCL is the only recorded head-to-head result, and it is substantial. The GTX 560's advantage comes despite the GTX 650 Ti having a higher FP32 rating (1,425.4 GFLOPS versus 1,088.6 GFLOPS) and a higher texture rate (59.39 GTexel/s versus 45.36 GTexel/s). This suggests the OpenCL workload favors the GTX 560's memory subsystem or scheduling characteristics rather than raw shader throughput.

The Verdict

The data is unambiguous in the single recorded benchmark: the GTX 560 wins the OpenCL test by 15%, and its average benchmark score of 9058 is 12.5% higher than the GTX 650 Ti's 8053. The GTX 560 also ranks higher in the database's overall percentile (45th versus 42nd). For any workload that resembles the OpenCL test, the GTX 560 is the stronger choice.

However, the GTX 650 Ti has its own advantages that are not reflected in the head-to-head score. It has a Vulkan score of 8229, which is higher than its OpenCL score, and it supports Vulkan 1.2.175, while the GTX 560 has no Vulkan support listed. The GTX 650 Ti also draws less power (110 W versus 150 W), requires a single 6-pin connector instead of two, and is a single-slot card versus the GTX 560's dual-slot design. Its dimensions are smaller (145 mm versus 210 mm), and its suggested PSU is 300 W versus 450 W.

The GTX 560 offers a wider memory bus (256 bit versus 128 bit) and higher bandwidth (128.0 GB/s versus 86.40 GB/s), which likely contributes to its OpenCL lead. The GTX 650 Ti counters with more shading units (768 versus 336), more TMUs (64 versus 56), and a smaller die (221 mm² versus 332 mm²) on a newer 28 nm process. The choice depends on whether the user prioritizes the measured OpenCL performance or the newer feature set and lower power footprint.

Architecture Differences

The GTX 560 is built on the GF114 chip using Fermi 2.0 architecture, manufactured on TSMC's 40 nm process. It packs 1,950 million transistors into a 332 mm² die, giving a transistor density of 5.9M per mm². The GTX 650 Ti uses the GK106S chip with Kepler architecture, on TSMC's 28 nm process, with 2,540 million transistors in a 221 mm² die, yielding a density of 11.5M per mm². The newer process allows nearly double the transistor density, which is typical of the generational jump from Fermi to Kepler.

Memory configurations differ markedly. The GTX 560 uses a 256-bit bus with 1024 MB of GDDR5 at 1000 MHz (4 Gbps effective), producing 128.0 GB/s of bandwidth. The GTX 650 Ti uses a 128-bit bus with the same 1024 MB of GDDR5 but at 1350 MHz (5.4 Gbps effective), resulting in 86.40 GB/s. The GTX 560's bandwidth advantage is 48% higher, which is significant for memory-bound workloads.

Shader and texture resources are distributed differently. The GTX 560 has 336 shading units, 56 TMUs, and 32 ROPs. The GTX 650 Ti has 768 shading units, 64 TMUs, but only 16 ROPs. The GTX 650 Ti's shading unit count is more than double, and its pixel rate is higher at 14.85 GPixel/s versus 11.34 GPixel/s. Its texture rate is 59.39 GTexel/s versus 45.36 GTexel/s. FP32 compute is also higher on the GTX 650 Ti at 1,425.4 GFLOPS versus 1,088.6 GFLOPS.

The bus interface differs: the GTX 560 uses PCIe 2.0 x16, while the GTX 650 Ti uses PCIe 3.0 x16. Display outputs are similar (2x DVI and 1x mini-HDMI), but the GTX 650 Ti supports HDMI 1.4a versus the GTX 560's 1.3a. Both support DirectX 12 (11_0) and OpenGL 4.6, but only the GTX 650 Ti lists Vulkan support (1.2.175). Power requirements are lower on the GTX 650 Ti: 110 W TDP versus 150 W, with a suggested PSU of 300 W versus 450 W.

FAQ

Q: Which card is faster in OpenCL benchmarks?

A: The GTX 560 scores 9058 in OpenCL, which is 15% higher than the GTX 650 Ti's 7877. The GTX 560 also has a higher average benchmark score of 9058 versus 8053.

Q: Does the GTX 650 Ti have any performance advantage?

A: The GTX 650 Ti has a Vulkan score of 8229, which is higher than its OpenCL score. The GTX 560 has no recorded Vulkan benchmark or Vulkan API support in the database.

Q: How do the memory systems compare?

A: The GTX 560 has a 256-bit bus with 128.0 GB/s bandwidth, while the GTX 650 Ti has a 128-bit bus with 86.40 GB/s bandwidth. Both have 1024 MB of GDDR5 memory.

Q: What are the power requirements for each card?

A: The GTX 560 has a 150 W TDP and requires two 6-pin power connectors with a suggested 450 W PSU. The GTX 650 Ti has a 110 W TDP, one 6-pin connector, and a suggested 300 W PSU.

Q: Which card has more compute resources?

A: The GTX 650 Ti has 768 shading units and 64 TMUs, versus 336 shading units and 56 TMUs on the GTX 560. The GTX 650 Ti also has higher FP32 performance at 1,425.4 GFLOPS versus 1,088.6 GFLOPS.

Q: Are there physical size differences?

A: The GTX 560 is a dual-slot card measuring 210 mm (8.3 inches), while the GTX 650 Ti is a single-slot card measuring 145 mm (5.7 inches).

Where Each One Wins

The GTX 560 wins the only recorded head-to-head test, the OpenCL benchmark, by 15%. Its 128.0 GB/s bandwidth and 256-bit bus give it a clear edge in memory-intensive compute workloads. The database shows it outperforming the GTX 650 Ti in average score (9058 versus 8053) and overall percentile (45 versus 42). For users running OpenCL-based applications, the GTX 560 is the better performer.

The GTX 650 Ti wins on efficiency and modern features. It consumes 40 W less power (110 W versus 150 W), needs only one 6-pin connector, and runs on a 300 W PSU versus 450 W. Its single-slot, 145 mm form factor is more compact. It has Vulkan support (1.2.175) and a Vulkan score of 8229, which is higher than its OpenCL result. The higher shading unit count (768) and FP32 rating (1,425.4 GFLOPS) suggest it may handle certain compute or shader-heavy tasks better, though the recorded OpenCL data does not confirm this.

The GTX 650 Ti also uses a newer 28 nm process with higher transistor density (11.5M per mm² versus 5.9M), and it supports PCIe 3.0 x16, doubling the bandwidth of the GTX 560's PCIe 2.0 interface. For systems with limited power delivery, space constraints, or a need for Vulkan compatibility, the GTX 650 Ti is the logical pick. For pure OpenCL throughput as measured in the database, the GTX 560 remains the winner.