AMD Radeon R7 350 vs NVIDIA GeForce GTX 560 SE Comparison

The AMD Radeon R7 350 and NVIDIA GeForce GTX 560 SE are two end-of-life graphics cards from different eras, and the benchmark data shows a clear, though modest, winner. The R7 350 scores 7792 in Geekbench OpenCL, while the GTX 560 SE scores 7171, giving the AMD card an 8.7% performance advantage in the only head-to-head test available. This places the R7 350 at the 40th percentile of all GPUs, while the GTX 560 SE sits just below at the 39th percentile, making both cards firmly entry-level by modern standards, but the R7 350 is the better performer.

The Verdict

The data points to one clear recommendation: choose the AMD Radeon R7 350 for raw compute performance. Its 8.7% lead in OpenCL benchmarks (7792 vs 7171) is consistent and meaningful, especially for tasks that leverage general-purpose GPU compute. The R7 350 also benefits from a much newer architecture and lower power demands, making it the more practical choice for a system that might run 24/7 or in a small form factor case.

The NVIDIA GeForce GTX 560 SE is only worth considering if you have a specific need for its unique traits: a 192-bit memory bus that provides 91.87 GB/s of bandwidth (27.6% more than the R7 350's 72.00 GB/s), or if you require dual DVI outputs on the card itself. However, its 150W TDP and dual-slot cooler with two 6-pin power connectors make it a much harder fit in modern systems, and its older Fermi 2.0 architecture lacks Vulkan support entirely.

For most builders, the R7 350 is the straightforward pick. It wins the benchmark, runs cooler (55W vs 150W TDP), requires no auxiliary power connectors, and fits in a single slot. The GTX 560 SE is a relic that only makes sense in a legacy system where its specific memory bandwidth or output configuration is absolutely required.

Architecture Differences

The two cards come from completely different design philosophies. The R7 350 uses AMD's GCN 1.0 architecture on the Cape Verde chip, built on a 28 nm process at TSMC. In contrast, the GTX 560 SE uses NVIDIA's Fermi 2.0 architecture on the GF114 chip, also from TSMC but on a larger 40 nm node.

The transistor counts tell the story of architectural efficiency. The R7 350 packs 1,500 million transistors into a 123 mm² die, achieving a density of 12.2M transistors per mm². The GTX 560 SE has more transistors — 1,950 million — but spreads them over a much larger 332 mm² die, resulting in only 5.9M transistors per mm². This nearly 2x density advantage for AMD reflects the newer manufacturing process and explains the massive difference in power consumption (55W vs 150W).

The compute layouts are also fundamentally different. The R7 350 uses 512 shading units with 32 TMUs and 16 ROPs. The GTX 560 SE uses 288 shading units but pairs them with 48 TMUs and 24 ROPs. This means NVIDIA's card has more texture and pixel throughput per shading unit, which explains its higher texture rate (35.33 GTexel/s vs 25.60 GTexel/s) despite the R7 350 having more raw shader cores.

Feature support is a major differentiator. The R7 350 supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The GTX 560 SE supports DirectX 12 (11_0) and OpenGL 4.6 but has no Vulkan support at all. This is a critical gap for any modern application that uses Vulkan for cross-platform rendering.

FAQ

Q: Which card is faster in compute workloads?

A: The AMD Radeon R7 350 wins the only head-to-head benchmark, scoring 7792 in Geekbench OpenCL versus 7171 for the GTX 560 SE, an 8.7% advantage.

Q: Do these cards support modern graphics APIs?

A: The R7 350 supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The GTX 560 SE supports DirectX 12 (11_0) and OpenGL 4.6 but lacks Vulkan support entirely.

Q: Which card has higher memory bandwidth?

A: The GTX 560 SE has higher memory bandwidth at 91.87 GB/s, thanks to its 192-bit bus, compared to the R7 350's 72.00 GB/s over a 128-bit bus. However, the R7 350 has 2 GB of memory versus 1 GB on the GTX 560 SE.

Q: What are the power requirements?

A: The R7 350 has a 55W TDP and requires no power connectors, with a suggested 250W PSU. The GTX 560 SE has a 150W TDP, needs two 6-pin power connectors, and suggests a 450W PSU.

Q: Which card is physically smaller?

A: The R7 350 is a single-slot card at 168 mm (6.6 inches) long. The GTX 560 SE is a dual-slot card at 210 mm (8.3 inches) long, making it significantly bulkier.

Q: How do these cards compare to other GPUs in the database?

A: The R7 350 sits at the 40th percentile, near the Intel Arc A310 (7550 score) and GeForce GTX 1650 (7472). The GTX 560 SE is at the 39th percentile, trading blows with the GeForce GTX 750 (7222) and AMD Radeon Vega 8 Mobile (7203).

Specification Differences

The two cards differ across nearly every specification field. The process node is a major gap: 28 nm for AMD versus 40 nm for NVIDIA. This leads to the transistor density difference of 12.2M/mm² versus 5.9M/mm², despite the GTX 560 SE having more total transistors (1,950 million vs 1,500 million) and a much larger die (332 mm² vs 123 mm²).

The memory subsystems are distinct: the R7 350 has 2 GB of GDDR5 on a 128-bit bus at 1125 MHz (4.5 Gbps effective), yielding 72.00 GB/s. The GTX 560 SE has 1 GB of GDDR5 on a 192-bit bus at 957 MHz (3.8 Gbps effective), yielding 91.87 GB/s.

The compute units differ: 512 shading units, 32 TMUs, and 16 ROPs on the R7 350; 288 shading units, 48 TMUs, and 24 ROPs on the GTX 560 SE. This results in the R7 350 having a higher pixel rate (12.80 GPixel/s vs 8.832 GPixel/s) but a lower texture rate (25.60 GTexel/s vs 35.33 GTexel/s). The FP32 performance is nearly identical at 819.2 GFLOPS for AMD and 847.9 GFLOPS for NVIDIA.

The power and physical specs are starkly different: 55W TDP with no power connectors and single-slot design for AMD; 150W TDP with two 6-pin connectors and dual-slot design for NVIDIA. The bus interface also differs: PCIe 3.0 x16 for AMD versus PCIe 2.0 x16 for NVIDIA. Display outputs are: 1x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 for the R7 350; 2x DVI, 1x mini-HDMI 1.3a for the GTX 560 SE.

Head-to-Head Benchmarks

The only direct comparison available is the Geekbench OpenCL test, and it decisively favors the AMD Radeon R7 350. The AMD card scores 7792, while the NVIDIA GeForce GTX 560 SE scores 7171, a delta of 8.7%. This is not a marginal win; it is a solid, reproducible lead in a compute-heavy workload.

This result is interesting because the theoretical FP32 numbers are nearly identical: 819.2 GFLOPS for the R7 350 versus 847.9 GFLOPS for the GTX 560 SE. Despite the NVIDIA card having a slight edge on paper, the AMD card wins in practice. This suggests the newer GCN architecture extracts more real-world performance from its compute units, likely due to better instruction scheduling and memory efficiency.

The GTX 560 SE's higher memory bandwidth (91.87 GB/s) and texture rate (35.33 GTexel/s) did not translate into a win in this test. The R7 350's advantages in pixel rate (12.80 GPixel/s) and its 2 GB frame buffer likely contributed to its success, as OpenCL workloads can be memory-capacity sensitive.

In the broader context of the database, the R7 350's 7792 score puts it ahead of the Intel Arc A310 (7550) and GeForce GTX 1650 (7472), while the GTX 560 SE's 7171 score places it just behind the GeForce GTX 750 (7222) and AMD Radeon Vega 8 Mobile (7203). The R7 350 is also closer to the Intel UHD Graphics 750 (7441) and AMD Radeon HD 8850M (7447) than it is to modern mid-range cards.

Where Each One Wins

The AMD Radeon R7 350 wins in compute performance, as proven by its 8.7% OpenCL lead. It is also the clear winner in power efficiency, physical footprint, and modern API support. The 55W TDP with no power connectors means it can be installed in virtually any system with a PCIe x16 slot, including older office PCs with small power supplies. Its single-slot design and shorter 168 mm length make it ideal for compact cases. The Vulkan 1.2.170 support means it can run modern Linux and Windows applications that rely on Vulkan, while the GTX 560 SE cannot. The 2 GB memory capacity also gives it more headroom for texture-heavy workloads.

The NVIDIA GeForce GTX 560 SE wins in memory bandwidth, texture throughput, and raw fill-rate capabilities. Its 91.87 GB/s bandwidth is 27.6% higher than the R7 350's, which could benefit certain bandwidth-bound tasks. The 48 TMUs and 24 ROPs give it a 35.33 GTexel/s texture rate and an 8.832 GPixel/s pixel rate that, while lower than the R7 350's pixel rate, still offers good texture filtering performance. Its dual DVI outputs are useful for connecting older monitors without adapters. The 192-bit memory bus is a legacy advantage that might matter in specific embedded or industrial applications that are optimized for wider buses.

For gaming, the data is sparse, but the compute lead suggests the R7 350 will generally perform better in modern titles that use compute shaders. The GTX 560 SE's higher bandwidth might help in older games that are bandwidth-limited, but its lack of Vulkan support and higher power draw make it a poor choice for any new system. The R7 350 is the safer, more capable card for almost any use case, with the GTX 560 SE only being relevant in specialized legacy scenarios.