AMD Radeon R7 350 vs NVIDIA GeForce GTX 460 Comparison

The NVIDIA GeForce GTX 460 and AMD Radeon R7 350 are two legacy cards from different eras, and the benchmark data reflects a closer contest than their release dates might suggest. The GTX 460, a Fermi-based card from mid-2010, holds a narrow edge in raw compute, while the R7 350, a GCN 1.0 part from 2016, counters with a much larger memory buffer and a far more efficient design. The data shows the GTX 460 wins the only head-to-head test, the Geekbench OpenCL benchmark, by a margin of 1.7%, scoring 7925 against the R7 350's 7792. However, the R7 350's overall average benchmark score of 7425 is dragged down by a separate Vulkan score of 7057, which the GTX 460 lacks entirely. For a builder choosing between these two end-of-life parts, the decision hinges on whether you prioritize raw compute throughput (GTX 460) or efficiency, memory capacity, and modern API support (R7 350).

The Verdict

Pick the NVIDIA GeForce GTX 460 if your workload is purely compute-bound and you can tolerate its older platform requirements. Its Geekbench OpenCL score of 7925 beats the R7 350's 7792, and it sits at the 41st percentile among all GPUs, slightly above the R7 350's 40th percentile. The GTX 460 also has a higher texture fill rate of 37.80 GTexel/s versus 25.60 GTexel/s on the AMD card, which historically benefits certain texture-heavy compute tasks. Its nearest rivals, the NVIDIA Quadro K4100M at 7906 and the GeForce GTX 880M at 8040, show it trading blows with mobile workstation parts from a later generation.

Pick the AMD Radeon R7 350 if you need a card that fits in a small, low-power system. Its 55 W TDP is dramatically lower than the GTX 460's 160 W, and it requires no external power connectors, while the GTX 460 needs two 6-pin connectors and a 450 W suggested PSU versus the R7 350's 250 W suggestion. The R7 350 also brings 2 GB of memory versus 768 MB, a crucial advantage for modern game textures or larger datasets. Its 512 shading units outnumber the GTX 460's 336, and it supports Vulkan 1.2.170, which the GTX 460 does not. The data indicates the R7 350 is the more practical card for a modern, low-profile build, even if it loses the single OpenCL compute test.

Architecture Differences

The two cards represent fundamentally different design philosophies separated by six years of GPU evolution. The GTX 460 uses the GF104 chip based on NVIDIA's Fermi architecture, built on a 40 nm process at TSMC. It packs 1,950 million transistors onto a 332 mm² die, yielding a transistor density of 5.9M per mm². The R7 350 uses AMD's Cape Verde chip with the GCN 1.0 architecture, manufactured on a 28 nm process, also at TSMC. It contains 1,500 million transistors on a much smaller 123 mm² die, achieving a density of 12.2M per mm²—more than double the GTX 460's density.

The compute layouts diverge sharply. The GTX 460 has 336 shading units, 56 texture mapping units (TMUs), and 24 raster output units (ROPs). The R7 350 counters with 512 shading units but only 32 TMUs and 16 ROPs. This explains the GTX 460's higher texture rate (37.80 GTexel/s) and the R7 350's higher pixel rate (12.80 GPixel/s versus 9.450 GPixel/s). The GTX 460 also leads in FP32 compute at 907.2 GFLOPS, compared to the R7 350's 819.2 GFLOPS. Neither card has ray tracing or tensor cores, and both list no FP16 performance.

Memory architecture also differs significantly. The GTX 460 uses a 192-bit memory bus with 768 MB of GDDR5, running at 900 MHz (3.6 Gbps effective), delivering 86.40 GB/s of bandwidth. The R7 350 uses a narrower 128-bit bus but pairs it with 2 GB of GDDR5 at 1125 MHz (4.5 Gbps effective), yielding 72.00 GB/s. The R7 350's lower bandwidth is offset by its larger capacity, a trade-off that favors the newer card for memory-hungry applications.

Head-to-Head Benchmarks

The only direct benchmark comparison in the data is the Geekbench OpenCL test, and it is a close call. The GTX 460 scores 7925, beating the R7 350's 7792 by a 1.7% delta. This is a slim margin, well within the noise of driver variations or thermal conditions. The GTX 460's nearest rivals in the overall database include the Quadro K4100M at 7906 (0.2% behind) and the GTX 880M at 8040 (1.4% ahead), placing it in a cluster of mid-tier mobile and desktop parts. The R7 350's nearest rivals are notably different: the Intel UHD Graphics 750 at 7441 (0.2% behind) and the NVIDIA GeForce GTX 1650 at 7472 (0.6% behind). This suggests the R7 350 competes against integrated graphics and entry-level discrete cards, while the GTX 460 sits closer to older dedicated workstation GPUs.

The R7 350's average benchmark score of 7425 is lower than its OpenCL score because it also has a Geekbench Vulkan score of 7057. That Vulkan result, which the GTX 460 cannot produce due to lacking Vulkan support, pulls the AMD card's average down. In a Vulkan-based workload, the R7 350 would likely outperform the GTX 460, but that test is absent from the head-to-head data. The GTX 460 has no Vulkan score listed, so its average equals its OpenCL score of 7925. This means the GTX 460's percentile ranking of 41 is based solely on OpenCL, while the R7 350's 40th percentile reflects a mix of OpenCL and Vulkan performance.

Specification Differences

| Specification | NVIDIA GeForce GTX 460 | AMD Radeon R7 350 |

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

| Process Node | 40 nm | 28 nm |

| Transistors | 1,950 million | 1,500 million |

| Die Size | 332 mm² | 123 mm² |

| Transistor Density | 5.9M / mm² | 12.2M / mm² |

| Memory Size | 768 MB | 2 GB |

| Memory Bus Width | 192 bit | 128 bit |

| Memory Clock | 900 MHz / 3.6 Gbps | 1125 MHz / 4.5 Gbps |

| Memory Bandwidth | 86.40 GB/s | 72.00 GB/s |

| Shading Units | 336 | 512 |

| TMUs | 56 | 32 |

| ROPs | 24 | 16 |

| Pixel Rate | 9.450 GPixel/s | 12.80 GPixel/s |

| Texture Rate | 37.80 GTexel/s | 25.60 GTexel/s |

| FP32 | 907.2 GFLOPS | 819.2 GFLOPS |

| TDP | 160 W | 55 W |

| Slot Width | Dual-slot | Single-slot |

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

| Suggested PSU | 450 W | 250 W |

| Bus Interface | PCIe 2.0 x16 | PCIe 3.0 x16 |

| Display Outputs | 2x DVI, 1x mini-HDMI 1.3a | 1x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 |

| Vulkan Support | None | 1.2.170 |

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

| Dimensions | 210 mm (8.3 inches) | 168 mm (6.6 inches) |

| Release Date | 2010-07-11 | 2016-07-05 |

| Launch MSRP | 199 USD | None |

The R7 350 is smaller, cooler, and more power-efficient, with a 42 mm shorter length. The GTX 460 requires a dual-slot cooler and external power, while the R7 350 fits in a single slot with no power connectors. The GTX 460's only listed display output is DVI and mini-HDMI, whereas the R7 350 adds DisplayPort 1.2. The R7 350 also supports a newer DirectX feature level (11_1 versus 11_0) and Vulkan, which the GTX 460 lacks entirely.

FAQ

Q: Which card has the higher raw compute performance?

A: The NVIDIA GeForce GTX 460 leads in FP32 compute with 907.2 GFLOPS, compared to the R7 350's 819.2 GFLOPS. It also wins the Geekbench OpenCL test, 7925 versus 7792, a 1.7% margin.

Q: Does the AMD R7 350 support modern APIs that the GTX 460 does not?

A: Yes, the R7 350 supports Vulkan 1.2.170, while the GTX 460 has no Vulkan support listed. The R7 350 also supports DirectX 12 (11_1), whereas the GTX 460 is limited to DirectX 12 (11_0). Both cards support OpenGL 4.6.

Q: Why is the R7 350's average benchmark score lower than its OpenCL score?

A: The R7 350 has two benchmark results: 7792 in Geekbench OpenCL and 7057 in Geekbench Vulkan. Its average score of 7425 is the mean of those two figures, while the GTX 460 has only an OpenCL score, so its average equals 7925.

Q: Which card requires less power and is easier to install?

A: The R7 350 is significantly more efficient, with a 55 W TDP versus the GTX 460's 160 W. It is a single-slot card with no power connectors and a suggested PSU of 250 W, while the GTX 460 is dual-slot, requires two 6-pin connectors, and needs a 450 W PSU.

Q: How does the memory configuration affect real-world usage?

A: The GTX 460 has higher memory bandwidth at 86.40 GB/s, but only 768 MB of VRAM. The R7 350 has 2 GB of VRAM but lower bandwidth at 72.00 GB/s. The larger capacity on the R7 350 is more suitable for modern games and large textures, while the GTX 460's bandwidth is better for older, smaller workloads.

Q: Which card has a higher pixel fill rate, and what does that mean?

A: The R7 350 has a higher pixel rate of 12.80 GPixel/s, versus the GTX 460's 9.450 GPixel/s. This indicates the R7 350 can fill more pixels per second, which can benefit resolution-heavy rendering tasks, even though the GTX 460 wins on texture rate (37.80 GTexel/s versus 25.60 GTexel/s).

Where Each One Wins

NVIDIA GeForce GTX 460 wins on: Raw FP32 compute (907.2 GFLOPS), texture throughput (37.80 GTexel/s), memory bandwidth (86.40 GB/s), and the only head-to-head OpenCL benchmark (7925 vs 7792). It also has a higher transistor count (1,950 million) and a larger die (332 mm²), which historically correlates with more complex shading capabilities. Its nearest rivals are older workstation and high-end mobile GPUs, suggesting it was a stronger part in its era.

AMD Radeon R7 350 wins on: Power efficiency (55 W TDP versus 160 W), memory capacity (2 GB versus 768 MB), pixel fill rate (12.80 GPixel/s), and modern API support with Vulkan 1.2.170. It is also physically smaller (168 mm versus 210 mm) and requires no external power connectors, making it the clear choice for compact builds or systems with weak power supplies. Its 512 shading units outnumber the GTX 460's 336, which can help in compute tasks that scale with shader count rather than raw clock speed. The R7 350's nearest rivals include integrated graphics and entry-level discrete cards, placing it in a more accessible performance tier for budget-conscious upgrades.