AMD Radeon RX 480 vs NVIDIA GeForce GTX 980 Ti Comparison

Where Each One Wins

The recorded benchmark data splits these two cards along clear API and workload lines. The AMD Radeon RX 480 takes one decisive victory, while the NVIDIA GeForce GTX 980 Ti claims three wins, including the most modern graphics test in the database.

The RX 480 dominates in the Metal benchmark with a score of 51057 against the GTX 980 Ti's 19520, a 161.6% advantage. That is not a marginal lead; it is a generational gap in API efficiency. Metal is Apple's graphics framework, and the GCN 4.0 architecture's 1:1 FP16 support appears to translate directly into a massive advantage in this workload.

The GTX 980 Ti wins the remaining three tests. In 3DMark Steel Nomad DX12, it scores 1321 versus 966, a 26.9% lead. In Geekbench OpenCL, it posts 43513 against 37998, a 12.7% margin. In Geekbench Vulkan, the gap narrows to just 3.7%, with scores of 47724 and 45968 respectively. The Vulkan result is close enough that real-world differences would be hard to separate, but the DX12 and OpenCL margins are substantial.

The wins breakdown is 3 for NVIDIA, 1 for AMD. However, the magnitude of AMD's Metal win is larger than any single NVIDIA victory. A 161.6% delta outweighs a 26.9% delta in raw competitive terms, even if the NVIDIA card wins more tests overall.

Architecture Differences

These two GPUs come from different foundries, process nodes, and design philosophies. The RX 480 uses the Ellesmere chip on a 14 nm process from GlobalFoundries, packing 5,700 million transistors into a 232 mm² die. The GTX 980 Ti uses the GM200 chip on a 28 nm process from TSMC, with 8,000 million transistors spread across a much larger 601 mm² die.

Transistor density tells the story: the RX 480 achieves 24.6M transistors per mm², while the GTX 980 Ti manages only 13.3M per mm². The newer process node gives AMD a density advantage that partially offsets NVIDIA's larger chip.

The GTX 980 Ti has more raw compute resources. It carries 2816 shading units, 176 texture mapping units, and 96 raster operations units. The RX 480 counters with 2304 shading units, 144 TMUs, and 32 ROPs. NVIDIA's ROP count is triple that of AMD, which helps explain the pixel rate difference: 103.3 GPixel/s versus 40.51 GPixel/s.

Cache and memory configurations differ sharply. The RX 480 uses 8 GB of GDDR5 on a 256 bit bus, delivering 256.0 GB/s of bandwidth. The GTX 980 Ti uses 6 GB of GDDR5 on a wider 384 bit bus, delivering 336.6 GB/s. The wider bus and higher bandwidth give NVIDIA an advantage in bandwidth-heavy workloads, but AMD offers 2 GB more capacity.

Clock speeds favor AMD. The RX 480 runs at 1120 MHz base and 1266 MHz boost, while the GTX 980 Ti runs at 1000 MHz base and 1076 MHz boost. Memory clocks also differ: 2000 MHz (8 Gbps effective) on AMD versus 1753 MHz (7 Gbps effective) on NVIDIA.

FP32 compute is close: 5.834 TFLOPS for AMD versus 6.060 TFLOPS for NVIDIA. The GTX 980 Ti has a 3.9% edge in raw floating point. The RX 480 offers FP16 at 5.834 TFLOPS (1:1), while the GTX 980 Ti has no listed FP16 support. That explains the Metal benchmark blowout.

Power and physical requirements differ meaningfully. The RX 480 carries a 150 W TDP with a single 6-pin connector and a 450 W suggested PSU. The GTX 980 Ti demands 250 W, needs one 6-pin and one 8-pin connector, and suggests a 600 W PSU. The RX 480 is also shorter at 240 mm versus 267 mm, and narrower at 95 mm height versus 111 mm.

API support shows a split. Both support DirectX 12, but AMD implements 12_0 while NVIDIA implements 12_1. Both support OpenGL 4.6. Vulkan support is present on both, with NVIDIA at 1.4 and AMD at 1.3. Display outputs differ: AMD offers one HDMI 2.0b and three DisplayPort 1.4a, while NVIDIA offers one DVI, one HDMI 2.0, and three DisplayPort 1.2.

The Verdict

The data points to different buyers for each card. The AMD Radeon RX 480 is the choice for anyone working in Metal-based environments. Its 51057 Metal score versus 19520 for NVIDIA is not a close call. If your software stack uses Metal, the RX 480 is categorically faster, and the 161.6% lead means the GTX 980 Ti is not competitive in that context.

The NVIDIA GeForce GTX 980 Ti wins the modern DX12 test by 26.9%, the OpenCL test by 12.7%, and the Vulkan test by 3.7%. For DX12 gaming workloads and OpenCL compute, the GTX 980 Ti is the stronger card. The Vulkan margin is slim enough that the RX 480 remains viable, but the data shows NVIDIA ahead.

The RX 480 offers lower power draw at 150 W versus 250 W, a smaller physical footprint, 2 GB more memory, and a newer process node. The GTX 980 Ti offers higher bandwidth, more ROPs, and stronger compute throughput at 6.060 TFLOPS versus 5.834 TFLOPS.

For Metal-centric workflows, the RX 480 is the only rational pick. For DX12 gaming and general OpenCL compute, the GTX 980 Ti has the edge. The average benchmark scores reflect this split: the RX 480 averages 33997 across all tests, placing it in the 78th percentile of all GPUs. The GTX 980 Ti averages 28020, sitting in the 73rd percentile. The RX 480's higher average comes almost entirely from the Metal outlier.

The GTX 980 Ti's nearest rivals cluster around 28000 average score, including the AMD Radeon Pro W5500X at 27973 and the AMD FirePro S7150 at 28117. The RX 480's nearest rivals hover near 34000, including the AMD Radeon HD 7950 at 33951 and the AMD Radeon RX 7700S at 33849. This places the two cards in different performance tiers despite their contemporaneous release.

FAQ

Q: Which card is faster in DirectX 12?

A: The NVIDIA GeForce GTX 980 Ti. In 3DMark Steel Nomad DX12, it scores 1321 against the RX 480's 966, a 26.9% advantage.

Q: Why is the AMD Radeon RX 480 so much faster in Metal?

A: The RX 480 scores 51057 in Geekbench Metal versus 19520 for the GTX 980 Ti, a 161.6% lead. The RX 480 supports FP16 at 1:1 ratio (5.834 TFLOPS), while the GTX 980 Ti has no listed FP16 support, which likely drives the gap.

Q: Which card has more memory?

A: The AMD Radeon RX 480 has 8 GB of GDDR5, while the NVIDIA GeForce GTX 980 Ti has 6 GB. However, the GTX 980 Ti has higher bandwidth at 336.6 GB/s versus 256.0 GB/s due to its 384 bit bus.

Q: Which card requires less power?

A: The AMD Radeon RX 480 has a 150 W TDP and needs only a single 6-pin connector. The GTX 980 Ti has a 250 W TDP and requires one 6-pin plus one 8-pin connector.

Q: How close is the Vulkan performance?

A: Very close. The GTX 980 Ti scores 47724 in Geekbench Vulkan, only 3.7% ahead of the RX 480's 45968. This is the smallest margin among the head-to-head tests.

Q: Which card has a higher average benchmark score?

A: The AMD Radeon RX 480 averages 33997 across all recorded tests, placing in the 78th percentile. The GTX 980 Ti averages 28020, placing in the 73rd percentile.

Head-to-Head Benchmarks

The largest win belongs to the AMD Radeon RX 480 in Geekbench Metal. The RX 480 posts 51057, obliterating the GTX 980 Ti's 19520 by 161.6%. This is the defining result of the comparison. No other test comes close to this margin. The RX 480's FP16 capability appears to be the differentiator, as the GTX 980 Ti has no FP16 support listed in the database.

The GTX 980 Ti's biggest win comes in 3DMark Steel Nomad DX12. It scores 1321 against 966, a 26.9% lead. This is a substantial margin for a modern graphics workload. The GTX 980 Ti's higher ROP count (96 versus 32) and higher pixel rate (103.3 GPixel/s versus 40.51 GPixel/s) likely contribute to this result.

In Geekbench OpenCL, the GTX 980 Ti wins by 12.7%, scoring 43513 against 37998. This margin reflects the GTX 980 Ti's higher FP32 throughput at 6.060 TFLOPS versus 5.834 TFLOPS. The NVIDIA card also has more shading units (2816 versus 2304) and more TMUs (176 versus 144).

The closest contest is Geekbench Vulkan. The GTX 980 Ti scores 47724, the RX 480 scores 45968, a 3.7% difference. This is within the range where driver optimizations or workload variance could flip the result. Both cards support Vulkan, with NVIDIA at version 1.4 and AMD at 1.3.

The average benchmark score tells a separate story. The RX 480 averages 33997, which is 21.3% higher than the GTX 980 Ti's 28020. This is driven entirely by the Metal result. Without Metal, the GTX 980 Ti would have the higher average. The database records the RX 480 in the 78th percentile of all GPUs, compared to the 73rd percentile for the GTX 980 Ti.

Specification Differences

| Specification | AMD Radeon RX 480 | NVIDIA GeForce GTX 980 Ti |

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

| Chip | Ellesmere | GM200 |

| Architecture | GCN 4.0 | Maxwell 2.0 |

| Process Node | 14 nm | 28 nm |

| Transistors | 5,700 million | 8,000 million |

| Die Size | 232 mm² | 601 mm² |

| Transistor Density | 24.6M / mm² | 13.3M / mm² |

| Base Clock | 1120 MHz | 1000 MHz |

| Boost Clock | 1266 MHz | 1076 MHz |

| Memory Clock | 2000 MHz (8 Gbps) | 1753 MHz (7 Gbps) |

| Memory Size | 8 GB | 6 GB |

| Memory Type | GDDR5 | GDDR5 |

| Memory Bus | 256 bit | 384 bit |

| Memory Bandwidth | 256.0 GB/s | 336.6 GB/s |

| Shading Units | 2304 | 2816 |

| TMUs | 144 | 176 |

| ROPs | 32 | 96 |

| Pixel Rate | 40.51 GPixel/s | 103.3 GPixel/s |

| Texture Rate | 182.3 GTexel/s | 189.4 GTexel/s |

| FP32 | 5.834 TFLOPS | 6.060 TFLOPS |

| FP16 | 5.834 TFLOPS (1:1) | Not listed |

| TDP | 150 W | 250 W |

| Power Connectors | 1x 6-pin | 1x 6-pin + 1x 8-pin |

| Suggested PSU | 450 W | 600 W |

| DirectX | 12 (12_0) | 12 (12_1) |

| Vulkan | 1.3 | 1.4 |

| Display Outputs | 1x HDMI 2.0b, 3x DisplayPort 1.4a | 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.2 |

| Length | 240 mm (9.4 inches) | 267 mm (10.5 inches) |

| Height | 95 mm (3.7 inches) | 111 mm (4.4 inches) |

| Width | 35 mm (1.4 inches) | 40 mm (1.6 inches) |

| Launch MSRP | 229 USD | 649 USD |