AMD Radeon Pro 570X vs NVIDIA GeForce GTX 980 Ti Comparison

The Verdict

The recorded data splits these two GPUs cleanly by workload. The AMD Radeon Pro 570X wins the only head-to-head test where the two overlap in the Geekbench Metal suite, scoring 40,066 against 19,520 for the NVIDIA GeForce GTX 980 Ti, a 105.3% advantage. That is a decisive margin, and it points to a clear recommendation for users whose software stack runs on Apple's Metal API. The database places the Radeon Pro 570X at the 76th percentile of all GPUs, with an average benchmark score of 32,176, which puts it 0.9% ahead of the Intel Arc Pro A30M and 0.7% behind the AMD FirePro S10000.

The NVIDIA GeForce GTX 980 Ti, by contrast, dominates the other two comparison points. In Geekbench OpenCL it scores 43,513 versus 27,604, a 36.6% lead for NVIDIA. In Geekbench Vulkan it scores 47,724 versus 28,859, a 39.5% lead. Its average benchmark score of 28,020 places it at the 73rd percentile, with its nearest rival being the AMD Radeon Pro W5500X at 27,973, a 0.2% gap. The GTX 980 Ti also carries a 6 GB frame buffer versus 4 GB, and a 384-bit memory bus versus 256-bit, which matters for texture-heavy workloads.

Who should pick which: the AMD Radeon Pro 570X is the choice for Metal-centric workflows, where its throughput advantage is massive. The NVIDIA GeForce GTX 980 Ti is the choice for OpenCL and Vulkan compute, where it leads by over a third. The GTX 980 Ti also offers a larger memory pool and higher raw pixel and texture rates, making it the stronger candidate for traditional rasterization workloads. The Radeon Pro 570X, however, is an IGP with no power connectors and a 150 W TDP, which suits compact, power-constrained systems, while the GTX 980 Ti is a dual-slot card with a 250 W TDP and a 600 W suggested PSU.

Architecture Differences

The two GPUs come from different architectural generations and foundries. The AMD Radeon Pro 570X uses the Ellesmere chip built on GCN 4.0 architecture, fabricated on a 14 nm process at GlobalFoundries. It integrates 5,700 million transistors on a 232 mm² die, yielding a transistor density of 24.6 million per square millimeter. The NVIDIA GeForce GTX 980 Ti uses the GM200 chip built on Maxwell 2.0 architecture, fabricated on a 28 nm process at TSMC. It integrates 8,000 million transistors on a 601 mm² die, yielding a transistor density of 13.3 million per square millimeter. The AMD chip is therefore smaller and denser, while the NVIDIA chip packs more total transistors into a much larger physical area.

Clock behavior also differs. The Radeon Pro 570X runs at a 1000 MHz base and 1105 MHz boost, while the GTX 980 Ti runs at a 1000 MHz base and 1076 MHz boost. The AMD part has a higher boost clock, but the NVIDIA part compensates with a wider memory interface and higher memory bandwidth. The Radeon Pro 570X's memory runs at 1695 MHz with 6.8 Gbps effective throughput on a 256-bit bus, delivering 217.0 GB/s. The GTX 980 Ti's memory runs at 1753 MHz with 7 Gbps effective throughput on a 384-bit bus, delivering 336.6 GB/s. That is a 55% bandwidth advantage for NVIDIA, which directly affects fill-rate-bound scenes.

Compute resources differ substantially. The Radeon Pro 570X has 1,792 shading units, 112 texture mapping units, and 32 ROPs. The GTX 980 Ti has 2,816 shading units, 176 TMUs, and 96 ROPs. The NVIDIA part has 57% more shading units, 57% more TMUs, and triple the ROP count. Pixel rate is 35.36 GPixel/s for AMD versus 103.3 GPixel/s for NVIDIA, a 3x gap. Texture rate is 123.8 GTexel/s for AMD versus 189.4 GTexel/s for NVIDIA. FP32 throughput is 3.960 TFLOPS for AMD versus 6.060 TFLOPS for NVIDIA. The AMD part supports FP16 at the same 3.960 TFLOPS, while the GTX 980 Ti has no recorded FP16 figure.

API support differs in one key aspect: the Radeon Pro 570X supports DirectX 12 (12_0) and Vulkan 1.3, while the GTX 980 Ti supports DirectX 12 (12_1) and Vulkan 1.4. Both support OpenGL 4.6. The GTX 980 Ti's higher DirectX feature level and newer Vulkan version give it an edge in compatibility with recent titles. The AMD part's display outputs are portable-device dependent, reflecting its IGP nature, while the GTX 980 Ti offers 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon Pro 570X has an average benchmark score of 32,176, which is 14.8% higher than the NVIDIA GeForce GTX 980 Ti's average of 28,020. The AMD part also sits at the 76th percentile of all GPUs, three points higher than the GTX 980 Ti's 73rd percentile.

Q: How large is the AMD Radeon Pro 570X's Metal advantage?

A: In Geekbench Metal, the Radeon Pro 570X scores 40,066 versus 19,520 for the GTX 980 Ti, a 105.3% lead. This is the single largest margin in the head-to-head data and suggests the AMD part is far better suited to Metal-optimized applications.

Q: Does the GTX 980 Ti outperform the Radeon Pro 570X in any benchmark?

A: Yes, in two of the three recorded tests. The GTX 980 Ti leads by 36.6% in Geekbench OpenCL (43,513 versus 27,604) and by 39.5% in Geekbench Vulkan (47,724 versus 28,859).

Q: What memory advantages does the GTX 980 Ti offer?

A: The GTX 980 Ti has 6 GB of GDDR5 memory, a 384-bit bus, and 336.6 GB/s bandwidth. The Radeon Pro 570X has 4 GB of GDDR5, a 256-bit bus, and 217.0 GB/s bandwidth. The NVIDIA part offers 50% more capacity and 55% more bandwidth.

Q: Which GPU is more suitable for compact systems?

A: The AMD Radeon Pro 570X is an IGP with no power connectors, a 150 W TDP, and portable-device-dependent display outputs. The GTX 980 Ti is a dual-slot card requiring 1x 6-pin and 1x 8-pin power connectors, a 250 W TDP, and a 600 W suggested PSU. The AMD part is clearly the lower-power, space-efficient option.

Q: What are the transistor density differences?

A: The AMD Radeon Pro 570X packs 5,700 million transistors into 232 mm², giving 24.6 million transistors per mm². The NVIDIA GTX 980 Ti packs 8,000 million transistors into 601 mm², giving 13.3 million per mm². The AMD part's 14 nm process enables nearly double the density.

Specification Differences

The two GPUs differ across nearly every measurable specification. The AMD Radeon Pro 570X uses the Ellesmere chip with GCN 4.0 architecture, while the NVIDIA GeForce GTX 980 Ti uses the GM200 chip with Maxwell 2.0. Manufacturing process: 14 nm at GlobalFoundries for AMD, 28 nm at TSMC for NVIDIA. Transistor count: 5,700 million for AMD, 8,000 million for NVIDIA. Die size: 232 mm² for AMD, 601 mm² for NVIDIA. Transistor density: 24.6M per mm² for AMD, 13.3M per mm² for NVIDIA.

Clock speeds: both have a 1000 MHz base, but AMD boosts to 1105 MHz while NVIDIA boosts to 1076 MHz. Memory clocks: AMD runs at 1695 MHz with 6.8 Gbps effective, NVIDIA at 1753 MHz with 7 Gbps effective. Memory capacity: 4 GB for AMD, 6 GB for NVIDIA. Memory bus: 256-bit for AMD, 384-bit for NVIDIA. Bandwidth: 217.0 GB/s for AMD, 336.6 GB/s for NVIDIA.

Compute units: AMD has 1,792 shading units, 112 TMUs, and 32 ROPs, while NVIDIA has 2,816 shading units, 176 TMUs, and 96 ROPs. Pixel rates: 35.36 GPixel/s for AMD, 103.3 GPixel/s for NVIDIA. Texture rates: 123.8 GTexel/s for AMD, 189.4 GTexel/s for NVIDIA. FP32: 3.960 TFLOPS for AMD, 6.060 TFLOPS for NVIDIA. FP16: 3.960 TFLOPS for AMD, none recorded for NVIDIA.

Power and physical specs: AMD has a 150 W TDP, IGP slot width, and no power connectors, while NVIDIA has a 250 W TDP, dual-slot width, and 1x 6-pin plus 1x 8-pin connectors. The NVIDIA card requires a 600 W suggested PSU. Dimensions: the GTX 980 Ti is 267 mm (10.5 inches) long, 111 mm (4.4 inches) high, and 40 mm (1.6 inches) wide; the Radeon Pro 570X has no recorded dimensions due to its IGP form factor.

API differences: AMD supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.3, while NVIDIA supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. Display outputs: AMD's are portable-device dependent, NVIDIA's are 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2. Release dates: AMD launched on 2019-03-17, NVIDIA on 2015-06-01. The GTX 980 Ti has a recorded launch MSRP of 649 USD.

Head-to-Head Benchmarks

The head-to-head data consists of three Geekbench tests, and the results split sharply along API lines. The AMD Radeon Pro 570X takes Geekbench Metal with a score of 40,066, versus 19,520 for the NVIDIA GeForce GTX 980 Ti. That is a 105.3% delta, meaning the AMD part more than doubles the NVIDIA score in this test. This is the defining result for the Radeon Pro 570X; the database shows no other test where the AMD part wins, and the margin here is so large that it drives the AMD part's higher average score despite losing the other two tests.

The NVIDIA GeForce GTX 980 Ti wins Geekbench OpenCL with 43,513 against 27,604, a 36.6% delta. This is a substantial lead, but it is roughly one-third of the margin AMD achieved in Metal. The GTX 980 Ti also wins Geekbench Vulkan with 47,724 against 28,859, a 39.5% delta. That is the largest NVIDIA margin in the dataset, and it aligns with the GTX 980 Ti's newer Vulkan 1.4 support versus the Radeon Pro 570X's Vulkan 1.3.

The win tally is 1 for AMD and 2 for NVIDIA. However, the average benchmark scores tell a different story: the AMD part averages 32,176, which is 14.8% higher than the NVIDIA part's 28,020. This happens because the Metal score is heavily weighted in the average, and the AMD part's 105.3% advantage in that test outweighs its two losses. The percentile ranking reflects this: AMD sits at 76, NVIDIA at 73. The nearest rivals for AMD are all within about 1%: the AMD FirePro S10000 at 32,388 (0.7% ahead), the AMD Radeon RX 7900 GRE at 32,456 (0.9% ahead), the Intel Arc Pro A30M at 31,894 (0.9% behind), and the AMD FirePro S9300 X2 at 32,540 (1.1% ahead). The nearest rivals for NVIDIA are similarly tight: the AMD Radeon Pro W5500X at 27,973 (0.2% behind), the AMD FirePro S7150 at 28,117 (0.3% ahead), the AMD Radeon RX 7800M at 27,883 (0.5% behind), and the AMD Radeon Pro Vega 20 at 27,839 (0.7% behind).

Interpreting the data: the Radeon Pro 570X is a specialist. Its Metal performance is exceptional, but its OpenCL and Vulkan scores are well below the GTX 980 Ti. The GTX 980 Ti is the more balanced compute card, with strong OpenCL and Vulkan results and a much larger memory bus. The choice between them depends entirely on the API and workload environment. In a Metal-centric ecosystem, the AMD part is clearly superior. In OpenCL or Vulkan environments, the NVIDIA part holds a commanding lead. For raw fill-rate and bandwidth-bound tasks, the GTX 980 Ti's 103.3 GPixel/s pixel rate, 189.4 GTexel/s texture rate, and 336.6 GB/s bandwidth make it the more capable rasterizer. The Radeon Pro 570X's 3.960 TFLOPS FP32 and FP16 throughput is competitive for compute, but its 32 ROPs and 217.0 GB/s bandwidth limit its output-stage performance.