AMD Radeon Pro W5700 vs NVIDIA GeForce GTX 980 Ti Comparison

The benchmark data presents a clear generational shift. The AMD Radeon Pro W5700 decisively outperforms the NVIDIA GeForce GTX 980 Ti across every shared workload, driven by a fundamentally newer architecture and a much more efficient manufacturing process. While the GTX 980 Ti was a flagship in its era, the W5700’s results show it is the superior card for modern compute and graphics tasks.

Head-to-Head Benchmarks

The AMD Radeon Pro W5700 wins all three head-to-head comparisons by substantial margins. The largest gap appears in the Geekbench Metal test, where the W5700 scores 89,557 against the GTX 980 Ti's 19,520. This represents a staggering 78.2% deficit for the NVIDIA card, meaning the AMD card delivers over 4.5 times the performance in this Apple-centric API. This is not a marginal improvement; it is a complete overhaul of compute capability in a Metal environment.

In the Geekbench OpenCL test, the W5700 again takes a commanding lead with a score of 74,613 compared to the GTX 980 Ti's 43,513. The delta here is 41.7% in favor of the AMD card. This result indicates that the W5700's advantage is not limited to a single API, but extends to general-purpose GPU compute workloads that rely on OpenCL. The GTX 980 Ti’s score is respectable, but it is firmly in the older generation's performance class.

The Vulkan test shows the closest, yet still one-sided, contest. The AMD Radeon Pro W5700 scores 70,706, while the NVIDIA GeForce GTX 980 Ti manages 47,724. This 32.5% delta confirms the W5700's superiority even in a modern, low-overhead graphics API. While the gap is smaller than in Metal, it is still a decisive victory for the AMD card, suggesting its architectural efficiency translates across different software interfaces.

Looking at the aggregate data, the GTX 980 Ti has an average benchmark score of 28,020, which places it in the 73rd percentile of all GPUs. The W5700's average score is lower at 25,726, placing it in the 71st percentile. This apparent contradiction is explained by the fact that the W5700's average includes several low-scoring Passmark tests (DirectX 9, 10, 11, 12, G2D) which are not part of the head-to-head comparison. In the specific Geekbench tests where they are directly compared, the W5700 is overwhelmingly faster. The GTX 980 Ti does not have any Passmark scores listed, so its average is derived only from its higher-performing Geekbench and 3DMark results.

The Verdict

The data is unambiguous: the AMD Radeon Pro W5700 is the superior GPU for professional and modern compute workloads. Its performance in Metal, OpenCL, and Vulkan is categorically higher than the NVIDIA GeForce GTX 980 Ti. Anyone working in environments that leverage these APIs—such as video editing, 3D rendering, or GPU-accelerated computation—should choose the W5700 without hesitation based on these benchmark results.

The NVIDIA GeForce GTX 980 Ti, while historically significant, is an end-of-life product from 2015. Its performance profile, as shown by the 78.2% deficit in Metal and the 41.7% deficit in OpenCL, is simply not competitive with the newer RDNA 1.0 architecture. Its only advantages lie in its higher average benchmark score, which is skewed by the absence of low-scoring legacy tests in its data set. For any current application that uses these modern APIs, the W5700 is the only logical choice.

There is no scenario in this data where the GTX 980 Ti wins. It has zero wins in the head-to-head benchmarks. The W5700 wins all three. The decision is not about preference; it is about raw, measured performance. The W5700 is the clear winner for any user prioritizing compute performance in the tested APIs.

Architecture Differences

The two cards represent distinct architectural eras. The NVIDIA GeForce GTX 980 Ti is built on the Maxwell 2.0 architecture, using the GM200 chip. This is a mature design from the GeForce 900 generation, fabricated on a 28 nm process at TSMC. The AMD Radeon Pro W5700, in contrast, uses the RDNA 1.0 architecture with the Navi 10 chip, part of the Radeon Pro Navi generation, and is built on a much more advanced 7 nm process.

This process difference is stark. The GTX 980 Ti has 8,000 million transistors on a 601 mm² die, resulting in a transistor density of 13.3M per mm². The W5700 packs 10,300 million transistors onto a much smaller 251 mm² die, achieving a density of 41.0M per mm². This means the AMD card fits nearly 30% more transistors into less than half the space, a direct consequence of the newer manufacturing node.

The architectural differences extend to the compute units. The GTX 980 Ti has 2,816 shading units, 176 texture mapping units (TMUs), and 96 render output units (ROPs). The W5700 has fewer of each: 2,304 shading units, 144 TMUs, and 64 ROPs. Despite having fewer units, the W5700 achieves higher performance due to its higher clock speeds and architectural efficiency. The W5700 also supports FP16 compute at a 2:1 rate, offering 17.33 TFLOPS, while the GTX 980 Ti has no listed FP16 capability.

The memory subsystems also reflect the generational gap. The GTX 980 Ti uses 6 GB of GDDR5 on a 384-bit bus, while the W5700 uses 8 GB of GDDR6 on a 256-bit bus. The newer GDDR6 memory operates at a higher effective speed, giving the W5700 a memory bandwidth of 448.0 GB/s, significantly higher than the GTX 980 Ti's 336.6 GB/s. This is achieved with a narrower bus, showcasing the efficiency of the newer memory standard.

Specification Differences

The two cards differ on nearly every core specification. The process node is a major differentiator, with the GTX 980 Ti using 28 nm and the W5700 using 7 nm. This leads to differences in die size (601 mm² vs 251 mm²) and transistor count (8,000 million vs 10,300 million). The W5700 also has a significantly higher transistor density.

Clock speeds are also distinct. The GTX 980 Ti has a base clock of 1000 MHz and a boost clock of 1076 MHz. The W5700 operates at a higher base clock of 1400 MHz and a much higher boost clock of 1880 MHz. This accounts for the W5700's higher pixel rate (120.3 GPixel/s vs 103.3 GPixel/s) and texture rate (270.7 GTexel/s vs 189.4 GTexel/s).

Memory configuration and power are also different. The W5700 has 8 GB of GDDR6 memory versus 6 GB of GDDR5, and it has higher bandwidth (448.0 GB/s vs 336.6 GB/s). Despite its higher performance, the W5700 has a lower TDP of 205 W compared to the GTX 980 Ti's 250 W, and it suggests a 550 W power supply versus 600 W. The W5700 also uses a PCIe 4.0 x16 interface, while the GTX 980 Ti uses PCIe 3.0 x16. Finally, the display outputs differ, with the W5700 offering 5x mini-DisplayPort 1.4a and a USB Type-C port, while the GTX 980 Ti offers a DVI, HDMI 2.0, and 3x DisplayPort 1.2.

FAQ

Q: Which GPU is faster in the Geekbench Metal test?

A: The AMD Radeon Pro W5700 is dramatically faster, scoring 89,557 compared to the NVIDIA GeForce GTX 980 Ti's 19,520. This is a 78.2% difference in favor of the AMD card.

Q: Does the NVIDIA GTX 980 Ti win any benchmark against the AMD W5700?

A: No. In the head-to-head benchmarks, the AMD Radeon Pro W5700 wins all three tests. The GTX 980 Ti has zero wins.

Q: How does the memory bandwidth compare between the two?

A: The AMD Radeon Pro W5700 has a memory bandwidth of 448.0 GB/s, which is significantly higher than the NVIDIA GeForce GTX 980 Ti's 336.6 GB/s, despite the AMD card using a narrower 256-bit bus.

Q: Is the AMD Radeon Pro W5700 more power-efficient?

A: Yes, the data indicates so. The W5700 has a TDP of 205 W, which is lower than the GTX 980 Ti's 250 W, and it also recommends a lower power supply wattage (550 W vs 600 W).

Q: What is the transistor density difference?

A: The AMD Radeon Pro W5700 has a transistor density of 41.0M per mm², which is over three times higher than the NVIDIA GeForce GTX 980 Ti's 13.3M per mm².

Q: Which card has more shading units?

A: The NVIDIA GeForce GTX 980 Ti has more shading units, with 2,816 compared to the AMD Radeon Pro W5700's 2,304. However, the AMD card still performs faster in benchmarks.

Where Each One Wins

The AMD Radeon Pro W5700 is the clear winner in all modern compute and graphics workloads represented in the data. Its massive leads in Metal (78.2%), OpenCL (41.7%), and Vulkan (32.5%) make it the superior choice for tasks that rely on these APIs. This includes GPU-accelerated rendering, video encoding, machine learning inference, and any professional application that leverages OpenCL or Vulkan for compute. Its higher memory bandwidth and capacity (8 GB vs 6 GB) also make it better suited for large datasets.

The NVIDIA GeForce GTX 980 Ti has no wins in any of the head-to-head benchmarks. Its only statistical advantage is a higher average benchmark score (28,020 vs 25,726) and a higher percentile rank (73rd vs 71st). This is a misleading metric, as the GTX 980 Ti's score is derived only from high-performing Geekbench and 3DMark tests, while the W5700's average is dragged down by multiple low-scoring Passmark legacy API tests. For any real-world application using the tested APIs, the GTX 980 Ti is not competitive. It is an end-of-life product that cannot match the architectural efficiency and raw performance of the newer RDNA 1.0 card.