AMD Radeon Pro WX 5100 vs NVIDIA GeForce GTX 660 Comparison

The benchmark data presents a clear overall winner: the AMD Radeon Pro WX 5100 dominates the NVIDIA GeForce GTX 660 across every tested workload. However, the story is more nuanced when considering the age, architecture, and intended use case of each card. The GTX 660 is a legacy gaming GPU from the Kepler era, while the WX 5100 is a professional Polaris-based card. The performance gap is substantial, but the data also reveals that both cards occupy a similar tier in the overall GPU landscape, as indicated by their close average benchmark scores and percentiles.

Head-to-Head Benchmarks

The AMD Radeon Pro WX 5100 secures a clean sweep, winning all three head-to-head benchmark comparisons. The most dramatic victory occurs in the Geekbench Metal test, where the WX 5100 scores 29003 against the GTX 660's 4305. This represents a staggering 85.2% deficit for the NVIDIA card, indicating a massive performance gulf in this specific API workload. The WX 5100's modern GCN architecture is clearly far better optimized for Apple's Metal API, making it the only sensible choice for any application leveraging that framework.

The gap narrows considerably in the other two tests, though the WX 5100 remains firmly ahead. In Geekbench OpenCL, the AMD card scores 24217, which is 53.1% higher than the GTX 660's 11347. Similarly, in Geekbench Vulkan, the WX 5100 achieves 26909 points, besting the GTX 660's 11415 by a 57.6% margin. These results show that while the WX 5100's advantage is most pronounced in Metal, it also delivers roughly double the performance in general-purpose compute APIs. The GTX 660, despite its Kepler architecture being capable of running these workloads, is simply outclassed by the newer, more powerful silicon in the AMD card.

Contextualizing these scores against their peers reinforces the WX 5100's dominance. The GTX 660's average benchmark score of 9022 places it just 0.2% behind the NVIDIA TITAN V CEO Edition and 0.4% behind the GeForce GTX 560. This suggests that while the GTX 660 is not a weak card, its performance level is comparable to those older or specialized parts. The WX 5100, with an average score of 8863, is 1.3% ahead of the NVIDIA GeForce RTX 3050 A Mobile and 1.4% ahead of the GeForce GTX 460 v2. In a direct head-to-head, the WX 5100's 8863 average is 1.8% higher than the GTX 660's 9022, a modest but consistent edge that aligns with its benchmark sweep.

The Verdict

The data is unambiguous: the AMD Radeon Pro WX 5100 is the superior graphics card. It wins every single benchmark comparison, and its wins are not marginal. The 85.2% lead in Metal and the 53-57% leads in OpenCL and Vulkan demonstrate a generational leap in compute capability. For any user running applications that utilize these APIs, the WX 5100 offers performance that the GTX 660 simply cannot approach. The verdict is for the WX 5100.

However, the choice is not solely about raw speed. The GTX 660, despite its age, remains a functional card with an average benchmark score of 9022, placing it in the 45th percentile of all GPUs. This is nearly identical to the WX 5100's 44th percentile. The data shows that while the WX 5100 wins the head-to-head, both cards sit in the same performance bracket relative to the entire GPU market. A user with a legacy system that only needs basic acceleration and is constrained by older power or cooling infrastructure might still consider the GTX 660, but they would be sacrificing a significant amount of performance. For all new workloads, the WX 5100 is the definitive pick.

Where Each One Wins

Based on the benchmark results, the AMD Radeon Pro WX 5100 wins in every measurable category. Its most significant advantage is in Metal-based applications, where its 29003 score is over six times higher than the GTX 660's 4305. This makes it the clear choice for macOS environments or any software that relies on Metal for rendering or compute. The WX 5100 also wins decisively in general compute workloads, as evidenced by its 53.1% lead in OpenCL and 57.6% lead in Vulkan. This suggests superior performance in tasks like video encoding, 3D rendering, and scientific simulations that leverage these APIs.

The NVIDIA GeForce GTX 660 has no benchmark wins to claim. Its only potential advantage lies in its historical context and physical properties. It is a dual-slot card with a 140 W TDP and requires a 6-pin power connector, while the WX 5100 is a single-slot, 75 W card that draws power solely from the PCIe slot. The GTX 660 also features a 2 GB memory buffer, which is smaller than the WX 5100's 8 GB. While these are not performance wins, they may be relevant for system compatibility in older machines. In terms of actual performance, the GTX 660's only "win" is avoiding a complete shutout in the average score, where its 9022 is only 1.8% lower than the WX 5100's 8863.

FAQ

Q: Which card is faster in Geekbench Metal?

A: The AMD Radeon Pro WX 5100 is significantly faster, scoring 29003 compared to the NVIDIA GeForce GTX 660's 4305. This represents an 85.2% performance deficit for the GTX 660.

Q: How do the two cards compare in OpenCL performance?

A: The Radeon Pro WX 5100 wins with a score of 24217, which is 53.1% higher than the GeForce GTX 660's score of 11347.

Q: What is the performance difference in Vulkan?

A: The AMD card leads again, scoring 26909 against the NVIDIA card's 11415. This is a 57.6% advantage for the Radeon Pro WX 5100.

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce GTX 660 has a slightly higher average score of 9022, compared to the AMD Radeon Pro WX 5100's 8863. This is a 1.8% difference in favor of the GTX 660.

Q: Do both cards have the same level of overall performance?

A: The data suggests they are in a similar tier. The GTX 660 sits in the 45th percentile of all GPUs, while the WX 5100 is in the 44th percentile. Their average scores are also within 1.8% of each other.

Q: Which card has a higher memory capacity?

A: The AMD Radeon Pro WX 5100 has 8 GB of memory, while the NVIDIA GeForce GTX 660 has 2 GB.

Architecture Differences

The two cards are built on fundamentally different architectures from different eras. The NVIDIA GeForce GTX 660 is based on the Kepler architecture, specifically the GK106 chip, manufactured on a 28 nm process at TSMC. This older design packs 2,540 million transistors into a 221 mm² die, resulting in a transistor density of 11.5M per mm². In contrast, the AMD Radeon Pro WX 5100 uses the GCN 4.0 architecture, built on the Ellesmere chip. It is fabricated on a more advanced 14 nm process at GlobalFoundries, which allows AMD to pack 5,700 million transistors into a similar 232 mm² die. This results in a significantly higher transistor density of 24.6M per mm², more than double that of the GTX 660.

These architectural differences translate into vastly different compute resources. The WX 5100 features 1792 shading units, 112 texture mapping units (TMUs), and 32 render output units (ROPs). The GTX 660, by comparison, has only 960 shading units, 80 TMUs, and 24 ROPs. This explains the WX 5100's superior compute throughput. The AMD card also supports FP16 compute at a 1:1 ratio with FP32, offering 3.892 TFLOPS in both precision levels. The GTX 660 only lists FP32 performance at 1.981 TFLOPS, with no FP16 capability mentioned. The API support also differs, with the WX 5100 supporting DirectX 12 (12_0) and Vulkan 1.3, while the GTX 660 is limited to DirectX 12 (11_0) and Vulkan 1.2.175.

Specification Differences

The specification sheets reveal several key differences beyond raw compute. The most significant is memory: the AMD Radeon Pro WX 5100 is equipped with an 8 GB GDDR5 frame buffer on a 256-bit bus, yielding a bandwidth of 160.0 GB/s. The NVIDIA GeForce GTX 660 has a much smaller 2 GB GDDR5 buffer on a 192-bit bus, providing just 144.2 GB/s of bandwidth. This gives the WX 5100 both a larger capacity and higher bandwidth, which is crucial for high-resolution textures and large datasets.

The cards also differ significantly in power and physical design. The GTX 660 has a 140 W TDP and requires a 1x 6-pin power connector, with a suggested PSU of 300 W. It is a dual-slot card measuring 241 mm in length. The WX 5100, on the other hand, is far more efficient with a 75 W TDP, requiring no external power connectors and only a 250 W suggested PSU. It is a single-slot card that is much shorter at 173 mm in length and 112 mm in height. Display outputs differ as well: the GTX 660 offers 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2, while the WX 5100 provides 4x DisplayPort 1.4a outputs. Finally, the release dates and launch MSRPs are distinct, with the GTX 660 launching in 2012 with a launch MSRP of 229 USD, and the WX 5100 launching in 2016 with a launch MSRP of 499 USD.