AMD FirePro W5170M vs NVIDIA GeForce GTX 660 Comparison

The NVIDIA GeForce GTX 660 and AMD FirePro W5170M are both end-of-life mobile and desktop graphics solutions built on a 28 nm TSMC process, yet they serve fundamentally different market segments. The GTX 660 is a dual-slot desktop card from the GeForce 600 generation, while the W5170M is an MXM module designed for professional mobile workstations. In the two available head-to-head benchmark comparisons, the GTX 660 wins decisively in both, but the W5170M's architecture offers certain features that may appeal to specific professional workloads. This analysis examines the benchmark data, architectural differences, and use-case scenarios based solely on the provided specifications.

Head-to-Head Benchmarks

The head-to-head data presents a clear picture of overall performance superiority for the NVIDIA GeForce GTX 660. In the Geekbench OpenCL test, the GTX 660 scores 11,347 against the FirePro W5170M's 8,140, yielding a 39.4% advantage for the NVIDIA card. This is a substantial margin that indicates the GTX 660 handles general-purpose compute workloads with significantly greater efficiency. The OpenCL result is particularly relevant for professional applications that leverage GPU acceleration for tasks like rendering, simulation, or data processing.

The Vulkan benchmark tells a similar story, though with a narrower gap. The GTX 660 achieves 11,415 points versus 9,050 for the AMD FirePro W5170M, a 26.1% difference. Vulkan performance is increasingly important for modern graphics APIs and cross-platform compatibility. The GTX 660's lead here suggests it has better low-level API optimization, which could translate to smoother performance in Vulkan-based titles or compute applications. It is importantly the GTX 660 supports Vulkan 1.2.175, while the W5170M supports Vulkan 1.2.170, a minor version difference that likely contributes to the performance gap.

The average benchmark score reinforces this trend: the GTX 660 averages 9,022 across all its benchmark results, while the W5170M averages 8,595. This places the GTX 660 at the 45th percentile of all GPUs, slightly ahead of the W5170M's 44th percentile. While the 45th percentile is hardly impressive in absolute terms, it represents a 5.0% average advantage for the NVIDIA card. The GTX 660's nearest rival, the NVIDIA TITAN V CEO Edition, scores 9,037 with a delta of -0.2%, meaning the GTX 660 is essentially tied with that high-end card in average performance — a surprising result that speaks to the benchmark suite's weighting. Similarly, the GeForce GTX 560 scores 9,058, just 0.4% above the GTX 660, showing that the older generation is nearly equivalent in these tests.

For the FirePro W5170M, its nearest rival is the NVIDIA Quadro P2200 at 8,686, which is 1.1% higher, and the Intel Arc A380 at 8,558, which is 0.4% lower. This places the W5170M in a tightly contested performance band where small architectural differences can flip results. The AMD Radeon HD 8870M and NVIDIA GeForce MX330 both trail by 1.6%, scoring 8,462 and 8,458 respectively.

Architecture Differences

The two GPUs employ fundamentally different architectures from their respective manufacturers. The GTX 660 is built on NVIDIA's Kepler architecture using the GK106 chip, while the W5170M uses AMD's GCN 1.0 architecture with the Tropo chip. Both are manufactured by TSMC at 28 nm, but the similarities end there.

The transistor counts differ significantly: the GTX 660 houses 2,540 million transistors on a 221 mm² die, yielding a transistor density of 11.5M per mm². In contrast, the W5170M contains 1,500 million transistors on a much smaller 123 mm² die, achieving a higher density of 12.2M per mm². The smaller die for the AMD part is notable given it is a mobile-oriented product, suggesting a focus on power efficiency and thermal management, although no TDP figure is provided for the W5170M.

Core configuration diverges sharply. The GTX 660 features 960 shading units, 80 texture mapping units (TMUs), and 24 render output units (ROPs). The W5170M is equipped with 640 shading units, 40 TMUs, and 16 ROPs. This translates to a 50% advantage for the GTX 660 in shading units, a 100% advantage in TMUs, and a 50% advantage in ROPs. These differences directly impact pixel fill rate and texture fill rate: the GTX 660 delivers 20.64 GPixel/s and 82.56 GTexel/s, while the W5170M manages 14.80 GPixel/s and 37.00 GTexel/s.

Clock speeds also favor the NVIDIA card. The GTX 660 runs at a base clock of 980 MHz with a boost of 1,032 MHz, whereas the W5170M operates at 900 MHz base and 925 MHz boost. This 8-12% clock advantage compounds with the core count advantage to produce the GTX 660's higher FP32 throughput of 1.981 TFLOPS versus 1,184.0 GFLOPS for the AMD part.

Memory architecture presents another key differentiator. Both cards feature 2 GB of GDDR5, but the GTX 660 uses a 192-bit bus with memory clocked at 1,502 MHz (6 Gbps effective), producing 144.2 GB/s of bandwidth. The W5170M has a narrower 128-bit bus with memory at 1,125 MHz (4.5 Gbps effective), yielding just 72.00 GB/s — exactly half the bandwidth of the GTX 660. This bandwidth disparity is critical for texture-heavy workloads and high-resolution rendering.

The physical form factors could not be more different. The GTX 660 is a dual-slot desktop card measuring 241 mm (9.5 inches) in length, requiring a 1x 6-pin power connector and a 300 W suggested PSU. The W5170M is an MXM Module with no power connectors, using the MXM-A (3.0) bus interface, making it suitable for laptop or compact workstation integration. Display outputs also differ: the GTX 660 provides 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2, while the W5170M's outputs are described as "Portable Device Dependent," reflecting its mobile nature.

Where Each One Wins

Based on the benchmark data, the NVIDIA GeForce GTX 660 wins all direct comparisons. Its 39.4% lead in OpenCL makes it the clear choice for compute-heavy applications that rely on OpenCL acceleration, such as scientific computing, video encoding, or certain CAD tasks. The 26.1% Vulkan advantage indicates superior performance in modern game engines and Vulkan-based professional tools. The GTX 660's higher pixel rate (20.64 vs 14.80 GPixel/s) and texture rate (82.56 vs 37.00 GTexel/s) suggest it handles fill-rate-limited scenarios, like high-resolution texture mapping or multi-sample anti-aliasing, far better than the W5170M.

The AMD FirePro W5170M does not win any benchmark category, but its architecture offers specific advantages not captured in the raw scores. Its higher transistor density (12.2M vs 11.5M per mm²) indicates a more compact design that could translate to better power efficiency, though no TDP is listed to confirm this. The W5170M supports DirectX 12 (11_1) compared to the GTX 660's DirectX 12 (11_0), giving it a slightly more modern feature level in Microsoft's graphics API. This could matter for compatibility with newer DirectX 12 titles that require 11_1 features, even if the GTX 660 generally outperforms it in raw benchmarks.

The W5170M's MXM form factor is its primary advantage in real-world deployment. It is designed for portable workstations where the GTX 660's dual-slot, 241 mm length and 6-pin power requirement are simply not viable. For mobile professionals who need a discrete GPU in a laptop chassis, the W5170M is the only option between these two. Its lower clock speeds and reduced core count suggest it generates less heat, making it easier to cool in a confined laptop enclosure.

In terms of ecosystem positioning, the GTX 660's average benchmark score of 9,022 places it at the 45th percentile, while the W5170M sits at the 44th percentile with 8,595. This 1-percentile difference is statistically minor, meaning the two cards are broadly comparable in the wider GPU landscape despite the GTX 660's clear head-to-head wins. The GTX 660's nearest rival delta of -0.2% (against the TITAN V CEO Edition) is essentially a tie, while the W5170M's closest competitor, the Quadro P2200, is 1.1% faster.

Specification Differences

The following specifications differ between the two GPUs:

• Chip: GK106 (NVIDIA) vs Tropo (AMD)
• Architecture: Kepler vs GCN 1.0
• Generation: GeForce 600 vs FirePro Mobile (Wx100M)
• Transistors: 2,540 million vs 1,500 million
• Die Size: 221 mm² vs 123 mm²
• Transistor Density: 11.5M / mm² vs 12.2M / mm²
• Base Clock: 980 MHz vs 900 MHz
• Boost Clock: 1,032 MHz vs 925 MHz
• Memory Clock: 1,502 MHz (6 Gbps effective) vs 1,125 MHz (4.5 Gbps effective)
• Memory Bus Width: 192 bit vs 128 bit
• Memory Bandwidth: 144.2 GB/s vs 72.00 GB/s
• Shading Units: 960 vs 640
• TMUs: 80 vs 40
• ROPs: 24 vs 16
• Pixel Rate: 20.64 GPixel/s vs 14.80 GPixel/s
• Texture Rate: 82.56 GTexel/s vs 37.00 GTexel/s
• FP32: 1.981 TFLOPS vs 1,184.0 GFLOPS
• TDP: 140 W vs not specified
• Slot Width: Dual-slot vs MXM Module
• Power Connectors: 1x 6-pin vs None
• Suggested PSU: 300 W vs not specified
• Bus Interface: PCIe 3.0 x16 vs MXM-A (3.0)
• Display Outputs: 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 vs Portable Device Dependent
• DirectX Support: 12 (11_0) vs 12 (11_1)
• Vulkan Support: 1.2.175 vs 1.2.170
• Release Date: 2012-09-05 vs 2014-08-24
• Predecessor: GeForce 500 vs FirePro Mobility
• Successor: GeForce 700 vs Radeon Pro Mobile
• Launch MSRP: 229 USD (GTX 660 only; W5170M has none listed)

FAQ

Q: Which GPU has higher raw compute performance?

A: The NVIDIA GeForce GTX 660 delivers 1.981 TFLOPS of FP32 compute, compared to 1,184.0 GFLOPS for the AMD FirePro W5170M. This translates to a 67.3% advantage for the GTX 660 in theoretical peak throughput.

Q: What explains the GTX 660's 39.4% OpenCL benchmark lead?

A: The GTX 660 has 960 shading units versus 640 for the W5170M, a 50% advantage in core count, combined with a higher boost clock (1,032 MHz vs 925 MHz). Additionally, the GTX 660 offers 144.2 GB/s memory bandwidth versus 72.00 GB/s for the W5170M, which is critical for OpenCL workloads that frequently access large data sets.

Q: Is the AMD FirePro W5170M better suited for any specific environment?

A: Yes, the W5170M is an MXM Module with no power connectors and a bus interface of MXM-A (3.0), making it suitable for portable workstations. The GTX 660 requires a dual-slot desktop chassis, a 1x 6-pin power connector, and a 300 W PSU, so it cannot be used in mobile form factors.

Q: How do the two GPUs compare in the overall GPU landscape?

A: The GTX 660 sits at the 45th percentile of all GPUs with an average benchmark score of 9,022, while the W5170M is at the 44th percentile with 8,595. The GTX 660's nearest rival, the NVIDIA TITAN V CEO Edition, is just 0.2% behind, whereas the W5170M's closest competitor, the NVIDIA Quadro P2200, is 1.1% ahead.

Q: Do both GPUs support modern graphics APIs?

A: Both support DirectX 12 and OpenGL 4.6. The GTX 660 supports DirectX 12 (11_0) and Vulkan 1.2.175, while the W5170M supports DirectX 12 (11_1) and Vulkan 1.2.170. The W5170M has a slightly higher DirectX feature level, but the GTX 660 has a newer Vulkan version.

Q: What were the release timelines and production status?

A: The GTX 660 was released on 2012-09-05 with a launch MSRP of 229 USD, while the W5170M followed on 2014-08-24 with no listed MSRP. Both are end-of-life products, and the GTX 660's predecessor is GeForce 500 with successor GeForce 700, while the W5170M's predecessor is FirePro Mobility and its successor is Radeon Pro Mobile.