NVIDIA GeForce GTX 660 vs NVIDIA Quadro K4100M Comparison

FAQ

Q: What are the average benchmark scores for the NVIDIA GeForce GTX 660 and the NVIDIA Quadro K4100M?

A: The GeForce GTX 660 has an average benchmark score of 9022, while the Quadro K4100M has an average benchmark score of 7906. This places the GTX 660 in the 45th percentile of all GPUs, compared to the 41st percentile for the Quadro K4100M.

Q: How do the two cards compare in the Geekbench OpenCL test?

A: The GeForce GTX 660 scores 11347 in Geekbench OpenCL, which is 24% higher than the Quadro K4100M's score of 9149. This is the GTX 660's strongest win in the head-to-head comparison.

Q: Which GPU wins the Geekbench Metal test?

A: The Quadro K4100M wins the Geekbench Metal test with a score of 6662, which is 35.4% higher than the GeForce GTX 660's score of 4305. This is a substantial margin in favor of the Quadro.

Q: What are the memory capacities and bus widths of these two GPUs?

A: The GeForce GTX 660 has 2 GB of GDDR5 memory on a 192-bit bus, while the Quadro K4100M has 4 GB of GDDR5 memory on a 256-bit bus. The Quadro K4100M offers double the memory capacity and a wider memory interface.

Q: What are the respective transistor counts and die sizes?

A: The GeForce GTX 660 uses the GK106 chip with 2,540 million transistors on a 221 mm² die. The Quadro K4100M uses the GK104 chip with 3,540 million transistors on a 294 mm² die, making it a larger and more complex chip.

Q: Are both GPUs based on the same architecture?

A: Yes, both the GeForce GTX 660 and the Quadro K4100M are based on NVIDIA's Kepler architecture and are manufactured by TSMC on a 28 nm process node. They also share the same API support for DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175.

Architecture Differences

Both GPUs trace their lineage to the Kepler architecture, but they implement it through different physical chips. The GeForce GTX 660 is built around the GK106 chip, while the Quadro K4100M uses the GK104 chip. The GK104 is the larger of the two, packing 3,540 million transistors across a 294 mm² die. The GK106 in the GTX 660 contains 2,540 million transistors on a 221 mm² die. Transistor density is close between them, with the GK104 at 12.0M per mm² and the GK106 at 11.5M per mm².

The clock behavior of the two cards differs significantly. The GeForce GTX 660 runs at a base clock of 980 MHz with a boost clock of 1032 MHz. The Quadro K4100M operates at a flat 706 MHz for both base and boost, meaning it does not have the same dynamic clock headroom. This clock disparity is a major reason the GTX 660 posts higher raw throughput figures despite having fewer functional units.

Memory configurations also diverge. The GTX 660 has 2 GB of GDDR5 on a 192-bit bus, with memory clocked at 1502 MHz (6 Gbps effective) delivering 144.2 GB/s of bandwidth. The Quadro K4100M carries 4 GB of GDDR5 on a 256-bit bus, but its memory runs at 800 MHz (3.2 Gbps effective), yielding 102.4 GB/s. The Quadro has more capacity and a wider bus, but the GTX 660 has significantly higher memory bandwidth.

In terms of execution resources, the Quadro K4100M actually holds the advantage. It has 1152 shading units, 96 texture mapping units, and 32 raster output pipelines. The GeForce GTX 660 has 960 shading units, 80 TMUs, and 24 ROPs. This means the Quadro K4100M has more cores available for compute work, yet its lower clock speed limits the realized performance from those units.

The physical form factor and power delivery also separate the two products. The GTX 660 is a dual-slot card measuring 241 mm in length (9.5 inches), requiring a single 6-pin power connector and a 300 W suggested power supply. The Quadro K4100M is an MXM module, needs no power connectors, and carries a 100 W TDP compared to the GTX 660's 140 W. The Quadro's mobile-oriented MXM-B (3.0) interface and portable-device-dependent display outputs reflect its workstation laptop positioning, while the GTX 660 is a desktop card with 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2 outputs.

Head-to-Head Benchmarks

The head-to-head benchmark data contains only two tests, and each GPU takes one of them. The GeForce GTX 660 wins the Geekbench OpenCL test by a wide margin, scoring 11347 against the Quadro K4100M's 9149. That is a 24% advantage for the GTX 660, and it is the larger of the two deltas in this comparison. OpenCL performance is often a proxy for general compute throughput, and the GTX 660's higher clocks and faster memory bandwidth help explain its dominance here.

The Quadro K4100M answers in the Geekbench Metal test, where it scores 6662 versus the GTX 660's 4305. The delta is 35.4% in favor of the Quadro, which is even more lopsided than the GTX 660's OpenCL win. Metal is a graphics API used primarily on Apple platforms, and the Quadro K4100M's larger shading unit count, 1152 versus 960, likely contributes to its strong showing in this workload.

Looking at the overall averages, the GeForce GTX 660 maintains a lead in the aggregate. Its average benchmark score of 9022 sits above the Quadro K4100M's 7906, a difference of roughly 14%. The GTX 660's nearest rivals in the database include the NVIDIA TITAN V CEO Edition at 9037 (a 0.2% gap) and the GeForce GTX 560 at 9058 (a 0.4% gap). The Quadro K4100M, by contrast, sits near the GeForce GTX 460, which scores 7925, a 0.2% difference in the Quadro's favor, and the Quadro P5000 at 8039, which is 1.7% ahead.

Benchmark results indicate that the GTX 660 is the stronger all-around performer in this pairing, despite the Quadro K4100M winning the Metal test outright, because the GTX 660's OpenCL victory is backed by a higher average score and a better percentile rank. The Quadro's Metal win is undeniable, but it is confined to a single test.

Specification Differences

The two GPUs differ across nearly every specification category. The chip design is different: GK106 for the GTX 660 versus GK104 for the Quadro K4100M. The GTX 660 has a higher base clock of 980 MHz versus 706 MHz, and a boost clock of 1032 MHz versus a fixed 706 MHz. Memory clocks also favor the GTX 660, at 1502 MHz (6 Gbps effective) compared to 800 MHz (3.2 Gbps effective).

Memory capacity and bus width favor the Quadro K4100M, with 4 GB on a 256-bit bus versus 2 GB on a 192-bit bus. However, bandwidth favors the GTX 660, at 144.2 GB/s versus 102.4 GB/s. The Quadro K4100M has more shading units (1152 versus 960), more TMUs (96 versus 80), and more ROPs (32 versus 24).

Pixel rate and texture rate favor the GTX 660, at 20.64 GPixel/s and 82.56 GTexel/s respectively. The Quadro K4100M records 16.94 GPixel/s and 67.78 GTexel/s. FP32 compute also favors the GTX 660, at 1.981 TFLOPS versus 1.627 TFLOPS.

Power and physical specifications differ substantially. The GTX 660 has a 140 W TDP, while the Quadro K4100M has a 100 W TDP. The GTX 660 is a dual-slot desktop card with a 241 mm length (9.5 inches), a 6-pin power connector, and a 300 W suggested PSU. The Quadro K4100M is an MXM module with no power connectors and no suggested PSU listed. Display outputs are 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2 for the GTX 660, versus portable-device-dependent outputs for the Quadro.

The release timeline also separates them. The GeForce GTX 660 launched on 2012-09-05, while the Quadro K4100M launched on 2013-07-22. The GTX 660 has a launch MSRP of 229 USD, and the Quadro K4100M has a launch MSRP of 1,499 USD. The GTX 660 is part of the GeForce 600 generation with a GeForce 500 predecessor and GeForce 700 successor. The Quadro K4100M belongs to the Quadro Kepler-M (Kx100M) generation, with a Quadro Fermi-M predecessor and Quadro Maxwell-M successor.

Where Each One Wins

The GeForce GTX 660 wins in raw compute and rasterization throughput. Its FP32 rating of 1.981 TFLOPS tops the Quadro K4100M's 1.627 TFLOPS. Its pixel rate of 20.64 GPixel/s and texture rate of 82.56 GTexel/s both exceed the Quadro's 16.94 GPixel/s and 67.78 GTexel/s. Memory bandwidth is another GTX 660 strength, at 144.2 GB/s versus 102.4 GB/s, which matters for any workload that streams large data sets. The Geekbench OpenCL score of 11347, a 24% advantage over the Quadro's 9149, positions the GTX 660 as the better choice for general-purpose compute tasks where OpenCL is the relevant API.

The Quadro K4100M wins in scenarios that benefit from more memory and more execution units. Its 4 GB VRAM is double the GTX 660's 2 GB, and its 256-bit bus is wider, even if bandwidth is lower. The 1152 shading units outnumber the GTX 660's 960, and the 96 TMUs and 32 ROPs are also higher counts. The Geekbench Metal score of 6662, a 35.4% lead over the GTX 660's 4305, shows that the Quadro architecture scales better in Metal-based graphics workloads. Its lower 100 W TDP and MXM form factor make it suited to portable workstation deployments where the GTX 660's dual-slot desktop footprint and 140 W TDP would not apply.

The Verdict

The data points to a clear split based on workload type. If the task is OpenCL compute, the GeForce GTX 660 is the stronger option. It delivers 24% higher OpenCL scores, higher FP32 throughput, higher pixel rate, higher texture rate, and faster memory bandwidth. Its average benchmark of 9022 also places it above the Quadro K4100M's 7906 overall. The GTX 660's nearest rival, the NVIDIA TITAN V CEO Edition, sits only 0.2% away from its average score, which shows the GTX 660 is competitive with much newer hardware in the database's measurements.

If the task runs through the Metal API, the Quadro K4100M takes the edge. Its Metal score of 6662 beats the GTX 660 by 35.4%, and its larger memory pool and wider bus provide capacity advantages that the GTX 660 cannot match. The Quadro K4100M also fits into portable systems as an MXM module with a 100 W TDP, making it the only option of the two for that form factor.

Users with desktop systems and a priority on compute performance should choose the GeForce GTX 660. Users who need a mobile workstation GPU with 4 GB of memory and strong Metal results should choose the Quadro K4100M. The choice is not about which is universally faster, it is about which API and form factor match the intended use case.