GPU Comparison

# NVIDIA GeForce GTX 660 vs NVIDIA GeForce MX330

The NVIDIA GeForce GTX 660 and NVIDIA GeForce MX330 represent two very different eras of GPU design, with the desktop-oriented Kepler part from 2012 squaring off against a mobile Pascal chip from 2020. The benchmark data reveals a clear overall winner: the GTX 660 takes both head-to-head tests, winning the Geekbench OpenCL test with a score of 11,347 against the MX330's 7,896 (a 43.7% advantage) and the Geekbench Vulkan test with 11,415 versus 9,019 (a 26.6% lead). The GTX 660 also holds a higher average benchmark score of 9,022 compared to the MX330's 8,458, and it sits at the 45th percentile among all GPUs versus the MX330's 43rd percentile. Yet the MX330 counters with dramatically superior efficiency and newer feature support, making this comparison more nuanced than raw performance alone suggests.

Head-to-Head Benchmarks

The GTX 660's most decisive victory comes in the Geekbench OpenCL test, where it posts a score of 11,347 against the MX330's 7,896. That 43.7% delta is substantial, and it reflects the GTX 660's much larger compute infrastructure, 960 shading units versus 384, along with 80 texture mapping units and 24 ROPs compared to 24 TMUs and 16 ROPs on the MX330. The GTX 660 also generates far higher raw throughput: its FP32 rating of 1.981 TFLOPS nearly doubles the MX330's 1,224.2 GFLOPS, and its texture rate of 82.56 GTexel/s more than doubles the MX330's 38.26 GTexel/s.

The Vulkan test narrows the gap somewhat but still favors the older card. The GTX 660 scores 11,415 versus 9,019 for the MX330, a 26.6% margin. Interestingly, the MX330's Vulkan score of 9,019 is actually higher than its OpenCL result of 7,896, suggesting the Pascal architecture's newer API implementation, Vulkan 1.4 support versus 1.2.175 on the GTX 660, helps it close the performance gap in that workload. The GTX 660's Vulkan score of 11,415 is nearly identical to its OpenCL score of 11,347, indicating consistent performance across both compute APIs.

The MX330 does claim one notable victory: pixel rate. Its 25.50 GPixel/s exceeds the GTX 660's 20.64 GPixel/s, a consequence of its higher boost clock of 1,594 MHz versus 1,032 MHz on the GTX 660. That means in pixel-bound workloads that don't stress the shading units heavily, the MX330 can actually outperform its larger rival. However, this advantage is narrow and doesn't translate into a win in either of the actual benchmark tests.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 660 averages 9,022 across all benchmark tests, placing it at the 45th percentile of all GPUs. The NVIDIA GeForce MX330 averages 8,458 and sits at the 43rd percentile. The GTX 660's nearest rival by average score is the NVIDIA TITAN V CEO Edition at 9,037 (a 0.2% difference), while the MX330's closest competitor is the AMD Radeon HD 8870M at 8,462 (a 0% delta).

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

A: The GTX 660 scores 11,347 in Geekbench OpenCL, which is 43.7% higher than the MX330's 7,896. This is the largest performance gap between the two cards in any benchmark test.

Q: Does the MX330 have any architectural advantages despite losing the benchmarks?

A: Yes. The MX330 is built on a 14 nm process at Samsung compared to the GTX 660's 28 nm TSMC node, and it supports DirectX 12 (12_1) versus the GTX 660's DirectX 12 (11_0). The MX330 also supports Vulkan 1.4, while the GTX 660 is limited to Vulkan 1.2.175. Its power draw of 10 W is dramatically lower than the GTX 660's 140 W TDP.

Q: What is the memory configuration difference between the two cards?

A: Both have 2 GB of GDDR5 memory, but the GTX 660 uses a 192-bit bus with 144.2 GB/s of bandwidth, while the MX330 uses a 64-bit bus with 56.06 GB/s. The MX330's memory runs at 7 Gbps effective versus 6 Gbps effective on the GTX 660, but the narrower bus limits overall bandwidth.

Q: Which card has better transistor density?

A: The MX330 packs 1,800 million transistors into a 74 mm² die, yielding 24.3M transistors per mm². The GTX 660 has 2,540 million transistors on a 221 mm² die, resulting in 11.5M per mm². The MX330's newer 14 nm process enables more than double the transistor density.

Q: How does the MX330 compare to its nearest rivals?

A: The MX330's average score of 8,458 places it roughly level with the AMD Radeon HD 8870M (8,462, 0% delta) and AMD Radeon 880M (8,436, 0.3% higher). It trails the Intel Arc A380 by 1.2% (8,558) and beats the NVIDIA GeForce GTX 675MX by 0.4% (8,427).

Where Each One Wins

The GTX 660 is the clear choice for any workload that stresses raw compute throughput. Its 960 shading units, 80 TMUs, and 1.981 TFLOPS of FP32 performance make it far better suited for GPU-accelerated computing tasks, OpenCL-heavy applications, and Vulkan-based games that can leverage its higher texture fill rate of 82.56 GTexel/s. The benchmark data confirms this: it wins both head-to-head tests decisively, with margins of 43.7% in OpenCL and 26.6% in Vulkan. Its 144.2 GB/s of memory bandwidth also gives it a massive advantage in bandwidth-sensitive scenarios, more than double the MX330's 56.06 GB/s. For desktop users with a PCIe 3.0 x16 slot and a 300 W suggested PSU, the GTX 660 offers substantially more performance per frame.

The MX330 wins in every category related to efficiency and portability. Its 10 W TDP is a tiny fraction of the GTX 660's 140 W, and it requires no power connectors at all, the GTX 660 needs a single 6-pin connector. The MX330's IGP slot width means it's designed for integration into laptops and compact systems, whereas the GTX 660 is a dual-slot card measuring 241 mm (9.5 inches) in length. The MX330 also offers newer API support with DirectX 12 (12_1) and Vulkan 1.4, which could matter for future software compatibility. Its higher pixel rate of 25.50 GPixel/s also suggests it can hold its own in simple 2D rendering or pixel-shader-light workloads. For users prioritizing battery life, low heat output, and a compact form factor, the MX330 is the only sensible option, but they should expect roughly 26-44% lower performance in compute-heavy tasks.

Specification Differences

The two cards differ substantially across nearly every specification category. The GTX 660 uses the GK106 chip with 2,540 million transistors on a 221 mm² die, while the MX330 uses the GP108B with 1,800 million transistors on a 74 mm² die. The GTX 660's base clock is 980 MHz with a boost of 1,032 MHz; the MX330 runs at 1,531 MHz base and 1,594 MHz boost. Memory speeds differ too: the GTX 660 runs at 1,502 MHz (6 Gbps effective), while the MX330 runs at 1,752 MHz (7 Gbps effective).

The bus interface also differs: the GTX 660 uses PCIe 3.0 x16, while the MX330 uses PCIe 3.0 x4. Display outputs are another differentiator, with the GTX 660 offering 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2, while the MX330's outputs are described simply as "Portable Device Dependent." Power requirements are starkly different: the GTX 660 has a 140 W TDP, a dual-slot cooler, and a single 6-pin power connector, while the MX330 has a 10 W TDP, an IGP form factor, and no power connectors. The GTX 660 requires a 300 W suggested PSU; the MX330 lists no suggested PSU at all. Physical dimensions exist only for the GTX 660 (241 mm / 9.5 inches long), as the MX330's mobile design has no listed dimensions.

Architecture Differences

The architectural gulf between these two GPUs spans three generations of NVIDIA design philosophy. The GTX 660 is built on the Kepler architecture using TSMC's 28 nm process, while the MX330 uses the Pascal architecture on Samsung's 14 nm node. That process shrink enables the MX330 to achieve a transistor density of 24.3M per mm², more than double the GTX 660's 11.5M per mm², despite having fewer total transistors (1,800 million versus 2,540 million).

The compute resource allocation differs dramatically. The GTX 660 fields 960 shading units, 80 TMUs, and 24 ROPs, producing a pixel rate of 20.64 GPixel/s and a texture rate of 82.56 GTexel/s. The MX330 has 384 shading units, 24 TMUs, and 16 ROPs, with a pixel rate of 25.50 GPixel/s and a texture rate of 38.26 GTexel/s. Neither card has ray tracing cores or tensor cores, but the MX330 does list FP16 support at 19.13 GFLOPS (1:64 ratio), a feature absent from the GTX 660's specifications.

API support reveals the generational gap most clearly. The MX330 supports DirectX 12 (12_1) and Vulkan 1.4, while the GTX 660 is limited to DirectX 12 (11_0) and Vulkan 1.2.175. Both support OpenGL 4.6. The GTX 660 belongs to the GeForce 600 generation with a release date of September 2012, while the MX330 comes from the GeForce MX (3xx) generation released in February 2020. The GTX 660 has a listed successor (GeForce 700) and predecessor (GeForce 500), while the MX330 lists neither. Both are end-of-life products, but the GTX 660 carries a launch MSRP of 229 USD, the MX330 has no launch MSRP listed.