AMD Radeon 660M vs NVIDIA GeForce GTX 660 Ti Comparison

The NVIDIA GeForce GTX 660 Ti and the AMD Radeon 660M represent two very different approaches to graphics processing: a dedicated desktop card from 2012 and an integrated mobile GPU from 2022. The benchmark data shows a clear, though nuanced, performance hierarchy, with the older discrete card leading in both available tests, but the margin varies significantly depending on the workload.

Head-to-Head Benchmarks

The head-to-head comparison consists of two Geekbench tests: OpenCL and Vulkan. The data shows a decisive victory for the NVIDIA GeForce GTX 660 Ti in the OpenCL test. It scores 15113 points against the AMD Radeon 660M's 12876 points, a substantial delta of 17.4%. This is the largest performance gap between the two in any metric, indicating that for general-purpose compute workloads using OpenCL, the GTX 660 Ti's dedicated hardware provides a significant advantage over the Radeon 660M's integrated solution.

The Vulkan test presents a much tighter contest. Here, the NVIDIA GeForce GTX 660 Ti scores 15012 points, while the AMD Radeon 660M comes very close with 14748 points. The delta is just 1.8% in favor of the NVIDIA card. This near-parity suggests that in modern graphics API workloads, the Radeon 660M's newer architecture and features manage to close the gap considerably, despite its older counterpart's dedicated memory and higher raw compute specifications. The GTX 660 Ti still wins, but the margin is remarkably slim.

Looking at the average benchmark score across both tests, the NVIDIA GeForce GTX 660 Ti posts a figure of 15063, which places it in the 57th percentile of all GPUs. Its nearest rival, the AMD Radeon Pro 560X, scores 15082, a delta of -0.1%, meaning the GTX 660 Ti is essentially tied with it. The AMD Radeon 660M, with an average score of 13812, sits in the 55th percentile. Its closest competitor is the NVIDIA RTX A2000 Mobile, which scores 13821, a delta of -0.1%, again showing a statistical tie. The difference in average score between the two reviewed products is 1251 points, a gap of roughly 9%.

The win count is 2-0 in favor of the NVIDIA GeForce GTX 660 Ti. However, the data's story is more complex than a simple sweep. The 17.4% OpenCL margin is a statement of raw compute power, while the 1.8% Vulkan margin confirms architectural efficiency. The GTX 660 Ti's 2.634 TFLOPS of FP32 performance and 144.2 GB/s of memory bandwidth are clear assets, but the Radeon 660M's RDNA 2.0 architecture, with its 6 ray tracing cores, allows it to remain competitive in a modern API context.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce GTX 660 Ti has a higher average benchmark score of 15063, compared to the AMD Radeon 660M's 13812.

Q: How large is the performance difference in the OpenCL benchmark?

A: The NVIDIA GeForce GTX 660 Ti leads by 17.4% in the Geekbench OpenCL test, scoring 15113 versus the AMD Radeon 660M's 12876.

Q: Is the AMD Radeon 660M competitive in any benchmark?

A: Yes, in the Geekbench Vulkan test, the AMD Radeon 660M scores 14748, which is only 1.8% behind the NVIDIA GeForce GTX 660 Ti's score of 15012.

Q: How does the AMD Radeon 660M's average score compare to its nearest rival?

A: The AMD Radeon 660M's average score of 13812 is statistically tied with its nearest rival, the NVIDIA RTX A2000 Mobile, which scores 13821 (a delta of -0.1%).

Q: What is the percentile ranking for each GPU?

A: The NVIDIA GeForce GTX 660 Ti is in the 57th percentile of all GPUs, while the AMD Radeon 660M is in the 55th percentile.

Q: Does the AMD Radeon 660M have any ray tracing cores?

A: Yes, the AMD Radeon 660M has 6 ray tracing cores, while the NVIDIA GeForce GTX 660 Ti has none.

The Verdict

The data points to a straightforward performance hierarchy: the NVIDIA GeForce GTX 660 Ti is the faster GPU in this comparison. It wins both benchmarks, with its most significant victory coming in the OpenCL test where it is 17.4% ahead. Its average benchmark score of 15063 is over 1200 points higher than the AMD Radeon 660M's 13812, a clear margin that reflects a raw performance advantage. For any workload that is heavily dependent on compute throughput, the GTX 660 Ti is the superior choice based on these numbers.

However, the verdict is not without its caveats. The Vulkan result, where the GTX 660 Ti wins by only 1.8%, shows that the AMD Radeon 660M is not far behind in modern graphics API performance. The Radeon 660M's 55th percentile ranking, while lower than the GTX 660 Ti's 57th, still places it in a similar tier. The choice between the two should be guided by the specific application. If the priority is maximum compute performance, the GTX 660 Ti is the clear winner. If the workload is primarily modern gaming or applications that leverage the Vulkan API, the Radeon 660M's performance is close enough to be a viable contender.

Specification Differences

The specification sheets for these two GPUs reveal fundamental differences. The NVIDIA GeForce GTX 660 Ti is a discrete card with 2 GB of dedicated GDDR5 memory on a 192-bit bus, providing 144.2 GB/s of bandwidth. In contrast, the AMD Radeon 660M is an integrated graphics processor (IGP) with a system-shared memory configuration, where the size, type, and bus width are all "System Shared," and bandwidth is "System Dependent." The GTX 660 Ti has a TDP of 150 W and requires two 6-pin power connectors, while the Radeon 660M has a TDP of just 40 W and requires no power connectors. The GTX 660 Ti is a dual-slot card measuring 241 mm, while the Radeon 660M is an IGP with no defined dimensions. The GTX 660 Ti offers a fixed set of display outputs (2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2), whereas the Radeon 660M's outputs are "Portable Device Dependent." The GTX 660 Ti's interface is PCIe 3.0 x16, and the Radeon 660M's is PCIe 4.0 x8. The GTX 660 Ti has a launch MSRP of 299 USD; the Radeon 660M has none listed.

Architecture Differences

Architecturally, these two GPUs are from different eras. The NVIDIA GeForce GTX 660 Ti is built on the Kepler architecture using the GK104 chip, manufactured on a 28 nm process at TSMC. It contains 3,540 million transistors on a 294 mm² die. The AMD Radeon 660M uses the RDNA 2.0 architecture with the Rembrandt chip, built on a much more advanced 6 nm process, also at TSMC. It packs 13,100 million transistors into a smaller 208 mm² die, resulting in a transistor density of 63.0M / mm² compared to the GTX 660 Ti's 12.0M / mm².

The compute resources differ significantly. The GTX 660 Ti has 1344 shading units, 112 texture mapping units (TMUs), and 24 raster output units (ROPs). The Radeon 660M has just 384 shading units, 24 TMUs, and 16 ROPs. The GTX 660 Ti has a higher texture rate of 109.8 GTexel/s, while the Radeon 660M is lower at 45.60 GTexel/s. However, the Radeon 660M has a slightly higher pixel rate of 30.40 GPixel/s versus the GTX 660 Ti's 27.44 GPixel/s. The Radeon 660M also features 6 ray tracing cores, a capability the GTX 660 Ti lacks entirely. Clock speeds differ, with the GTX 660 Ti running at a base of 915 MHz and a boost of 980 MHz, while the Radeon 660M runs higher at 1500 MHz base and 1900 MHz boost. The Radeon 660M also supports FP16 at 2.918 TFLOPS (2:1), a feature not listed for the GTX 660 Ti. API support also differs, with the Radeon 660M offering DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the GTX 660 Ti supports DirectX 12 (11_0) and Vulkan 1.2.175.

Where Each One Wins

The NVIDIA GeForce GTX 660 Ti is the clear winner in raw compute performance. Its 17.4% lead in the OpenCL benchmark and its higher average score of 15063 establish it as the superior choice for applications that leverage general-purpose compute on the GPU. Its dedicated 2 GB of GDDR5 memory and higher memory bandwidth of 144.2 GB/s are likely contributors to this performance advantage. This makes it the better option for tasks like video encoding, scientific simulations, or any workload that stresses the GPU's processing cores. The data also shows it holds a lead in the Vulkan API, although that lead is much smaller.

The AMD Radeon 660M's wins are more subtle. While it does not win any benchmark outright, its performance in the Vulkan test, coming within 1.8% of the GTX 660 Ti, is its major strength. This suggests that in modern gaming workloads or applications built on the Vulkan API, the Radeon 660M is nearly as fast as the older NVIDIA card. Its architectural advantages, including a more modern 6 nm process, a higher boost clock of 1900 MHz, and support for ray tracing, make it a more future-proof choice for newer software. Its significantly lower TDP of 40 W and lack of power connectors also make it the only logical choice for a portable, low-power system, whereas the GTX 660 Ti's 150 W TDP and dual-slot size require a desktop chassis with adequate power and space.