Boosting simulation performance: AnyLogic users gain from Mac ARM support

A man holding macbook with Mac ARM chip

Apple silicon has significantly shifted the computing landscape, particularly for Mac users. Apple’s switch from Intel's older chips to ARM-based M-series chips has promised significant improvements in simulation performance. This change isn't just about speeding up processors. It focuses on enhancing how applications leverage ARM's unique capabilities.

In the AnyLogic 8.9 release, we added full support for Macs on ARM architecture. Today, we'll explore how AnyLogic performs on both ARM and x86-64 Macs and why native support for ARM is a game-changer for you as a Mac user.

Why is native ARM support crucial for simulation performance?

Running an application designed for x86-64 architecture on Apple silicon chips requires Rosetta 2 translation. It converts the code initially made for Intel processors. This process increases the run time and may lead to slower performance than native applications.

Native support for Apple silicon optimizes the application to leverage the capabilities of M-series chips directly. In the case of AnyLogic, it results in faster simulation speed, improved power efficiency, and a better user experience.

AnyLogic performance and simulation speed: Mac ARM vs. x86-64

For those of you who use AnyLogic on Mac ARM machines, you may find that the experience with AnyLogic has become significantly smoother and faster. Let's have a look at some real examples of comparing the AnyLogic performance on Mac x86-64 and Mac ARM.

AnyLogic speed

Faster app launch

Comparison of AnyLogic launch on Mac ARM and Mac x86-64
(click to play the videos)

As you can see, the ARM version is almost two times faster than the x86-64. On ARM, AnyLogic launches in 12 seconds, whereas on x86-64, it takes 21 seconds. This cut in launch time means you can work more quickly, enhancing productivity and reducing downtime.

Faster model compilation

Comparison of the model compilation speed on Mac ARM and Mac x86-64
(click to play the videos)

Building a model on a Mac x86-64 takes 12 seconds, but it's twice as fast on a Mac ARM, only 6 seconds. This significant improvement allows you to iterate and test your models more efficiently, speeding up the development process.

Better experiment performance

The performance gains go beyond just launching and compiling. AnyLogic on ARM shows substantial improvements when running experiments or simulations.

Faster model runtime in a single run

Smoothness comparison of a model run on Mac ARM vs. Mac x86-64
(click to play the videos)

Faster simulation speeds provide quicker insights. This helps you complete complex projects sooner, process real-time data, and make faster decisions.

Multi-run performance difference

Multi-run simulation performance difference
(click to play the videos)

In scenarios involving multiple runs or batch processing, the ARM version significantly outperforms its x86-64 counterpart. This leads to better resource usage and the ability to handle more prominent and complex models efficiently.

Future trends in simulation software and hardware

The future of modeling software aims to enhance simulation speed and performance through deeper integration with AI, ML, and cloud computing, which necessitates advanced hardware capabilities to efficiently handle complex scenarios. In response to these advancements, we focus on improving parallel processing capabilities to accelerate simulation performance, which promises unprecedented efficiencies and cuts wait times.

Apple's shift to ARM-based M-series chips has dramatically enhanced simulation performance and model development speed for AnyLogic Mac users. With native support for ARM architecture, the software achieves faster compilation and increased simulation speeds. Looking ahead, continuous advancements in technology are expected to further improve simulation performance and capabilities.

To stay ahead of AnyLogic innovations and gain valuable insights, be sure to register for the AnyLogic Conference 2024.

Visit Conference page

Related posts