Evaluation of x86 and ARM architectures using compute-intensive workloads

In the 1990s, the tech giants, Intel and ARM, were oriented towards very different goals. Intel's x86 concentrated on peak performance for PCs and servers, while ARM focused on increasing energy efficiency mainly for mobile devices and wearables. But today, after three decades, with a massive ad-vancement in technology, there is a need to work towards similar goals. Intel has started manufacturing processors for handheld devices, and ARM has begun delving into servers. Since both x86 and ARM-based processors are now competitors, it is essential to compare their performance for a variety of applications. Multiple studies have been conducted in the past using actual hardware measurements, to compare the performance of different ISAs. In this work, a comparative study is done using the gem5 simulator with MinorCPU, an in-order CPU model. Different memory models are also used to observe how memory technology affects performance. The Sieve of Eratosthenes algorithm has been used to carry out the comparative study. Clocks-per-Instruction (CPI), instruction mix, Ll and L2 cache miss rate, DRAM bandwidth, and average power are the parameters used for the comparison.