Why your computer stopped getting faster and started getting more cores instead
In short: For decades, engineers boosted performance by cranking up clock speeds. But they eventually hit a wall of heat and power consumption. The solution was to stop building one faster processor and start packing multiple cores onto a single chip, fundamentally changing how software must be written to keep up.
By the early 2000s, the era of rapid clock speed increases hit a hard limit known as the power wall. As operating frequencies climbed into the gigahertz range, processors began generating exponentially more heat, making further speed gains unsustainable. Simultaneously, engineers faced the memory wall—where processor speeds outpaced memory—and the instruction-level parallelism (ILP) wall, where it became increasingly difficult to find enough independent tasks within a single stream of code to keep a high-performance core busy.
The industry pivoted to multi-core architectures, placing two or more central processing units on a single integrated circuit. This design offers significant advantages: signals travel shorter distances between cores, allowing for faster cache coherency and lower energy consumption compared to separate chips. By reusing proven core designs, manufacturers also reduced the risk of design errors. However, this shift introduced a new bottleneck: software. Unless applications are specifically written to use multiple threads, they cannot distribute work across cores, leaving much of the hardware's potential power untapped.
The challenge of parallel programming remains a significant hurdle. Writing code that splits tasks across cores is notoriously difficult, often introducing subtle bugs related to thread safety and shared data. While some tasks, like video rendering, can be easily parallelized, others—such as entropy decoding—are inherently serial. As we look toward a future of many-core processors with hundreds of units, the ability of developers to refactor software to exploit these resources has become the single greatest constraint on computer performance, threatening to create an insurmountable ceiling if the software model does not evolve.
Opens the Intellect+ feed with this card first. Swipe up for the next one.
Sources
- Summary by Intellect · source: Multi-core processor
To cite this card, link to https://intellectplus.co.uk/app/technology-multi-core-processor.
How we source, check and correct cards: Editorial standards.
General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.
Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?
Related
More in Computing and information technology · More in Technology · Latest · A–Z · All topics