Intel Core processors automatically throttle their performance based on the processor workload and their thermal environment, reducing power consumption and also protecting hardware from overheating if not adequately cooled. However, for demanding defense applications where predictability is essential, performance throttling may have adverse effects. This paper presents the basics of Intel processor throttling, and the advanced cooling techniques needed to ensure consistent performance for defense applications.
Basic Concepts of Processor Throttling
Processor throttling (sometimes called dynamic frequency scaling') is used in computer architectures to adjust the clock frequency, or instructions executed per unit of time, of a processor. Simply put, the processor's clock is slowed down. Throttling back the clock frequency causes a processor to run more slowly, do less work, use less power and, as a consequence, generate less heat.
The benefits of throttling are directly related to these reductions. Reduced power use will extend battery life for laptops and decrease energy costs for IT systems. Decreases in heat generation reduce cooling costs and can also reduce environmental noise when demands on cooling systems (fans and air conditioning) are reduced.
Most significantly, reduced heat generation due to throttling may allow systems to avoid catastrophic failure from overheating. This is one of the primary reasons for introducing throttling into processor architectures like the Intel x86 and why it is often a 'hard-wired' certainty for some processors. When the die temperature reaches a given imminent failure' threshold, throttling is initiated with no user or system programming control. However, most x86 processors have multiple levels of performance throttling below this 'imminent failure' threshold where throttling is within programming control.
Processor Throttling: Commercial vs. Defense Environments
Today's Intel x86 processors are very attractive to defense designers for two primary reasons:
- Industry-leading price vs. performance, which continues to set the pace with recent additions of multiple processing cores, AVX2 vector (math) processing engines and powerful integrated Graphics Processing Units (GPUs), and;
- Broad software support, including commonly used operating systems, middleware, and tools.
Download the white paper to learn more.