HyperThreading

Hyper-Threading is Intel's proprietary simultaneous multithreading technology that allows a single physical CPU core to present two logical processors to the operating system. By exposing two logical processors, the operating system can schedule more than one thread per core, potentially improving utilization of processor resources and throughput on multi-threaded workloads.

The two logical processors share most of the core's execution resources, including pipelines and caches. The technology keeps two architectural states, enabling the processor to issue instructions from two threads in parallel. It does not double the performance of a core, and benefits are highly workload dependent.

Hyper-Threading was introduced by Intel in 2002 on some Xeon processors and later extended to mainstream Core CPUs. It is configurable in the system BIOS or UEFI and can be disabled where security or policy requires. It is commonly used in servers and desktop systems with multi-threaded tasks.

Performance gains vary; HT tends to improve throughput for multi-threaded applications but can offer little or even reduce performance for heavily memory-bound or single-threaded workloads due to shared resources. Security considerations have led some deployments to disable SMT/Hyper-Threading in sensitive environments to mitigate side-channel risks. Power and heat also increase modestly when enabled.