Hyper-Threading

Hyper-Threading is Intel’s implementation of simultaneous multithreading. A supported physical core exposes more than one architectural execution context to software. The operating system can schedule independent threads onto those logical processors. The contexts share substantial core machinery, so logical-processor count is different from physical-core count.

Separate architectural state allows independent instruction streams to make progress. Intel’s 2002 technical journal describes the architecture of its early implementation, including duplicated architectural state and shared physical execution resources. A thread can use resources that another thread leaves available. The potential gain comes from utilization rather than creating a second complete core.

Shared execution resources can also create contention. Two threads may need the same functional units, caches, or memory bandwidth. If one thread already uses the limiting resource heavily, another may add little throughput or increase individual task latency. The result is workload dependent and cannot be predicted from the presence of the feature alone.

Scheduling and measurement need an accurate hardware topology. Software should distinguish logical siblings from separate physical cores and separate sockets. A throughput-oriented service can prefer a different policy from a latency-sensitive task. Benchmarking should report whether the feature is enabled and which execution contexts the workload uses.

The technology complements physical multi-core integration. A processor can contain several cores, each with multiple logical contexts, or use a different combination. Feature support and behavior vary by model. Hyper-Threading is best understood as a resource-sharing design technique evaluated through the actual application’s throughput, latency, and operating requirements.

Sources and further reading. [Intel Technology Journal, Hyper-Threading, 2002](https://www.intel.com/content/dam/www/public/us/en/documents/research/2002-vol06-iss-1-intel-technology-journal.pdf); [Intel technical explanation](https://www.intel.com/content/www/us/en/gaming/resources/hyper-threading.html).

Read Hyper-Threading as a simultaneous-multithreading technique that shares core resources between independently scheduled instruction streams.