TLP means that two or more software program execution threads are processed simultaneously, directly inside a single processor. Hyper-Threading can benefit both individual multi-threaded applications and the multi-tasking of single-threaded applications. It can also apply to operating system code itself, including services and background housekeeping threads. In all cases, to avoid performance problems, it's best if the two different threads do not frequently use the same processor resources at the same time.
Hyper-Threading works in a form of simultaneous multi-threading technology (SMT), Hyper-Threading technology allows multiple threads of software applications to be run simultaneously on one processor by duplicating the architectural state on each processor while the same processor execution resources is shared. For multiprocessor-capable software applications, the Hyper-Threading based processor is considered two separate logical processors on which the software applications can run without modification. Also, each logical processor responds to interrupts independently. The first logical processor can track one software thread, while the second logical processor tracks another software thread simultaneously. Because the two idle if only one thread was executing. This results in an increased utilization of the execution resources within each physical processor.
Hyper-Threading can provide sizeable performance benefits as seen in Intel benchmarks and certain industry benchmarks. However, cases can exist where contention for processor resources or poorly written multithreaded applications (where one thread is executing far more work than a second thread) can result in reduced performance in Hyper-Threaded mode. In fact, the most common cause for HT not providing benefit to two application threads executed simultaneously also shows up in a real dual-processor environment, which points to poor multithreaded software design.
