speculative execution


Also found in: Wikipedia.

speculative execution

(processor)
A technique allows a superscalar processor to keep its functional units as busy as possible by executing instructions before it is known that they will be needed.

The Intel P6 uses speculative execution.

Compare branch prediction, speculative evaluation.
References in periodicals archive ?
The authors wrote, "Meltdown and Spectre represent a class of recently discovered cache timing side-channel attacks that leverage the effects of out-of-order and speculative execution on cache state.
Speculative execution predicts what code a program will use next.
The root cause of the weaknesses is a concept called speculative execution, whereby the CPU preemptively executes instructions that it anticipates ahead of time.
Security issues related to modern microprocessors and speculative execution have been discovered.
AMD (AMD) CTO Mark Papermaster said: "The public disclosure on January 3rd that multiple research teams had discovered security issues related to how modern microprocessors handle speculative execution has brought to the forefront the constant vigilance needed to protect and secure data.
In the case of computers, speculative execution is used to
Intel on Wednesday confirmed a report stating that its semiconductors contain a vulnerability based around a chip-processing technique called speculative execution.
Speculative execution is the lay term given to the method where a CPU predicts which calculations it needs to do beforehand based on what is happening and what it might need to do.
According to Intel, the exploit is a way for an attacker to observe the content of privileged memory, exploiting a CPU technique called speculative execution to circumvent expected privilege levels.
intelligence community-sponsored research, BluVector has developed two revolutionary detection technologies, which are replacing signatures with self-evolving machine learning and speculative execution.
Speculative Execution (SE) has been presented as an approach to the problem above by backing up those delayed jobs from low-performance machines to higher ones.
To overcome this challenges we proposed some mechanisms that mainly focuses on the resource utilization by using the dynamic allocation technique to meet their deadlines and Smart Speculative Execution strategy by running straggled tasks in the fast worker nodes there by maximizing the Data locality which in turn preserves the time and cost.