speculation (speculative loading)
Speculation (also known as speculative loading ), is a process implemented in Explicitly Parallel Instruction Computing ( EPIC ) processors and their compiler s to reduce processor-memory exchanging bottlenecks or latency by putting all the data into memory in advance of an actual load instruction. Invalid changes or exceptions to the load are delayed and cross-checked until the processor finally resolves them.
Intel's IA-64 architecture is based on EPIC and the first processor in the IA-64 line, the Itanium , uses speculation.