This preview shows page 1. Sign up to view the full content.
Unformatted text preview: explicit parallelism, predication and speculation.
These terminologies are explained below.
1. Explicit Parallelism. EPIC technology breaks through the sequential nature of today's
conventional processor architectures by allowing the software to communicate explicitly
to the processor when operations can be done in parallel. For this it uses tighter coupling
between the compiler and the processor, and enables the compiler to extract maximum
parallelism in the original code and explicitly describe it to the processor. For example
in case of IA-64, all instructions are of 32-bit fixed length. At compile time, the
compiler detects which of the instructions can be executed in parallel. It then reorders
them and groups them in such a way that instructions belonging to separate groups can be
executed in parallel. At runtime, this explicit parallelism information provided by the
compiler is exploited by the processor to execute the instructions faster.
Predication. Predication is a technique for obtaining improved performance by
reducing the number of branches and branch mispredicts. Once again the help of
compiler is first taken to reorder the instructions to reduce the number of branches as
much as possible at compile time. The conventional processors use "branch prediction"
technique in which the processor predicts which way a branch will fork and speculatively
executes instructions along the predicted path. At the time of execution of the branch
instruction, if the prediction is found to be found to be correct, the processor gains
performance improvement because the instructions lying in the path to be executed now
have already been executed and their results can be directly used now by the processor.
However, if the prediction is found to be wrong, the results of execution of the predicted
path are discarded and the instructions of the correct path are now taken up for execution.
However, the EPIC technology uses "branch predication" instead of "branch prediction".
In this technique, instead...
View Full Document
This document was uploaded on 04/07/2014.
- Spring '14