Back to search
Georgia Institute of Technology
Neural Methods for Resolving Hard-to-Predict Branches
Abstract
dc:description.abstractThis work presents a new category of branch predictors designed to be addendums to existing state of the art prediction mechanisms. We call these neural network inspired predictors Shallow Online Neural (SON) Predictors as they utilize easily quantizable shallow networks and exhibit online training as opposed to other related works. This predictor is apt as both a branch prediction scheme and as a TAGE confidence predictor.
Degree
thesis:*- Level thesis:degree_level
- Masters
- Department dc:contributor.department
- Computer Science
- Grantor dc:publisher
- Georgia Institute of Technology
- Year dc:date.issued
- 2021
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Gupta, Pulkit
- Advisor dc:contributor.advisor
-
- Conte, Thomas M.
Subjects
dc:subject × 4Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- http://hdl.handle.net/1853/66176