Abstract
A general paradigm for converting any profile-insensitive list to a profile-sensitive scheduler is proposed. This technique is developed via a theoretical analysis of a simplified abstract model of the general problem of profile-driven scheduling over any cyclic code region, yielding a scoring measure for ranking branch instructions. The ranking digests the profile information and has the useful property that scheduling with respect to rank is probably good for minimizing the expected completion time of the region, within the limits of abstraction. While the ranking scheme is computationally intractable in the most general case, this is practicable for super blocks.
Original language | English (US) |
---|---|
Pages (from-to) | 58-67 |
Number of pages | 10 |
Journal | Proceedings of the Annual International Symposium on Microarchitecture |
State | Published - 1996 |
Externally published | Yes |
Event | Proceedings of the 1996 29th Annual IEEE/ACM International Symposium on Microarchitecture, MICRO-29 - Paris, Fr Duration: Dec 2 1996 → Dec 4 1996 |
ASJC Scopus subject areas
- Hardware and Architecture
- Software