ISLPED'06 Proceedings of the 2006 International Symposium on Low Power Electronics and Design 2006
DOI: 10.1109/lpe.2006.4271804
|View full text |Cite
|
Sign up to set email alerts
|

Selective Writeback: Exploiting Transient Values for Energy-Efficiency and Performance

Abstract: Today's superscalar microprocessors use large, heavily-ported physical register files (RFs) to increase the instruction throughput. The high complexity and power dissipation of such RFs mainly stem from the need to maintain each and every result for a large number of cycles after the result generation. We observed that a significant fraction (about 45%) of the result values are delivered to their consumers via the bypass network (consumed "on-the-fly") and are never read out from the destination registers. In … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1

Citation Types

0
1
0

Year Published

2007
2007
2008
2008

Publication Types

Select...
1
1

Relationship

1
1

Authors

Journals

citations
Cited by 2 publications
(1 citation statement)
references
References 23 publications
0
1
0
Order By: Relevance
“…The major drawback of the late allocation schemes is in the form of non-trivial increases in the datapath complexity due to the need to: (a) support several levels of register mapping tables, (b) perform various associative searches on the rename table and issue queue after the reassignment of mappings and (c) avoid potential deadlocks. The second set of techniques aim at reducing the register file pressure by using the early deallocation of physical registers [12], [14], [15], [16], [9], [31], [32]. While these mechanisms differ in the timing and manner of register deallocation, the additional logic needed to support precise state reconstruction and guarantee correctness of the execution is fairly complex, sometimes requiring additional accesses to the rename table [12] or register state checkpointing support [9,14].…”
Section: Register File Optimizationsmentioning
confidence: 99%
“…The major drawback of the late allocation schemes is in the form of non-trivial increases in the datapath complexity due to the need to: (a) support several levels of register mapping tables, (b) perform various associative searches on the rename table and issue queue after the reassignment of mappings and (c) avoid potential deadlocks. The second set of techniques aim at reducing the register file pressure by using the early deallocation of physical registers [12], [14], [15], [16], [9], [31], [32]. While these mechanisms differ in the timing and manner of register deallocation, the additional logic needed to support precise state reconstruction and guarantee correctness of the execution is fairly complex, sometimes requiring additional accesses to the rename table [12] or register state checkpointing support [9,14].…”
Section: Register File Optimizationsmentioning
confidence: 99%