Abstract:
Register files represent a substantial portion of the energy budget in modern processors, and are growing rapidly with the trend towards larger Instruction Level Parallelism (ILP). The energy cost of a register file access depends greatly on the register file circuitry used. This paper compares various register file circuitry techniques for their energy efficiencies, as a function of the architectural parameters such as the number of registers and the number of ports. The Port Priority Selection technique combined with differential reads and low-swing writes was found to be the most energy efficient and provided significant energy savings compared to traditional approaches in the case of large register files. The dependence of register file access energy upon technology scaling is also studied. However, as this paper shows, it appears that none of these will be enough to prevent centralized register files from becoming the dominant power component of next-generation superscalar computers, and alternative methods for inter-instruction communication need to be developed.
Citations
|
344
|
Complexityeffective superscalar processors
– Palacharla, Jouppi, et al.
- 1997
|
|
39
|
Register File Design Considerations in Dynamically Scheduled Processors
– Farkas, Jouppi, et al.
- 1996
|
|
37
|
Trends in low-power RAM circuit technologies
– Itoh, Sasaki, et al.
- 1995
|
|
33
|
Energy Consumption Modeling and Optimization for SRAM’s
– Evans, Franzon
- 1995
|
|
20
|
A 9-ns 1.4 gigabyte 17-ported CMOS register file
– Jolly
- 1991
|
|
20
|
A Three Dimensional Register File for Superscalar Processors
– Tremblay, Joy, et al.
- 1995
|
|
15
|
Future CMOS scaling and reliability
– Hu
- 1993
|
|
9
|
Technology Challenges for Integration Near and Below
– Asai, Wada
- 1997
|
|
8
|
A 14-port 3.8ns 116-word 64b Read-Renaming Register File
– Asato, Montoye, et al.
- 1995
|
|
7
|
et al., "A 200MHz 64b Dual-Issue CMOS Microprocessor
– Dobberpuhl
- 1992
|
|
6
|
Driving source-line cell architecture for sub-1-V high-speed low-power applications
– Mizuno, Nagano
- 1996
|
|
6
|
et al., "A 160-MHz, 32-b, 0.5-W CMOS RISC Microprocessor
– Montanaro
- 1996
|
|
6
|
Mismatch Sensitivity of a Simultaneously Latched CMOS Sense Amplifier
– Sarpeshkar, Wyatt, et al.
- 1991
|
|
4
|
SuperSPARC: A fully integrated superscalar processor
– Blanck, Krueger
- 1991
|
|
4
|
A Current Direction Sense Technique for Multiport SRAM’s
– Izumikawa, Yamashina
- 1996
|
|
3
|
A High-Speed Clamped Bit-Line currentMode Sense Amplifier
– Blalock, Jaeger
- 1991
|
|
3
|
Experimental Investigation of the Minimum Signal for Reliable Operation of DRAM Sense Amplifier
– Geib, Weber, et al.
- 1992
|
|
2
|
A High-Speed Sensing Scheme for 1T Dynamic RAM's Utilizing the Clamped Bit-Line Sense Amplifier
– Blalock, Jaeger
- 1992
|
|
2
|
A 1-V Operating 256-kb Full-CMOS SRAM
– Sekiyama, Seki, et al.
- 1992
|
|
1
|
et al, "SRAM with Multiple Sensing Schemes", Notre Dame CSE
– Brockman, Zyuban
- 1997
|
|
1
|
A 12.5-ns 16-Mb
– Ishibashi
- 1994
|
|
1
|
Automated Bias Control (ABC) Circuit for High-Performance VLSI's
– Kuroda, Fukunaga
- 1992
|
|
1
|
et al, "Sub-1-V Swing Internal Bus Architecture for Future Low-Power ULSI's
– Nakagome, Itoh
- 1993
|
|
1
|
Multiport Register File Memory Cell Configuration for Read Operation." Assignee: Sun Microsystems
– Aug
- 1997
|
|
