TY - GEN
T1 - Asymmetrically banked value-aware register files
AU - Wang, Shuai
AU - Yang, Hongyan
AU - Hu, Jie
AU - Ziavras, Sotirios G.
PY - 2007
Y1 - 2007
N2 - Designing high-performance low-power register files is of critical importance to the continuation of current performance advances in wide-issue and deeply-pipelined super-scalar microprocessors. In this paper, we propose a new microarchitecture, the asymmetrically-banked value-aware register file (AB-VARF), to exploit the prevailing narrow-width register values for low-latency and power-efficient register file designs. The register bit-widths of different banks in our AB-VARF register files are specifically customized to capture different narrow-width values. Augmented with a value width predictor, the register renaming logic is slightly tuned to rename predicted narrow-width registers to the corresponding narrow-width banks. Our experimental evaluation with SPEC CINT2000 benchmark suites shows that AB-VARF reduces the energy consumption by 92.6% over a conventional register file, on the average, at the cost of a 6.6% performance loss to an ideal 1-cycle monolithic register file.
AB - Designing high-performance low-power register files is of critical importance to the continuation of current performance advances in wide-issue and deeply-pipelined super-scalar microprocessors. In this paper, we propose a new microarchitecture, the asymmetrically-banked value-aware register file (AB-VARF), to exploit the prevailing narrow-width register values for low-latency and power-efficient register file designs. The register bit-widths of different banks in our AB-VARF register files are specifically customized to capture different narrow-width values. Augmented with a value width predictor, the register renaming logic is slightly tuned to rename predicted narrow-width registers to the corresponding narrow-width banks. Our experimental evaluation with SPEC CINT2000 benchmark suites shows that AB-VARF reduces the energy consumption by 92.6% over a conventional register file, on the average, at the cost of a 6.6% performance loss to an ideal 1-cycle monolithic register file.
UR - http://www.scopus.com/inward/record.url?scp=36349037023&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=36349037023&partnerID=8YFLogxK
U2 - 10.1109/ISVLSI.2007.27
DO - 10.1109/ISVLSI.2007.27
M3 - Conference contribution
AN - SCOPUS:36349037023
SN - 0769528961
SN - 9780769528960
T3 - Proceedings - IEEE Computer Society Annual Symposium on VLSI: Emerging VLSI Technologies and Architectures
SP - 363
EP - 368
BT - Proceedings - IEEE Computer Society Annual Symposium on VLSI
T2 - IEEE Computer Society Annual Symposium on VLSI: Emerging VLSI Technologies and Architectures, ISVLSI'07
Y2 - 9 March 2007 through 11 March 2007
ER -