TY - GEN
T1 - A media-oriented vector architectural extension with a high bandwidth cache system
AU - Gao, Ye
AU - Shoji, Naoki
AU - Egawa, Ryusuke
AU - Takizawa, Hiroyuki
AU - Kobayashi, Hiroaki
PY - 2012
Y1 - 2012
N2 - Although modern SIMD extensions enable to improve the performance of MMAs, the limited parallel-processing capability and low memory bandwidth are their drawbacks to execute next generation MMAs. This paper proposes a novel media-oriented vector architectural extension (MVPX) to aim at accelerating a wide range of next generation MMAs. MVPX consists of an out-of-order vector processing mechanism (OVPM) and a high bandwidth cache system (MVP-cache). OVPM enables MVPX to take advantage of high data level parallelism for various MMAs. Meanwhile, MVP-cache can reduce the data transfer from/to off-chip memory accesses, and thus increase the sustained memory bandwidth by exploiting reusable vector data. It has a low power/area overhead approach to increasing memory bandwidth, compared with approaches to increasing the off-chip memory bandwidth. As a result, MVPX expands the potential of GPPs for next generation media processing in a power efficient way.
AB - Although modern SIMD extensions enable to improve the performance of MMAs, the limited parallel-processing capability and low memory bandwidth are their drawbacks to execute next generation MMAs. This paper proposes a novel media-oriented vector architectural extension (MVPX) to aim at accelerating a wide range of next generation MMAs. MVPX consists of an out-of-order vector processing mechanism (OVPM) and a high bandwidth cache system (MVP-cache). OVPM enables MVPX to take advantage of high data level parallelism for various MMAs. Meanwhile, MVP-cache can reduce the data transfer from/to off-chip memory accesses, and thus increase the sustained memory bandwidth by exploiting reusable vector data. It has a low power/area overhead approach to increasing memory bandwidth, compared with approaches to increasing the off-chip memory bandwidth. As a result, MVPX expands the potential of GPPs for next generation media processing in a power efficient way.
KW - SIMD
KW - cache
KW - memory bandwidth
KW - multimedia application
KW - vector architecture
UR - http://www.scopus.com/inward/record.url?scp=84864057170&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84864057170&partnerID=8YFLogxK
U2 - 10.1109/COOLChips.2012.6216588
DO - 10.1109/COOLChips.2012.6216588
M3 - Conference contribution
AN - SCOPUS:84864057170
SN - 9781467312028
T3 - Symposium on Low-Power and High-Speed Chips - Proceedings for 2012 IEEE COOL Chips XV
BT - Symposium on Low-Power and High-Speed Chips - Proceedings for 2012 IEEE COOL Chips XV
T2 - 15th IEEE Symposium on Low-Powerand High-Speed Chips, COOL Chips XV
Y2 - 18 April 2012 through 20 April 2012
ER -