ACA Phase 2

1
Advanced Computer Architecture
????? ??? ??? ????? ????? ????? ?????????? ??
?????? ???? Cache ??? ???? ????? ?????
Instruction Cache ?? ???????????? ??????????

• ???? ????? ???? 810190113 ??????? ???? 810190240
• ??????? ??? ? ???????? ? ??????? ?????
• ?? 1390

2
Our Work
Victim Cache for Filter Cache Energy-aware
Filter Cache Architecture for Multicore
Processors , 2010
3
Filter Cache Architecture

• increases the execution time
• 2.7 (single-core), 6.9 (2- cores), 7.4
  (4-cores),
• 7.6 (8-cores), 7.9 (16-cores)
• Reduces the energy consumption
• 12.9 (single-core), 9.9 (2-cores), 9.2
  (4-cores),
• 9.0 (8-cores), 8.9 (16-cores)

4
Victim Cache Architecture

• increases the execution time
• 3.7 on Average
• Reduces the energy consumption
• 8.6 on average

5
Our Work
or
Double Size Filter Cache DFilter cache 1 KB
Victim Cache for Filter Cache Filter cache 512
B Victim cache 512 B
6
Our Work

• Reduce Power of Instruction Cache in Multicore
  processors
• multicore FrameWork
• Benchmark
• Power Tools

7

• First Phase
• - SimpleScalar
• -Extension Multiprocessor Enhancements of the
  SimpleScalar Tool Set
• Spec2000 Single-thread benchmark
• Cacti 3.0
• Second Phase
• Multi2Sim A Simulation Framework to Evaluate
  Multicore Multithreaded Processors 2007
• Configurable memory hierarchy, cache coherence
• ParMibench An Open-Source Benchmark for
  Embedded Multiprocessor Systems 2010
• - Cacti 6.5

8
(No Transcript)
9
Our Work
Functional Unit 4 integer ALU, 4 FP ALU , 1 int Mul/Div , 1 FP Mul/Div
Traditional Cache L1 Ins 32KB , L1 Data 32KB 4-way, 32bytes lines L2 128KB
Filter Cache 512B Fully Associative, 1 Cycle Latency
Double Size Filter Cache 1KB, Fully Associative, 1 Cycle Latency
Victim Cache for Filter Cache Filter Cache 512B , Victim Cache 512B , Fully Associative, 1 Cycle Latency
Susan.smoothing sha SearchString Susan.smoothing sha SearchString Susan.smoothing sha SearchString Susan.smoothing sha SearchString
1 Core 2 Core 4 Core
Traditional Cache
Filter Cache
Double Filter Cache
Victim for Filter Cache
12 Run for each Benchmark !
10
Result
Susan.Smoothing
11
Result
sha
12
Result
SearchString
13
References

• 1 Y.J.Park, H.J.Choi, C.H. Kim, J.M. Kim,
  Energy-aware Filter Cache Architecture for
  Multicore Processors , 2010
• 2 J. Kin, M. Gupta, and W. Mangione-Smith, The
  Filter Cache An Energy Efficient Memory
  Structure, Proceedings of the International
  Symposium on Microarchitecture, pp. 184-193,
  1997.
• 3 N. P. Jouppi, Improving Direct-Mapped Cache
  Performance by the Addition of a Small
  Fully-Associative Cache and Prefetch Buffers,
  Proceedings of 17th Annual International
  Symposium on Computer Architecture, pp. 364-373,
  1990.
• 4 Multiprocessor Enhancements of the
  SimpleScalar Tool Set, Naraig Manjikian , 2001
• 5 ParMiBench - An Open-Source Benchmark for
  Embedded Multiprocessor Systems , Syed Muhammad
  Zeeshan Iqbal, Yuchen Liang, and Hakan Grahn ,
  2010
• 6 Multi2Sim A Simulation Framework to Evaluate
  Multicore-Multithreaded Processors, R. Ubal and
  J. Sahuquillo and S. Petit and P. L\'opez , 2007
• 7 CACTI 6.0 A Tool to Model Large Caches ,
  Muralimanohar, Naveen Balasubramonian, Rajeev
  Jouppi, Norman P. , 2009