Pipeline spectroscopy
Authors: 
Thomas R. Puzak, A. Hartstein, P. G. Emma, V. Srinivasan, Arthur NadasAuthors Info & Claims
Thomas R. Puzak, A. Hartstein, P. G. Emma, V. Srinivasan, Arthur NadasAuthors Info & ClaimsPages 15 - es
Abstract
Pipeline Spectroscopy is a new technique that allows us to measure the cost of each cache miss. The cost of a miss is displayed (graphed) as a histogram, which represents a precise readout showing a detailed visualization of the cost of each cache miss throughout all levels of the memory hierarchy. We call the graphs 'spectrograms' because they reveal certain signature characteristics of the processor's memory hierarchy, the pipeline, and the miss pattern itself. We show that in a memory hierarchy with N cache levels (L1, L2, ..., LN, and memory) and a miss cluster of size C, there are (C+N/C) possible miss penalties. This represent all possible sums from all possible combinations of the miss latencies from each level of the memory hierarchy (L2, L3, ... Memory) for a given cluster size. Additionally, a theory is presented that describes the shape of a spectrogram, and we use this theory to predict the shape of spectrograms for larger miss clusters. Detailed analysis of a spectrograph leads to much greater insight in pipeline dynamics, including effects due to prefetching, and miss queueing delays.
References
[1]
A. Glew, "MLP yes! ILP no!," in ASPLOS Wild and Crazy Idea Session, October 1998.
[2]
V. Pai and S. Adve, "Code Transformations to Improve Memory Parallelism," in 32nd International Symposium on Microarchitecture, November 1999.
[3]
H. Zhou and T. Conte, "Enhancing Memory Level Parallelism via Recovery-Free Value Prediction," in International Conference on Supercomputing, June 2003.
[4]
D. Sorin et al, "Analytic Evaluation of Shared-Memory Systems with ILP Processors," in 25th International Symposium on Computer Architecture, 1998.
[5]
V. Pai, P. Ranganathan and S. Adve, "The Impact of Instruction- Level Parallelism on Multiprocessor Performance and Simulation Methodology," in HPCA February 1997.
[6]
P. Ranganathan, K. Gharachorloo, S. Adve and L. Barroso, "Performance of Database Workloads on Shared-Memory Systems with Out-of-Order Processors," in ASPLOS-VIII, 1998.
[7]
Y Chou, B. Fahs, and S Abraham, "Microarchitecture Optimizations for Exploiting Memory-Level Parallelism Exploiting Memory-Level Parallelism" in 31st International Symposium on Computer Architecture, 2004.
[8]
Yuan Chou, Lawrence Spracklen, Santosh G. Abraham. "Store Memory-Level Parallelism Optimizations for Commercial Applications," pp. 183--196, 38th MICRO 2005.
[9]
M. Qureshi, D. Lynch, O. Mutlu, Yale Patt, "A Case for MLP-Aware Cache Replacement" in 33rd ISCA June 2006
[10]
A. Zahir, V. Hummel, M. Kling, T Yeh, US. Patent 6,353,802, "Apparatus and Method for Cycle Accounting in Microprocessors"
[11]
B. Gaither, R. Smith, US Patent 6,892,173 B1, "Analyzing Effectiveness of a Computer Cache By Estimating a Hit Rate Based on Applying a Subset of Real-time Addresses to a Model of the Cache"
[12]
H. Ravichandran, US Patent 6,341,357 B1, "Apparatus and Method for Processor Performance Monitoring",
[13]
R. Trauben, US Patent 5,594,864, "Method and apparatus for unobtrusively monitoring Processor States and Characterizing Bottlenecks in a Pipeline Processor Executing Grouped Instructions"
[14]
G. Brooks, US Patent 5,845,310 "System and Methods For Performing Cache Latency Diagnostics in Scalable Parallel Processing Architectures Including Calculating CPU Idle Time and Counting Number of Cache Misses.
[15]
W. Flynn, US Patent 6,256,775 B1, "Facilities For Detailed Software Performance Analysis in a Multithreaded Processor"
[16]
F. Levine, B. McCredie, W. Starke, E. Welbon, US Patent 5,862,371, "Method and System for Instruction Trace Reconstruction Utilizing Performance monitor outputs and bus Monitoring"
[17]
J. Dean, J. E. Hicks, C. A. Waldspurger, W. E. Weihl, and G. Z. Chrysos. ProfileMe: Hardware support for instruction-level profiling on out-of-order processors. In MICRO'97: pages 292--302, 1997.
[18]
Brian A. Fields, Rastislav Bodik, Mark D. Hill, Chris J. Newburn., Interaction cost and shotgun profiling. ACM Transactions on Architecture and Code Optimization, Vol 1, No. 3. Sept 2004.
[19]
Tejas Karkhanis, James E. Smith, A First-Order Superscalar Processor Model. Proceedings of the 31st ISCA. pages 338--349, June 2004.
[20]
A. Hartstein and T. Puzak. The optimum pipeline depth for a microprocessor, 29th ISCA, pages 7--13 May 2002.
[21]
A. Hartstein and T. Puzak. Optimum power/performance pipeline depth. 36th Annual IEEE/ACM International Symposium on Microarchitecture In MICRO, Dec. 2003.
[22]
P. J. Denning, "The Working Set Model for Program Behavior", CACM 19(5) pp.285--294 (1976).
[23]
R. Bartoszynski, M Niewiadomska-Bugaj, Probability and Statistical Inference, (Wiley series in probability and statistics) 1996
[24]
J. S. Liptay, "Design of the IBM Enterprise System/9000 High-End Processor," IBM Journal of Research and Development, Vol. 36, No. 4, pp. 713--731, July, 1992.
[25]
D. B. Fite, J. E. Murray, D. P. Manley, M. M. McKeon, E. H. Fite, R. M. Salett, and T. Fossum, "Branch Prediction," U.S. Patent #5142634, assigned to Digital Equipment Corporation, Filed Feb. 3, 1989, Issued Aug. 25, 1992.
[26]
N. Suzuki, "Microprocessor Having Branch Prediction Function," U.S. Patent #5327536, assigned to NEC Corporation, Filed May 22, 1991, Issued July 5, 1994.
[27]
C. H. Perleberg, and A. J. Smith, "Branch Target Buffer Design and Optimization," IEEE Transactions on Computers, Vol. 42, Issue 4, pp. 396--412, Apr., 1993.
[28]
B. Calder, and D. Grunwald, "Fast and Accurate Instruction Fetch and Branch Prediction," 21st ISCA pp. 2--11, April 18--21, 1994.
Index Terms
- Pipeline spectroscopy
Recommendations
Pipeline spectroscopy
ecs'07: Experimental computer science on Experimental computer sciencePipeline Spectroscopy is a new technique that allows us to measure the cost of each cache miss. The cost of a miss is displayed (graphed) as a histogram, which represents a precise readout showing a detailed visualization of the cost of each cache miss ...
Increasing hardware data prefetching performance using the second-level cache
Techniques to reduce or tolerate large memory latencies are critical for achieving high processor performance. Hardware data prefetching is one of the most heavily studied solutions, but it is essentially applied to first-level caches where it can ...
NCID: a non-inclusive cache, inclusive directory architecture for flexible and efficient cache hierarchies
CF '10: Proceedings of the 7th ACM international conference on Computing frontiersChip-multiprocessor (CMP) architectures employ multi-level cache hierarchies with private L2 caches per core and a shared L3 cache like Intel's Nehalem processor and AMD's Barcelona processor. When designing a multi-level cache hierarchy, one of the key ...
Comments
Information & Contributors
Information
Published In
June 2007
218 pages
ISBN:9781595937513
DOI:10.1145/1281700
Copyright © 2007 ACM.
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]
Sponsors
In-Cooperation
Publisher
Association for Computing Machinery
New York, NY, United States
Publication History
Published: 13 June 2007
Check for updates
Author Tags
Qualifiers
- Article
Conference
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 157Total Downloads
- Downloads (Last 12 months)1
- Downloads (Last 6 weeks)1
Reflects downloads up to 23 Oct 2024
Other Metrics
Citations
View Options
Get Access
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in