Per-instruction cycle stacks through time-proportional event analysis
- Author
- Bjorn Gottschall, Lieven Eeckhout (UGent) and Magnus Jahre
- Organization
- Project
-
- Computer Architecture Scale Models
- Photonic Network-on-Wafer for Multi-Tile GPUs: From Architecture to Hardware Implementation
- Load Slice Core (Load Slice Core: A Power and Cost-Efficient Microarchitecture for the Future)
- Abstract
- Understanding what applications spend time on and why is critical for effective performance optimization. Unfortunately, current state-of-the-art performance analysis tools are generally unable to provide this information. The fundamental reason is that they lack time proportionality; i.e., in many cases, they do not attribute execution time to the instructions and performance events that the architecture is exposing the latency of. Time-proportional event analysis (TEA) creates per-instruction cycle stacks, which clearly and accurately explain what the application spends time on and why at the level of individual static instructions. TEA requires executing the application only once; it is accurate (with an average error of 2.1%); and its hardware implementation incurs negligible runtime, power, and area overheads of 1.1%, 0.1%, and 249 bits per core, respectively.
- Keywords
- Program processors, Runtime, Systematics, Codes, Out of order, Optimization, Performance evaluation, Optimization methods, Event detection
Downloads
-
tea top picks 24.pdf
- full text (Accepted manuscript)
- |
- open access
- |
- |
- 235.00 KB
-
(...).pdf
- full text (Published version)
- |
- UGent only
- |
- |
- 531.43 KB
Citation
Please use this url to cite or link to this publication: http://hdl.handle.net/1854/LU-01JX2DW8DEF4CT3ZSP6D45TQ7Y
- MLA
- Gottschall, Bjorn, et al. “Per-Instruction Cycle Stacks through Time-Proportional Event Analysis.” IEEE MICRO, vol. 44, no. 4, 2024, pp. 27–33, doi:10.1109/MM.2024.3407377.
- APA
- Gottschall, B., Eeckhout, L., & Jahre, M. (2024). Per-instruction cycle stacks through time-proportional event analysis. IEEE MICRO, 44(4), 27–33. https://doi.org/10.1109/MM.2024.3407377
- Chicago author-date
- Gottschall, Bjorn, Lieven Eeckhout, and Magnus Jahre. 2024. “Per-Instruction Cycle Stacks through Time-Proportional Event Analysis.” IEEE MICRO 44 (4): 27–33. https://doi.org/10.1109/MM.2024.3407377.
- Chicago author-date (all authors)
- Gottschall, Bjorn, Lieven Eeckhout, and Magnus Jahre. 2024. “Per-Instruction Cycle Stacks through Time-Proportional Event Analysis.” IEEE MICRO 44 (4): 27–33. doi:10.1109/MM.2024.3407377.
- Vancouver
- 1.Gottschall B, Eeckhout L, Jahre M. Per-instruction cycle stacks through time-proportional event analysis. IEEE MICRO. 2024;44(4):27–33.
- IEEE
- [1]B. Gottschall, L. Eeckhout, and M. Jahre, “Per-instruction cycle stacks through time-proportional event analysis,” IEEE MICRO, vol. 44, no. 4, pp. 27–33, 2024.
@article{01JX2DW8DEF4CT3ZSP6D45TQ7Y,
abstract = {{Understanding what applications spend time on and why is critical for effective performance optimization. Unfortunately, current state-of-the-art performance analysis tools are generally unable to provide this information. The fundamental reason is that they lack time proportionality; i.e., in many cases, they do not attribute execution time to the instructions and performance events that the architecture is exposing the latency of. Time-proportional event analysis (TEA) creates per-instruction cycle stacks, which clearly and accurately explain what the application spends time on and why at the level of individual static instructions. TEA requires executing the application only once; it is accurate (with an average error of 2.1%); and its hardware implementation incurs negligible runtime, power, and area overheads of 1.1%, 0.1%, and 249 bits per core, respectively.}},
author = {{Gottschall, Bjorn and Eeckhout, Lieven and Jahre, Magnus}},
issn = {{0272-1732}},
journal = {{IEEE MICRO}},
keywords = {{Program processors,Runtime,Systematics,Codes,Out of order,Optimization,Performance evaluation,Optimization methods,Event detection}},
language = {{eng}},
number = {{4}},
pages = {{27--33}},
title = {{Per-instruction cycle stacks through time-proportional event analysis}},
url = {{http://doi.org/10.1109/MM.2024.3407377}},
volume = {{44}},
year = {{2024}},
}
- Altmetric
- View in Altmetric
- Web of Science
- Times cited: