Proceedings of the 48th Design Automation Conference 2011
DOI: 10.1145/2024724.2024916
|View full text |Cite
|
Sign up to set email alerts
|

Threadmill

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1

Citation Types

0
1
0

Year Published

2012
2012
2018
2018

Publication Types

Select...
5
1

Relationship

0
6

Authors

Journals

citations
Cited by 31 publications
(1 citation statement)
references
References 12 publications
0
1
0
Order By: Relevance
“…The major downside, however, of random test programs is the difficulty to obtain the expected correct results (the correct output of these programs is unknown because they are randomly generated), which are required to determine the correctness of the output of the prototype chips being validated. To meet this requirement, validation process mainly resorts either to multi-pass consistency end-of-test checking methods (each testcase is executed multiple times ('passes') and the end-of-test values of some system resources (e.g., memory, registers) are compared for consistency) [98], or to golden responses generated by architectural simulators. However, the throughput difference between native chip execution and simulation (at different levels) is between 3 and 6 orders of magnitude.…”
Section: Reducing the Bug Detection Latency For Atm Post-silicon Validationmentioning
confidence: 99%
“…The major downside, however, of random test programs is the difficulty to obtain the expected correct results (the correct output of these programs is unknown because they are randomly generated), which are required to determine the correctness of the output of the prototype chips being validated. To meet this requirement, validation process mainly resorts either to multi-pass consistency end-of-test checking methods (each testcase is executed multiple times ('passes') and the end-of-test values of some system resources (e.g., memory, registers) are compared for consistency) [98], or to golden responses generated by architectural simulators. However, the throughput difference between native chip execution and simulation (at different levels) is between 3 and 6 orders of magnitude.…”
Section: Reducing the Bug Detection Latency For Atm Post-silicon Validationmentioning
confidence: 99%