Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Constant L1i-Cache Usage of Payloads #84

Open
2 tasks
marenz2569 opened this issue Oct 30, 2024 · 0 comments
Open
2 tasks

Constant L1i-Cache Usage of Payloads #84

marenz2569 opened this issue Oct 30, 2024 · 0 comments
Labels
backlog optimization improvements Backlog of things that improve the optimization e.g., runtime or effectiveness of the results

Comments

@marenz2569
Copy link
Member

Currently we create the payloads by setting a number of lines and the proportions of the instructions. Given a map of instruction name to proportion we create a sequence containing the number proportion times the particular instruction. The total length of the sequence is the sum of all proportions. This sequence will be repeated floor(line count / sequence size), resulting in a non constant count of instructions and therefore different sizes of L1i-Cache usage.

Please note the different power consumption based on changing sequence size in FIRESTARTER 2: Dynamic Code Generation for Processor Stress Tests Fig. 8.

DoD

  • Evaluate how the line count should be set i.e., based on constant L1i-Cache usage or a specific number of instructions.
  • Change the optimization to use either a constant L1i-Cache usage or a specific number of instructions if this is sufficient.
@marenz2569 marenz2569 added the backlog optimization improvements Backlog of things that improve the optimization e.g., runtime or effectiveness of the results label Oct 30, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
backlog optimization improvements Backlog of things that improve the optimization e.g., runtime or effectiveness of the results
Projects
None yet
Development

No branches or pull requests

1 participant