You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: content/posts/optimuzz-meets-llm/index.md
+2-1Lines changed: 2 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -56,7 +56,8 @@ Which target line should Optimuzz aim at? Provide your best suggestion of the ta
56
56
The LLM answered with the target line 2026, where the optimized instruction is returned for the update optimization condition.
57
57
From this, we confirmed that the LLM can effectively identify the target line of the optimization update.
58
58
Note that the example above is a summary from the commit message and code changes from the LLVM commit: https://github.com/llvm/llvm-project/commit/1f69d6354a4b088bda5ee62c31411a75eead8b0c.
59
-
The script that construct a prompt can be found at https://github.com/prosyslab/optimuzz/blob/master/tools/targetline.py.
59
+
The script that construct a prompt can be found at https://github.com/prosyslab/optimuzz/blob/master/tools/targetline.py, and
60
+
the example prompt with its expected output can be found at https://github.com/prosyslab/optimuzz/blob/master/tools/targetline.expected .
60
61
61
62
As a result, we have successfully integrated LLMs into our workflow.
62
63
Every time there is an update on an optimization in LLVM, we automatically fetch the commit information using the GitHub API.
0 commit comments