michaeljmarshall commented on code in PR #16145:
URL: https://github.com/apache/lucene/pull/16145#discussion_r3320739076


##########
lucene/CHANGES.txt:
##########
@@ -141,6 +141,12 @@ Improvements
 
 Optimizations
 ---------------------
+* GITHUB#16145: Always fire prefetch when ReadAdvice is RANDOM in 
MemorySegmentIndexInput. The power-of-two throttle in
+  MemorySegmentIndexInput assumes "consecutive cache hits → file is warm", 
which does not hold for
+  random-access patterns where each request visits unpredictable pages. A 
negative sentinel in the
+  shared prefetch counter signals RANDOM mode with zero overhead on the 
non-RANDOM hot path.
+  (Michael Marshall)

Review Comment:
   Sorry for that mistake. I had initially prototyped this with a sentinel 
implementation, but forgot to update this text. I'll get this fixed.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to