mikemccand commented on issue #15662:
URL: https://github.com/apache/lucene/issues/15662#issuecomment-3860089182

   > (its lifetime was only a few hours)
   
   Yikes -- poor baby mainboard.
   
   > > > recent biological failure
   > > 
   > > 
   > > Whoa, WHAT? Yikes -- looks painful 
[@uschindler](https://github.com/uschindler)!! Did you ... somehow break that 
little bone fragment off? Playing tennis!!?? Typing too much to genai? Maybe 
doing your [long-running plank 
exercises/positions](https://en.wikipedia.org/wiki/Plank_(exercise)) while 
waiting for Lucene tests to finish?
   > 
   > No pain at all. Its a followup from a 2012-repaired bankart repair, this 
time bone is the issue, the previous fix was stable for 13 years. I was tryingt 
too much things in parallel on a stepmaster machine and fall down while trying 
to hold myself with arms causing the shoulder to partially dislocate. As bone 
broke at that time the shoulder dislocated half an hour later by something 
similar to Planks :-) The only painful thing was the dislocated shoulder, 
everything else is feeling "normal". Pain is gone once it is back at its 
correct place.
   
   @uschindler I have been trying, ever since reading this comment yesterday, 
then on my commute home through crazy cold / snowy / icy / muddy Boston 
streets, then dinner, bedtime, Arduino hacking, dreams, to specifically picture 
the sequence of events you describe above!  I tried [Veo 
here](https://gemini.google.com/share/f4d913ab587e) but that's nowhere close to 
what I'm picturing (it is missing the "something similar to Planks" phase of 
your shoulder accident, even when I prompted it for that second phase.  Also, 
it's weird how the phone simply evaporates/teleports instead of falling on the 
floor and exploding like it normally would.)  I hope you get/got the Bony 
Bankart Lesion repaired!
   
   Anyway, bad news... I ran the bench on GitHub commit 
https://github.com/apache/lucene/commit/0a699fa4d2db9741c699a01c1b731bbd66613b6a
 (just before shared `AtomicInteger` counter for prefetch optimization) and 
it's STILL SLOW.
   
   Summary of current situation/timeline:
     * 01/22 normal nightly bench was last fast one
     * ... benchy broken for ~ 1 week ...
     * ... Mike does whole system upgrade; many packages upgraded including 
kernel ...
     * @shubhamsrkdev [fixes nightly benchy compilation 
errors](https://github.com/mikemccand/luceneutil/commit/3d5e3f8b545ff29c230835d16dc0941afc7306da)
 (thank you!)
     * 01/29 normal nightly bench was first slow one (~10% - 48% slowdown)
     * downgrade kernel to latest LTS (`6.12.67-1-lts`) and run one-off 
nightly: still slow
     * put kernel back to Arch bleeding edge (`6.18.7-arch1-1`) and roll Lucene 
changes back to 
https://github.com/apache/lucene/commit/0a699fa4d2db9741c699a01c1b731bbd66613b6a
 and run one-off nightly benchy: still slow
   
   I will next test Lucene rev 
https://github.com/apache/lucene/commit/d455fefc6b63f90216c938fda061286918fee37e


-- 
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