Tony Garnock-Jones
|
94646a92de
|
Remove a few small Runnable allocations
|
2023-10-28 18:22:53 +02:00 |
Tony Garnock-Jones
|
d1acf60d1b
|
Avoid double-scheduling for a nice win
|
2023-10-28 17:12:07 +02:00 |
Tony Garnock-Jones
|
6904ef76df
|
Rearrange to put the common case at the top
|
2023-10-28 16:27:51 +02:00 |
Tony Garnock-Jones
|
636da3f28f
|
Is there *really* a need to allocate fresh Turns every time? I think not
|
2023-10-28 16:18:42 +02:00 |
Tony Garnock-Jones
|
732bb0066d
|
Oops, forgot to remove the allocation
|
2023-10-28 16:16:16 +02:00 |
Tony Garnock-Jones
|
7d9174c363
|
More data-structure inlining
|
2023-10-28 16:09:06 +02:00 |
Tony Garnock-Jones
|
fd8b445482
|
Try to eke out a little more speed
|
2023-10-28 16:03:46 +02:00 |
Tony Garnock-Jones
|
5401d2d8e9
|
Reduce allocation slightly further
|
2023-10-28 15:45:27 +02:00 |
Tony Garnock-Jones
|
f55fd2c668
|
Good performance boost by avoiding hashmaps in the single-target case
|
2023-10-28 15:37:10 +02:00 |
Tony Garnock-Jones
|
637f1da5f4
|
Implement ring benchmark; repair some errors
|
2023-10-28 13:23:29 +02:00 |
Tony Garnock-Jones
|
c7718b3ddd
|
Make turns atomic, rather than reply-turn-per-event
|
2021-05-31 12:05:00 +02:00 |
Tony Garnock-Jones
|
d6c8a80995
|
Bump language level to JDK 16 to get records; update example1
|
2021-05-12 15:47:59 +02:00 |
Tony Garnock-Jones
|
af611e1b2b
|
First broad step toward novy
|
2021-05-12 15:47:59 +02:00 |