profan
|
b8c9d55258
|
add comment regarding chunking of generated code
|
2026-06-07 00:01:31 +01:00 |
profan
|
0ab99cdd10
|
hmm yes
|
2026-06-06 23:55:38 +01:00 |
profan
|
b654e46f69
|
wooe
|
2026-06-06 23:37:43 +01:00 |
profan
|
c8fb614c51
|
remove duplicated code
|
2026-06-06 23:29:00 +01:00 |
profan
|
ec098a72da
|
ensure eval and compilation do not conflict/overlap, add pending indicator
|
2026-06-06 22:59:13 +01:00 |
profan
|
6ac115fe0b
|
some cleanup
|
2026-06-06 22:02:02 +01:00 |
profan
|
54bb37852a
|
cache the compiled inner loop whenever the options change, so it does not have to recompile every time when evaluated
|
2026-06-06 21:21:15 +01:00 |
profan
|
006ad8c4f7
|
hmm yes
|
2026-06-06 20:55:45 +01:00 |
profan
|
6822d8f896
|
implement optional compilation of inner loop to CIL, unrolling the loop into a number of functions we invoke
|
2026-06-06 20:49:12 +01:00 |
profan
|
f7c3bde22f
|
ensure float path is used, no implicit double conversion
|
2026-06-05 21:18:47 +01:00 |
profan
|
31d83ec058
|
shave another 50-100ms off zeroing memory
|
2026-06-05 21:11:17 +01:00 |
profan
|
076bea641d
|
make visualization of operands nicer, so you can distinguish constants and addresses more easily
|
2026-06-05 20:40:40 +01:00 |
profan
|
abfc5433c5
|
things
|
2026-06-05 20:28:50 +01:00 |
profan
|
e47723b3c4
|
show stats about how long eval takes, and how long it takes per pixel
|
2026-06-05 20:12:24 +01:00 |
profan
|
da83b1857e
|
just a naming thing
|
2026-06-05 19:59:21 +01:00 |
profan
|
0129200964
|
support running both with/without simd optionally with a (really horrible) shared codepath
|
2026-06-05 19:57:40 +01:00 |
profan
|
838887f06c
|
yonks
|
2026-06-05 00:16:20 +01:00 |
profan
|
afb608ef7a
|
ensure evaluation cannot be racey
|
2026-06-05 00:14:35 +01:00 |
profan
|
0eda244ce2
|
add the ability to toggle parallelism
|
2026-06-05 00:09:12 +01:00 |
profan
|
c6f5411f91
|
add raylib based interactive visualization of the rendering
|
2026-06-04 23:49:15 +01:00 |
profan
|
3625fce0c9
|
implement basic constant folding, eliminating around 17 % of the instructions from the hot loop and folding it into the instructions
|
2026-05-31 17:19:15 +01:00 |
profan
|
5f00120b97
|
hmm
|
2026-05-25 17:51:38 +01:00 |
profan
|
b434bac921
|
ceremony
|
2026-05-25 17:35:18 +01:00 |
profan
|
640ae5aea5
|
format output slightly more nicely
|
2026-05-25 12:17:25 +01:00 |
profan
|
88cea8be53
|
claw back some time with skip locals init
|
2026-05-25 11:58:51 +01:00 |
profan
|
b576f44fea
|
massively speed up image construction
|
2026-05-25 11:47:36 +01:00 |
profan
|
84a1ad7330
|
iNITIal commit, simd accelerated interpreter
|
2026-05-25 11:38:28 +01:00 |