71aa8aaeea
Paril/mem write ( #412 )
...
* simplify light_surfaces storage
* memory on writes
* Revert "batch together traces on the whole surface - saves a small amount of time"
This reverts commit 72611b7e97
2024-03-13 20:01:05 -06:00
b89bac4e8c
common: split start of bspfile.hh into bspfile_common.hh
...
old layout was confusing CLion because bspfile.hh included
itself
2024-03-11 20:14:59 -06:00
275695a37c
qbsp: remove -transsky flag which was broken
...
unknown what it was useful for
2024-03-09 22:07:55 -07:00
aa3b8479b4
qbsp: shrink node_t from 280 to 208 bytes
...
- contentflags_t from 80 to 8 bytes
2024-03-09 19:37:23 -07:00
3d419853d3
store leaves on surf instead of surface light
2024-03-09 00:14:03 -05:00
6c1feb5af8
cut per-surface memory usage down by making streams re-usable per thread
2024-03-08 23:09:16 -05:00
6157945d4d
optimize Q2 brush info by using world-aligned defaults if they exist
...
optimize Q2 animations by using world-aligned defaults, since Q2 only needs the secondary texinfo for the texture
2024-03-05 11:11:20 -05:00
ee075b5781
qbsp: shrink node_t from 328 to 280 bytes
2024-02-27 23:34:47 -07:00
f3da00395d
style: replace (void) with ()
2024-02-26 21:07:22 -07:00
12b8066246
light: change remaining double -> float
...
tiny perf improvement on base1-test.map with light params:
-emissivequality high
before: 7.4s
after: 6.9s
no visible changes in base1.
2024-02-25 23:42:45 -07:00
1e53d45a6c
qbsp: fix comment
2024-02-25 15:54:58 -07:00
765d95e189
qbsp: q1: make liquids detail when -transwater is in use (i.e. by default)
...
Fixes #353
Fixes #244
2024-02-10 21:08:29 -07:00
aa970c0d18
qbsp: only pick up .wad textures when compiling for non-Q2 games
...
fixes #404
2024-02-04 19:41:45 -07:00
d512abd27d
qbsp: merge across liquids by default in q1
...
- change option name to -nomergeacrossliquids
- add docs
- allow using it in q2 as well (requested by jitspoe)
2024-01-25 23:04:31 -07:00
0aeb0acd94
qbsp: adjust tjunc logic
2024-01-24 00:22:27 -07:00
b66c95a53f
tests: add q2_mist_transwater
2024-01-23 20:36:34 -07:00
53850d94ac
fix crash in embree tracing due to new batching code
...
simplify structs for tracing
2024-01-21 08:52:16 -05:00
07e050f484
light: restore -bounce flag with no arguments as a shorthand for -bounce 1
2024-01-13 18:35:18 -07:00
1c692d6f47
qbsp: record portal type
2024-01-05 12:32:01 -07:00
8ab9960ca0
lightpreview: don't crash when vis is run more than once
2023-12-24 17:22:01 -07:00
8db1d32e9f
light: fix excessive bouncing with lit water, or _shadow -1
...
Fixes #397
2023-12-24 00:44:56 -07:00
6351cc50e6
rename: vec_t -> double
2023-12-23 21:09:16 -07:00
f0c8d92993
qbsp: add -missing_textures_as_zero_size flag to allow writing 0x0 textures when a texture is missing
...
this allows us to write the texture name, but it's non-standard.
light: warn when a surface light template doesn't match any faces in the bsp
2023-12-21 00:05:40 -07:00
1eaf2170f8
Merge branch 'brushbsp' of https://github.com/ericwa/ericw-tools into brushbsp
...
# Conflicts:
# light/trace_embree.cc
2023-12-07 14:13:36 -05:00
9ed4f3d042
bunch of double -> float
...
fix Q2 animated textures
2023-12-07 14:11:46 -05:00
b45871efdb
vis algorithm improvement (2h42 -> 57m on ad_dm5) ( #361 )
...
* vis algorithm improvement, new commandline option -targetchecks
* vis: only calculate num_expected_targetchecks if needed
Fixes a bit of performance regression when -targetchecks was added, when running
with -targetchecks 0, apparently due to the cost of all the std::popcount's.
gmsp3v2.bsp, 32 threads:
18s (brushbsp branch 9388d43b1c
2023-12-02 13:41:56 -07:00
b58b976693
tests: add qbsp content type tests
2023-11-26 00:23:41 -07:00
ad9fd7fc59
common: structured bspxbrushes reading/writing
2023-11-25 17:28:26 -07:00
ec452e2291
remove implicit conversion for qplane3
2023-11-19 19:41:58 -05:00
db0951dc40
light: support embree4 in addition to 3
2023-11-19 14:14:22 -07:00
d5ffbd9d33
qbsp: remove unused type
2023-11-18 23:24:03 -07:00
c77f4ba2dd
qbsp: fix misc_external_map + -wrbrushes
...
- fix world extents not including external maps
- thanks @DaZombieKiller for pointing this out
Fixes #372
2023-11-18 15:14:38 -07:00
22ad2f4141
light: cache list of lightsurfs that are actuall emissive
...
light.exe -visapprox none ad_sepulcher.bsp
45.5s -> 22.7s
2023-11-13 23:35:32 -07:00
5bcb2d4657
vis: remove MAX_PORTALS_ON_LEAF
...
reduces memory use a bit
gmsp3v2.bsp - 51.96s elapsed
2023-11-11 13:10:01 -07:00
37fc62ca0e
vis: remove dead code from header
2023-11-11 12:29:30 -07:00
8888506821
qbsp: fix -notex option
2023-11-09 23:57:07 -07:00
8845866c00
qvec: use fold expression for qv::dot
...
seems to help codegen on msvc.
gmsp3v2.bsp vis times: 51.9s , 52.1s
prev commit was 56s
2023-11-09 00:25:31 -07:00
0ea6b0989c
clean up vec stuff a bit
...
constexpr assignment operators
2023-11-09 01:38:17 -05:00
f1bf0f546d
add missing #pragma once's
2023-11-08 21:30:44 -07:00
c9edf4ddaa
vis: huge speedup from fixing false sharing with stat counters
...
gmsp3v2.bsp, 32 threads 3950x, 157s -> 61s
2023-11-08 00:05:08 -07:00
7e5ad9a39b
lightpreview: add support for Q1 visdata
2023-11-06 22:54:21 -07:00
e9532346c7
vis: partly fix perf regression of Oct 2021
...
- drops usage of winding_storage_hybrid
- makes pstack_t trivially default initializable again (like the C
original)
- we weren't making use of the unlimited size of winding_storage_hybrid;
the original vis already has an allowance for if a winding exceeds
the fixed stack allocation size (24) and we were still applying that
limit
before:
cmake-build-release\tests\tests.exe -r=xml -sf=*benchmark.cc
Testing started at 1:53 PM ...
| ns/op | op/s | err% | total | benchmark
|--------------------:|--------------------:|--------:|----------:|:----------
| 45.15 | 22,149,002.06 | 0.1% | 0.01 | `create pstack_t`
| 47.04 | 21,256,429.80 | 0.1% | 0.01 | `create pstack_t + 1x AllocStackWinding`
| 48.28 | 20,710,594.32 | 0.2% | 0.01 | `create pstack_t + 2x AllocStackWinding`
| 128.52 | 7,780,603.30 | 0.3% | 0.01 | `setup + ClipStackWinding`
Benchmark from: https://www.quaddicted.com/tools/vis_benchmark
cmake-build-release\vis\vis.exe -nostate -nopercent gmsp3v2.bsp
153.31s elapsed
after:
cmake-build-release\tests\tests.exe -r=xml -sf=*benchmark.cc
Testing started at 1:46 PM ...
| ns/op | op/s | err% | total | benchmark
|--------------------:|--------------------:|--------:|----------:|:----------
| 1.64 | 610,528,630.06 | 0.3% | 0.01 | `create pstack_t`
| 3.29 | 303,895,535.45 | 0.6% | 0.01 | `create pstack_t + 1x AllocStackWinding`
| 5.20 | 192,172,148.84 | 0.2% | 0.01 | `create pstack_t + 2x AllocStackWinding`
| 66.84 | 14,960,818.66 | 1.0% | 0.01 | `setup + ClipStackWinding`
Benchmark from: https://www.quaddicted.com/tools/vis_benchmark
cmake-build-release\vis\vis.exe -nostate -nopercent gmsp3v2.bsp
141.42s elapsed
2023-11-05 13:58:36 -07:00
b83c4a6440
use fold expr instead of for statements in vec3
...
fix a bunch of clang warnings
2023-11-04 19:01:08 -04:00
d6fbefaf0e
lightpreview: decompile hulls for leaf visualization
2023-10-19 00:10:42 -06:00
91734ac8f6
light: fix gcc9 build
2023-09-17 21:24:04 -06:00
e6d9605b86
light: fix lightgrid assertion failure if lightgrid gets a nan (separate bug)
2023-09-17 20:27:29 -06:00
d189145a01
build: fixes for mingw support ( #373 )
2023-09-16 13:15:31 -06:00
0509c9a790
fix multiple bounce interaction with lightgrid
2023-09-16 11:55:10 -04:00
147be8f87a
some optimizations in indirect lighting code
...
multiple bounce support
2023-09-16 00:41:05 -04:00
6baa1243cf
allow auto-gamedir to handle subdirectoried maps properly
...
fix include paths
2023-09-08 08:41:56 -04:00
Jonathan
Eric Wasylishen
Erika
qruf