Repository navigation
Converter: preset L in the preset table, with a synthetic two-preset test - #94
Merged
Merged
Conversation
…reset test Preset L uses preset M's map versions and record kinds; a map names it with "preset": "L" and gets its own no-model message (exit 12; the DLSS version named is provisional). run_tests.py adds a synthetic library file with an L block directly before an M block and maps for both: each converts its own block only, and a file without a preset's block fails with that preset's message. The real preset-L map file is not part of this change. Existing synthetic outputs are byte for byte those of main. Refs #89 Signed-off-by: phantoslayer21 <156617979+phantoslayer21@users.noreply.github.com>
phantoslayer21
marked this pull request as ready for review
November 13, 2026 22:05
This was referenced Nov 13, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Stacked on #90; merge after #90. #90 (the map's preset by name) is merged, and this pull request now targets main.
Its diff is only its own commit.
Preset L (Model L, #89) is a further preset of the same family as M. Its maps use preset M's record kinds and map
versions, and its parameters are a block of their own in the same library file as M's. With #90 that makes it one row
in the converter's preset table.
clean/converter/src/map.cppknown_presets: rowL, map versions 1 and 4. M comes first, so a map without the"preset"field is never read as L.clean/converter/src/api.cpp/.h:kMsgNoPresetL, the exit-12 message "This DLSS file has no preset-L model.Version 310.9 or newer is needed." The version line is provisional, pending confirmation of the first DLSS
version that ships a preset-L model; it is one line to change then.
tests/cases.pypreset_pair_case: a synthetic library file (our own sizes and bytes) with an L-like parameter blockdirectly before an M-like one, each with its own dither table and keyed constants elsewhere. The L map
(
"preset": "L") has one stage more, standing in for L's extra blocks. The M map has no"preset"field, as theshipped one.
tests/run_tests.pytest_preset_l:its key source is found at the offset where it was placed;
"preset"field reads as M (M's message) and still converts the same tensors;"preset"valuesK,F,Xandlin a version-1 map are converter data damaged.tests/unit_tests.cpp: the L row (versions, message),"preset":"L"in the version-1 and version-4 unit maps.clean/converter/README.md: the L row in the exit-code table and a short "Preset L" paragraph.Not in this pull request: the preset-L map file itself (
data/preset-L.v1.map.json). It is converter data from thespec side and comes separately. Until then, preset L is covered by the synthetic case above. Wiring L into the
runtime, launcher and model folder is separate Model L work.
Refs #89
Tested
Run locally on Windows, CPU only, Ninja Release
/MTwith MSVC. "Main" is d9c697e.ctest(unit, decompression-vectors, synthetic, synthetic5): 4/4 passed.run_tests.py: 3757 checks passed, 0 failed (main: 3663; the 94 more aretest_preset_l).syntheticandsynthetic5tests, every output fileoutside the new
preset-lfolder was hashed: 1615 files (63model.wgcmplus raw tensors), SHA-256 identical tomain's.
Not tested