magnum

Commit Graph

Author	SHA1	Message	Date
Vladimír Vondruš	2479b34165	AnySceneConverter: forgot to say this is since 2020.06, huh.	5 years ago
Vladimír Vondruš	2891ea3a0f	doc: mention also how Magnum handles errors. Or HOW IT EXPLODES STRAIGHT IN YOUR HANDS	5 years ago
Vladimír Vondruš	a2b8fb73c2	doc: freshen up plugin-related documentation.	5 years ago
Vladimír Vondruš	780ba2fd99	doc: prettier responsive button layout.	5 years ago
Vladimír Vondruš	2347a7f730	doc: updated changelog.	5 years ago
Vladimír Vondruš	83f18b2313	doc: finally document the library organization and essential tools.	5 years ago
Vladimír Vondruš	c13c61e17d	Trade: hint on related APIs in mutable MeshData access.	5 years ago
Vladimír Vondruš	2bd933d3ef	doc: add docs for GL context asserts and other troubleshooting.	5 years ago
Vladimír Vondruš	27b0527fc0	Trade: use consistent naming in an assert message.	5 years ago
Vladimír Vondruš	49a5756f70	Trade: minor.	5 years ago
Vladimír Vondruš	1c3f4844e9	Trade: doc++	5 years ago
Squareys	e72a440720	Math: typo.	5 years ago
Vladimír Vondruš	c59c48c75f	sceneconverter: list skins in --info.	5 years ago
Vladimír Vondruš	120286a930	Add Degh, Radh and Range*Dh typedefs. We have half-float vectors and matrices, so why not these as well. Not sure for what all is the angle precision usable, but at the very least it could be useful for compact meshlet occlusion cone / AABB representation or rough animations.	5 years ago
Vladimír Vondruš	047c539a17	ÄnyImageImporter: recognize BMP header magic. I needed to know the signature for a StbImageConverter test, so why not add it here as well now that I managed to figure that out.	5 years ago
Vladimír Vondruš	c1c02dea58	Trade: make Abstract{Image,Scene}Converter::doConvertToFile() protected. This is already done for the AbstractImporter and the new AbstractShaderConverter, as there's a common use case of checking just the filename for input/output path or file type detection and then delegating to the common implementation working directly on data.	5 years ago
Vladimír Vondruš	48838c442f	Adapt to Corrade changes.	5 years ago
Vladimír Vondruš	416704fa47	Any{Image,Scene,Shader}Converter: modernize the string code a bit. Only those use StringViews in the plugin API so far, importers not yet.	5 years ago
Vladimír Vondruš	6087107e0d	Trade: don't use deprecated APIs in deprecated APIs. Because MSVC IS STUPID and emits a ton of warnings here. All sane compilers are sane and silent. It doesn't fucking make sense to warn here, COME ON.	5 years ago
Vladimír Vondruš	99395eb69a	Doc++	5 years ago
Vladimír Vondruš	e6d673c279	Propagate configuration options in Any* plugins. Minor but very important convenience feature, especially useful when dealing with command-line apps. This now works: magnum-imageconverter a.png a.jpg -c jpegQuality=0.75 The AnyImageConverter gets the jpegQuality option and then automatically propagates it to the concrete plugin (which is either JpegImageConverter or StbImageConverter), possibly warning in case the target plugin doesn't recognize given option (i.e., doesn't list it in its default configuration). Previously the user had to always specify a concrete converter implementation using -C, which was rather annoying and nonintuitive.	5 years ago
Vladimír Vondruš	4d61cda4a3	AnyImageImporter: don't expect metadata to be present too early. The check that the plugin is loadable should happen before this, not after. All other importer plugins do that already.	5 years ago
Vladimír Vondruš	23712c414e	AnyShaderConverter: assert the metadata validity always. Practically the metadata shouldn't be null if we managed to load the plugin already, so this was probably harmless. But dumb.	5 years ago
Vladimír Vondruš	3452fd7a24	AnyShaderConverter: adapt tests to GlslangShaderCovnerter updates.	5 years ago
Vladimír Vondruš	50e75f90c8	AnyImageConverter: rename a test macro for clarity.	5 years ago
Vladimír Vondruš	15ddae576b	AnyImageConverter: add placeholder tests for compressed 2D support.	5 years ago
Vladimír Vondruš	d6a3be571b	AnySceneConverter,AnySceneImporter: expand tests with external plugins. The AnyShaderConverter tests are doing this already anyway, so why not here as well.	5 years ago
Vladimír Vondruš	22d1484814	Test: rename Any* plugin test cases for clarity.	5 years ago
Vladimír Vondruš	fb2eddb6bb	doc: updated credits and changelog.	5 years ago
Aaron Gokaslan	ca677c69fe	Fix typos	5 years ago
Vladimír Vondruš	ffdbae5974	ObjImporter: fix a GCC 11 warning. Ew, this code is terrible.	5 years ago
Vladimír Vondruš	7346dc2d01	package/ci: move a comment to where the macOS image is defined.	5 years ago
Vladimír Vondruš	f216320da6	GCC 11, your fancy new warnings are NOT helpful.	5 years ago
Vladimír Vondruš	202dff63ae	package/ci: MSYS, the ancient curses seem to have returned.	5 years ago
Vladimír Vondruš	c942e3e764	Shaders: Bitangent is disallowed only with InstancedObjectId. With just ObjectId it's fine. Sorry for a silly breakage!	5 years ago
Vladimír Vondruš	93fd668387	Shaders: wait, this should work, and not assert.	5 years ago
Vladimír Vondruš	3102678d19	doc: updated credits.	5 years ago
Aaron Gokaslan	2f31891420	Add noexcept here as well.	5 years ago
Aaron Gokaslan	712a3b635f	Platform: mark Windowless*Context moves as noexcept.	5 years ago
Vladimír Vondruš	ef77fac659	Animation: explicitly mark Track move construction/assignment as noexcept. Interestingly enough, GCC reported these were noexcept even without that annotation. Co-authored-by: Aaron Gokaslan <aaronGokaslan@gmail.com>	5 years ago
Aaron Gokaslan	1cfa2d0fce	GL: remove an unneeded std::move().	5 years ago
Vladimír Vondruš	37fc638d12	package/ci: had to upgrade to 10.15 because of MoltenVK. Sigh, that's like those projects that require GCC 13 to build.	5 years ago
Vladimír Vondruš	cc74784d40	Shaders: ability to disable Phong specular contribution. In cases when specular highlights are not desired, results in 30% speedup (on Intel) and ~25% speedup on AMD, compared to setting the specular color to transparent black. Testing was easy thanks to already having a ground truth image for this case.	5 years ago
Vladimír Vondruš	a305a1ed85	Shaders: benchmarks Phong UBOs with ARB_buffer_storage as well. Just to have at least one case where it's considered. Doesn't help that much, so probably not even worth bothering with.	5 years ago
Vladimír Vondruš	afd8d7c8f9	Shaders: introduce Phong::Flag::LightCulling. After several failed attempts to make UBO performance not suck on Intel Mesa and Windows drivers, I ended up hiding the dynamic aspect under a flag. That way it's still possible to get the proper perf in UBO workflows that don't do light culling, and for workflows where light culling matters the 2x slowdown might be still better than looping through several extra lights that don't contribute anything.	5 years ago
Vladimír Vondruš	47b5940a89	Shaders: microoptimize redundant square root. Just to have all options tried out, no significant impact this time again.	5 years ago
Vladimír Vondruš	0b7c7f8249	Shaders: hoist camera calculation calculation out of the loop. While (of course) having zero effect on a single-light scenario, with five lights it saves about 10% in the classic uniform case (on Intel). Not bad (also, FFS, what the compiler is doing if it's not able to optimize this?!).	5 years ago
Vladimír Vondruš	000c332eee	Shaders: this was a stupid thing to do. Doesn't actually seem to have any measurable impact (on Intel), so I guess the compiler can realize this as well or it was really unimpactful.	5 years ago
Vladimír Vondruš	bda01bb3a5	Shaders: benchmark Phong with 64 lights but just 5 used. Should be roughly the same perf as one with 5 set statically.	5 years ago
Vladimír Vondruš	df75ab926a	Shaders: calculate Phong light directions in the fragment shader. Hm, I wonder why I did it this way, the operation isn't really heavy to benefit from the hardware interpolators, and with a lot of lights we'd hit the maximum output count in the vertex shader. With classic uniforms this seems to be ~5% faster on both Intel and AMD cards. With UBOs this is ~15% (!) faster on AMD (I guess because the constant UBO access overhead is moved to just one stage instead of both?) but slower on Intel (of course, sigh... I assume due to UBO reads being slow and so when done for every fragment instead of every vertex it costs more?). Since this benefits the real GPUs while the card that was already awful is more awful I don't think that's a big deal. This change stays.	5 years ago

1 2 3 4 5 ...

8975 Commits (2479b34165ba141cb9233e98bfdf5bd037953493) All Branches Search

8975 Commits (2479b34165ba141cb9233e98bfdf5bd037953493)

All Branches