Remove Matrices #760

devshgraphicsprogramming · 2024-10-25T07:47:05Z

Description

Testing

TODO list:

include/nbl/builtin/hlsl/glsl_compat/math.hlsl

devshgraphicsprogramming · 2024-10-25T07:48:52Z

include/nbl/builtin/hlsl/glsl_compat/math.hlsl

+template<typename T>
+inline T radians(NBL_CONST_REF_ARG(T) degrees)
+{
+	static_assert(
+		is_floating_point<T>::value,
+		"This code expects the type to be either a double or a float."
+	);
+
+	return degrees * PI<T>() / T(180);
+}
+
+template<typename T>
+inline T degrees(NBL_CONST_REF_ARG(T) radians)
+{
+	static_assert(
+		is_floating_point<T>::value,
+		"This code expects the type to be either a double or a float."
+	);
+
+	return radians * T(180) / PI<T>();
+}


we have numbers ot something like that for PI and friends.

Also the _HLSL_VERSION case should forward to the inline SPIR-V from the std450 extended set

devshgraphicsprogramming · 2024-10-25T07:49:42Z

include/nbl/builtin/hlsl/math/quaternion/quaternion.hlsl

+	DEFINE_MUL_QUATERNION_BY_SCALAR_OPERATOR(uint32_t)
+	DEFINE_MUL_QUATERNION_BY_SCALAR_OPERATOR(uint64_t)


this is a mistake, only provide mul with T

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl

…ls.hlsl from #760

…H (ambiguity dependent type issues), reference #760

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl

…files#r1816728485 comment in Devsh-Graphics-Programming/Nabla#760 PR

…ng/Nabla/pull/760/files#r1816728485 for #760 PR, update examples_tests submodule

cmake/adjust/flags.cmake

AnastaZIuk · 2024-11-04T17:37:36Z

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl


 namespace nbl
 {
 namespace hlsl
 {

+// TODO: if `IdentityFloat32_t3x4` and `IdentityFloat32_t3x4` constexprs are ok, then I can expand them into templated struct, not doing it untill the concept is approved
+//template<typename T, uint32_t N, uint32_t M>
+//struct IdentityMatrix


add

template<typename T> concept MatrixType = requires { typename T::Base::row_type; typename T::Base::col_type; typename T::Base::value_type; } && std::is_base_of_v<glm::mat<T::Base::row_type::length(), T::Base::col_type::length(), typename T::Base::value_type>, typename T::Base>;

to matrix.hlsl just after definition (but ask @devshgraphicsprogramming if it should be named MatrixType)

then create diagonal utility

template<MatrixType Matrix> constexpr inline Matrix diagonal(float scalar) { return Matrix(scalar); }

you can also template the scalar actually and make it value_type of the Matrix, now note that

matrix<T, N, M>(scalar) creates a diagonal matrix with scalar put on the diagonal

if your N == M then matrix<T, N, M>(1.f) is the identity matrix you need

so following diagonal create identity function (not struct!), you should be able to

using any_square_matrix_t = float32_t3x3; constexpr auto bruh = hlsl::identity<any_square_matrix_t>()

have constraints on the template parameter, let's allow only for square matrices

but this is not C++ only file

so we can't use concepts..

so we can't use concepts..

hehe, you can! 492a0ad

typename T::Base::row_type; typename T::Base::col_type; typename T::Base::value_type;

this is specific to GLM and won't pass in plain HLSL

AnastaZIuk · 2024-11-04T17:40:20Z

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl

+//{
+//
+//};
+NBL_CONSTEXPR hlsl::float32_t3x4 IdentityFloat32_t3x4 =


lets not hardcode and follow this comment

AnastaZIuk · 2024-11-04T17:40:23Z

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl

+//};
+NBL_CONSTEXPR hlsl::float32_t3x4 IdentityFloat32_t3x4 =
+	hlsl::float32_t3x4(hlsl::float32_t4(1, 0, 0, 0), hlsl::float32_t4(0, 0, 1, 0), hlsl::float32_t4(0, 0, 1, 0));
+NBL_CONSTEXPR hlsl::float32_t4x4 IdentityFloat32_t4x4 =


lets not hardcode and follow this comment

AnastaZIuk · 2024-11-04T17:41:48Z

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl

 template<typename T>
-matrix<T, 4, 4> getMatrix3x4As4x4(const matrix<T, 3, 4>& mat)
+inline core::vectorSIMDf transformVector(const matrix<T, 4, 4>& mat, const core::vectorSIMDf& vec)


shouldn't we also remove SIMD vectors?

will do in other pr

AnastaZIuk · 2024-11-04T17:42:19Z

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl

-matrix<T, 4, 4> getMatrix3x4As4x4(const matrix<T, 3, 4>& mat)
+inline core::vectorSIMDf transformVector(const matrix<T, 4, 4>& mat, const core::vectorSIMDf& vec)
+{
+	core::vectorSIMDf output;


https://github.com/Devsh-Graphics-Programming/Nabla/pull/760/files#r1828132005

I will remove every occurance of the simd vector when I work on it, no need to spam xd

AnastaZIuk · 2024-11-04T18:05:52Z

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl

+namespace transformation_matrix_utils_impl
+{
+	template<typename T>
+	inline T determinant_helper(const matrix<T, 3, 3>& mat, vector<T, 3>& r1crossr2)


please document that your returned determinant is yielded by scalar triple product

AnastaZIuk · 2024-11-04T18:08:09Z

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl

-	const vector<T, 3> zaxis = glm::normalize(position - target);
-	const vector<T, 3> xaxis = glm::normalize(hlsl::cross(upVector, zaxis));
-	const vector<T, 3> yaxis = hlsl::cross(zaxis, xaxis);
+	static_assert(N >= 3 && M >= 3);


instead of doing static_assert lets have it in requires clause

can't do since it is not C++ only file

you can, NBL_FUNC_REQUIRES

AnastaZIuk · 2024-11-04T18:17:07Z

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl

-	const vector<T, 3>& target,
-	const vector<T, 3>& upVector)
+template<typename T, uint32_t N, uint32_t M>
+inline matrix<T, 3, 3> getSub3x3TransposeCofactors(const matrix<T, N, M>& mat)


please document it returns adjugate (transpose of the cofactor matrix) of sub 3x3 mat matrix

AnastaZIuk · 2024-11-04T18:27:08Z

include/nbl/core/math/glslFunctions.h

-NBL_FORCE_INLINE T dot(const T& a, const T& b);
+NBL_FORCE_INLINE T dot(const T& a, const T& b)
+{
+	static_assert(!(std::is_same_v<T, hlsl::float32_t2> || std::is_same_v<T, hlsl::float32_t3> || std::is_same_v<T, hlsl::float32_t4>));


move to requires clause

can't do since it is not C++ only file

you can, NBL_FUNC_REQUIRES

…anguage/inline and afterwards https://gudok.xyz/inline/ please

Signed-off-by: Ali Cheraghi <[email protected]>

FFT Fixes

Signed-off-by: Ali Cheraghi <[email protected]>

blake3: add reset()

Signed-off-by: Ali Cheraghi <[email protected]>

Prefix sum box blur

Signed-off-by: Ali Cheraghi <[email protected]>

video: update profiles

Fix false positive aspect mask error raising

Signed-off-by: Ali Cheraghi <[email protected]>

video: free command buffers when destroyed

devshgraphicsprogramming commented Oct 25, 2024

View reviewed changes

include/nbl/builtin/hlsl/glsl_compat/math.hlsl Outdated Show resolved Hide resolved

devshgraphicsprogramming commented Oct 25, 2024

View reviewed changes

AnastaZIuk reviewed Oct 25, 2024

View reviewed changes

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl Outdated Show resolved Hide resolved

AnastaZIuk reviewed Oct 25, 2024

View reviewed changes

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl Outdated Show resolved Hide resolved

AnastaZIuk reviewed Oct 25, 2024

View reviewed changes

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl Outdated Show resolved Hide resolved

AnastaZIuk reviewed Oct 25, 2024

View reviewed changes

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl Show resolved Hide resolved

AnastaZIuk reviewed Oct 25, 2024

View reviewed changes

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl Show resolved Hide resolved

AnastaZIuk added a commit that referenced this pull request Oct 25, 2024

steal include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_uti…

22f505c

…ls.hlsl from #760

AnastaZIuk added a commit that referenced this pull request Oct 25, 2024

fix getMatrix3x4As4x4 (return type issues) & buildCameraLookAtMatrixR…

51f4ced

…H (ambiguity dependent type issues), reference #760

AnastaZIuk reviewed Oct 25, 2024

View reviewed changes

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl Outdated Show resolved Hide resolved

AnastaZIuk reviewed Oct 25, 2024

View reviewed changes

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl Outdated Show resolved Hide resolved

AnastaZIuk reviewed Oct 25, 2024

View reviewed changes

include/nbl/builtin/hlsl/matrix_utils/transformation_matrix_utils.hlsl Outdated Show resolved Hide resolved

AnastaZIuk added a commit that referenced this pull request Oct 25, 2024

actually *this* addresses https://github.com/Devsh-Graphics-Programmi…

f1daa25

…ng/Nabla/pull/760/files#r1816728485 for #760 PR, update examples_tests submodule

Przemog1 changed the base branch from master to cameraz November 4, 2024 13:28

AnastaZIuk reviewed Nov 4, 2024

View reviewed changes

cmake/adjust/flags.cmake Outdated Show resolved Hide resolved

AnastaZIuk reviewed Nov 4, 2024

View reviewed changes

devsh and others added 4 commits November 11, 2024 17:05

Merge remote-tracking branch 'origin/master' into bindless_blit

1edbc38

Merge branch 'master' into more_fft_utils

7ca1e7d

@alichraghi small homework - read https://en.cppreference.com/w/cpp/l…

51cbd88

…anguage/inline and afterwards https://gudok.xyz/inline/ please

CVulkanSwapchain allow 2 more acquiresInFlight

6c0b3cb

Fletterio and others added 27 commits January 17, 2025 14:40

Update accessor concepts

eee9904

move E_TEXTURE_CLAMP into hlsl headers

83db408

Signed-off-by: Ali Cheraghi <[email protected]>

box_blur: move into here from example code

19973ac

Signed-off-by: Ali Cheraghi <[email protected]>

Address all PR comments

5518e01

Merge pull request #818 from Devsh-Graphics-Programming/more_fft_utils

0f8bcac

FFT Fixes

FFT readme update

2e52107

Update examples submodule ptr

ae5dbad

prefix_sum_blur: address reviews

d2adf17

Signed-off-by: Ali Cheraghi <[email protected]>

FFT readme update and accessor concept change to remove memory barriers

2782d8d

FFT readme update

a46b9b0

box_sampler: rework ETC_REPEAT

b65dc7f

Signed-off-by: Ali Cheraghi <[email protected]>

box_sampler: simplify ETC_MIRROR

6f74af2

Signed-off-by: Ali Cheraghi <[email protected]>

blake3: add reset()

75ffca1

Signed-off-by: Ali Cheraghi <[email protected]>

Merge pull request #822 from Devsh-Graphics-Programming/ali_blake_reset

43323a6

blake3: add reset()

box_sampler: update ETC_REPEAT

b1f2827

Signed-off-by: Ali Cheraghi <[email protected]>

Merge pull request #819 from Devsh-Graphics-Programming/ali_blur3

45d97fd

Prefix sum box blur

post Prefix Sum Bloom merge submodule update

bacb479

Update submodule pointer

f4b2626

video: update profiles

1e05ae4

Signed-off-by: Ali Cheraghi <[email protected]>

video: remove bad report and exclude android from core profile

68ca358

Signed-off-by: Ali Cheraghi <[email protected]>

Merge pull request #824 from Devsh-Graphics-Programming/ali_profiles

4786926

video: update profiles

Fix false positive aspect mask error raising

eac5159

Merge pull request #826 from Crisspl/quick-fix

72bd107

Fix false positive aspect mask error raising

cmake: no longer dies when NBL_SANITIZE_ADDRESS is ON

25ef700

Signed-off-by: Ali Cheraghi <[email protected]>

video: free command buffers when destroyed

d8e5751

Signed-off-by: Ali Cheraghi <[email protected]>

Merge pull request #830 from Devsh-Graphics-Programming/ali_cmd

8b67720

video: free command buffers when destroyed

Resolved conflicts, merged master

4a894fe

Przemog1 closed this Feb 7, 2025

Przemog1 deleted the rm_core_mat_vec branch February 7, 2025 10:12

devshgraphicsprogramming mentioned this pull request Feb 26, 2025

Remove SIMD vectors #837

Open

		DEFINE_MUL_QUATERNION_BY_SCALAR_OPERATOR(uint32_t)
		DEFINE_MUL_QUATERNION_BY_SCALAR_OPERATOR(uint64_t)

Remove Matrices #760

Remove Matrices #760

Uh oh!

Conversation

devshgraphicsprogramming commented Oct 25, 2024

Description

Testing

TODO list:

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AnastaZIuk Nov 4, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AnastaZIuk Nov 4, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AnastaZIuk Nov 4, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AnastaZIuk Nov 4, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

AnastaZIuk Nov 4, 2024 •

edited

Loading

AnastaZIuk Nov 4, 2024 •

edited

Loading

AnastaZIuk Nov 4, 2024 •

edited

Loading

AnastaZIuk Nov 4, 2024 •

edited

Loading