More fft utils #795
Merged
More fft utils #795
Conversation
This file contains 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
Comment on lines
16
to
28
| static inline uint32_t3 padDimensions(uint32_t3 dimensions, std::span<uint16_t> axes, bool realFFT = false) | ||
| { | ||
| uint16_t axisCount = 0; | ||
| for (auto i : axes) | ||
| { | ||
| dimensions[i] = core::roundUpToPoT(dimensions[i]); | ||
| if (realFFT && !axisCount++) | ||
| dimensions[i] /= 2; | ||
| } | ||
| return dimensions; | ||
| } | ||
|
|
||
| static inline uint64_t getOutputBufferSize(const uint32_t3& inputDimensions, uint32_t numChannels, std::span<uint16_t> axes, bool realFFT = false, bool halfFloats = false) |
There was a problem hiding this comment.
you could have just replaced uint32_t3 with vector<uint32_t,M> and made it available both for C++ and HLSL
also why is span<uint16_t> used, you could just deduce the axis count from N ?
roundUpToPoT could move to hlsl namespace actually
There was a problem hiding this comment.
roundUpToPoT is in core/math/intutil.h. I can create a builtin/math/intutil.hlsl, copy most functions over and refactor every usage of the functions in that file
There was a problem hiding this comment.
roundUpToPoTis incore/math/intutil.h. I can create abuiltin/math/intutil.hlsl, copy most functions over and refactor every usage of the functions in that file
ask @Przemog1 about the preferred location so it fits in with the spirit of #801
You don't actually need to do big refactor if you provide a "legacy" alias (make the old nbl::core::roundUpToPoT call nbl::hlsl::roundUpToPoT) and add the [[deprecated]] attribute
There was a problem hiding this comment.
I'm creating builtin/hlsl/math/intutil.hlsl, @Przemog1 lmk if you'd rather have it be named different or placed somewhere else
…ity for realt FFTs
…in functional.hlsl . Small change to FFT Indexing utils
… between cpp and hlsl. Only thing missing is to move intutils.h to a common header as well
this allows the offset accessor passed in the bigger FFTs to pass the concept requirements - Minor change to the offset accessor, I think it follows the original intent
that moved to .hlsl - Update the GLSL bitreverse to be in-line with the GLSL spec
Comment on lines
12
to
84
| template<typename Integer NBL_FUNC_REQUIRES(is_integral_v<Integer>) | ||
| NBL_CONSTEXPR_FORCED_INLINE_FUNC bool isNPoT(Integer value) | ||
| { | ||
| return value & (value - Integer(1)); | ||
| } | ||
|
|
||
| template<typename Integer NBL_FUNC_REQUIRES(is_integral_v<Integer>) | ||
| NBL_CONSTEXPR_FORCED_INLINE_FUNC bool isPoT(Integer value) | ||
| { | ||
| return !isNPoT<Integer>(value); | ||
| } | ||
|
|
||
|
|
||
| template<typename Integer NBL_FUNC_REQUIRES(is_integral_v<Integer>) | ||
| NBL_CONSTEXPR_FORCED_INLINE_FUNC Integer roundUpToPoT(Integer value) | ||
| { | ||
| return Integer(0x1u) << Integer(1 + hlsl::findMSB<Integer>(value - Integer(1))); // this wont result in constexpr because findMSB is not one | ||
| } | ||
|
|
||
| template<typename Integer NBL_FUNC_REQUIRES(is_integral_v<Integer>) | ||
| NBL_CONSTEXPR_FORCED_INLINE_FUNC Integer roundDownToPoT(Integer value) | ||
| { | ||
| return Integer(0x1u) << hlsl::findMSB<Integer>(value); | ||
| } | ||
|
|
||
| template<typename Integer NBL_FUNC_REQUIRES(is_integral_v<Integer>) | ||
| NBL_CONSTEXPR_FORCED_INLINE_FUNC Integer roundUp(Integer value, Integer multiple) | ||
| { | ||
| Integer tmp = (value + multiple - 1u) / multiple; | ||
| return tmp * multiple; | ||
| } | ||
|
|
||
| template<typename Integer NBL_FUNC_REQUIRES(is_integral_v<Integer>) | ||
| NBL_CONSTEXPR_FORCED_INLINE_FUNC Integer align(Integer alignment, Integer size, NBL_REF_ARG(Integer) address, NBL_REF_ARG(Integer) space) | ||
| { | ||
| Integer nextAlignedAddr = roundUp<Integer>(address, alignment); | ||
|
|
||
| Integer spaceDecrement = nextAlignedAddr - address; | ||
| if (spaceDecrement > space) | ||
| return 0u; | ||
|
|
||
| Integer newSpace = space - spaceDecrement; | ||
| if (size > newSpace) | ||
| return 0u; | ||
|
|
||
| space = newSpace; | ||
| return address = nextAlignedAddr; | ||
| } | ||
|
|
||
| #ifndef __HLSL_VERSION | ||
|
|
||
| // Have to wait for the HLSL patch for `is_enum`. Would also have to figure out how to do it without initializer lists for HLSL use. | ||
|
|
||
| //! Get bitmask from variadic arguments passed. | ||
| /* | ||
| For example if you were to create bitmask for vertex attributes | ||
| having positions inteeger set as 0, colors as 1 and normals | ||
| as 3, just pass them to it and use the value returned. | ||
| */ | ||
|
|
||
| template<typename BitmaskType NBL_FUNC_REQUIRES(is_integral_v<BitmaskType> || std::is_enum_v<BitmaskType>) | ||
| NBL_CONSTEXPR_FORCED_INLINE_FUNC uint64_t createBitmask(std::initializer_list<BitmaskType> initializer) | ||
| { | ||
| uint64_t retval{}; | ||
| for (const auto& it : initializer) | ||
| retval |= (1ull << it); | ||
| return retval; | ||
| } | ||
|
|
||
| #endif | ||
|
|
||
| } // end namespace hlsl | ||
| } // end namespace nbl |
Merged
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.
Description
Still adding more changes on this side
Testing
All thesting done on the FFT_Bloom example on its respective branch