-
-
Notifications
You must be signed in to change notification settings - Fork 314
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Moves Group Varint implementation out of inline.
- Loading branch information
1 parent
70ed8a3
commit 48e5f5f
Showing
4 changed files
with
168 additions
and
117 deletions.
There are no files selected for viewing
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
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
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
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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,148 @@ | ||
| //----------------------------------------------------------------------------// | ||
| // // | ||
| // ozz-animation is hosted at http://github.com/guillaumeblanc/ozz-animation // | ||
| // and distributed under the MIT License (MIT). // | ||
| // // | ||
| // Copyright (c) Guillaume Blanc // | ||
| // // | ||
| // Permission is hereby granted, free of charge, to any person obtaining a // | ||
| // copy of this software and associated documentation files (the "Software"), // | ||
| // to deal in the Software without restriction, including without limitation // | ||
| // the rights to use, copy, modify, merge, publish, distribute, sublicense, // | ||
| // and/or sell copies of the Software, and to permit persons to whom the // | ||
| // Software is furnished to do so, subject to the following conditions: // | ||
| // // | ||
| // The above copyright notice and this permission notice shall be included in // | ||
| // all copies or substantial portions of the Software. // | ||
| // // | ||
| // THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR // | ||
| // IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, // | ||
| // FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL // | ||
| // THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER // | ||
| // LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING // | ||
| // FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER // | ||
| // DEALINGS IN THE SOFTWARE. // | ||
| // // | ||
| //----------------------------------------------------------------------------// | ||
|
|
||
| #include "ozz/base/encode/group_varint.h" | ||
|
|
||
| namespace ozz { | ||
|
|
||
| namespace internal { | ||
|
|
||
| // Find the number of bytes (-1) required to store the integer value | ||
| OZZ_INLINE uint8_t tag(uint32_t _v) { | ||
| return (_v >= (1 << 24)) + (_v >= (1 << 16)) + (_v >= (1 << 8)); | ||
| } | ||
|
|
||
| // Copies 4 bytes integer value to a byte buffer. All 4 bytes are written, | ||
| // whatever the value. | ||
| // Value is stored in little endian order. memcpy isn't used to ensure load and | ||
| // store can be done from different endianness systems. | ||
| OZZ_INLINE void store(uint32_t _v, ozz::byte* _buffer) { | ||
| _buffer[0] = _v & 0xff; | ||
| _buffer[1] = (_v >> 8) & 0xff; | ||
| _buffer[2] = (_v >> 16) & 0xff; | ||
| _buffer[3] = (_v >> 24) & 0xff; | ||
| } | ||
|
|
||
| // Copies 4 bytes integer value from a byte buffer. All 4 bytes are read, | ||
| // whatever the value. | ||
| // Value is stored in little endian order. memcpy isn't used to ensure load and | ||
| // store can be done from different endianness systems. | ||
| OZZ_INLINE uint32_t load(const ozz::byte* _in) { | ||
| return static_cast<uint32_t>(_in[0]) | static_cast<uint32_t>(_in[1]) << 8 | | ||
| static_cast<uint32_t>(_in[2]) << 16 | | ||
| static_cast<uint32_t>(_in[3]) << 24; | ||
| } | ||
| } // namespace internal | ||
|
|
||
| inline ozz::span<ozz::byte> EncodeGV4(const ozz::span<const uint32_t>& _input, | ||
| const ozz::span<ozz::byte>& _buffer) { | ||
| assert(_input.size() == 4 && "Input size must be 4"); | ||
| assert(_buffer.size_bytes() >= 4 * sizeof(uint32_t) + 1 && | ||
| "Output buffer is too small."); | ||
| const uint8_t a_tag = internal::tag(_input[0]); | ||
| const uint8_t b_tag = internal::tag(_input[1]); | ||
| const uint8_t c_tag = internal::tag(_input[2]); | ||
| const uint8_t d_tag = internal::tag(_input[3]); | ||
|
|
||
| ozz::byte* out = _buffer.data(); | ||
|
|
||
| // Compute and store prefix. | ||
| *out++ = static_cast<ozz::byte>((d_tag << 6) | (c_tag << 4) | (b_tag << 2) | | ||
| a_tag); | ||
| internal::store(_input[0], out); | ||
| out += a_tag + 1; | ||
| internal::store(_input[1], out); | ||
| out += b_tag + 1; | ||
| internal::store(_input[2], out); | ||
| out += c_tag + 1; | ||
| internal::store(_input[3], out); | ||
| out += d_tag + 1; | ||
| return {out, _buffer.end()}; | ||
| } | ||
|
|
||
| inline ozz::span<const ozz::byte> DecodeGV4( | ||
| const ozz::span<const ozz::byte>& _buffer, | ||
| const ozz::span<uint32_t>& _output) { | ||
| assert(_buffer.size_bytes() >= 5 && "Input buffer is too small."); | ||
| assert(_output.size() == 4 && "Output size must be 4"); | ||
|
|
||
| const ozz::byte* in = _buffer.data(); | ||
| const uint8_t prefix = *in++; | ||
|
|
||
| constexpr uint32_t kMask[4] = {0xff, 0xffff, 0xffffff, 0xffffffff}; | ||
| const uint8_t k0 = prefix & 0x3; | ||
| _output[0] = internal::load(in) & kMask[k0]; | ||
| in += k0 + 1; | ||
| const uint8_t k1 = (prefix >> 2) & 0x3; | ||
| _output[1] = internal::load(in) & kMask[k1]; | ||
| in += k1 + 1; | ||
| const uint8_t k2 = (prefix >> 4) & 0x3; | ||
| _output[2] = internal::load(in) & kMask[k2]; | ||
| in += k2 + 1; | ||
| const uint8_t k3 = prefix >> 6; | ||
| _output[3] = internal::load(in) & kMask[k3]; | ||
| in += k3 + 1; | ||
| return {in, _buffer.end()}; | ||
| } | ||
|
|
||
| size_t ComputeGV4WorstBufferSize(const ozz::span<const uint32_t>& _stream) { | ||
| assert((_stream.size() % 4) == 0 && "Input stream must be multiple of 4"); | ||
| return _stream.size() * 4 + _stream.size() / 4; | ||
| } | ||
|
|
||
| ozz::span<ozz::byte> EncodeGV4Stream(const ozz::span<const uint32_t>& _stream, | ||
| const ozz::span<ozz::byte>& _buffer) { | ||
| assert((_stream.size() % 4) == 0 && "Input stream must be multiple of 4"); | ||
| assert(_buffer.size_bytes() >= ComputeGV4WorstBufferSize(_stream) && | ||
| "Output buffer is too small"); | ||
|
|
||
| ozz::span<ozz::byte> out = _buffer; | ||
| for (const uint32_t* data = _stream.begin(); data < _stream.end(); | ||
| data += 4) { | ||
| out = EncodeGV4({data, 4}, out); | ||
| } | ||
|
|
||
| return out; | ||
| } | ||
|
|
||
| ozz::span<const ozz::byte> DecodeGV4Stream( | ||
| const ozz::span<const ozz::byte>& _buffer, | ||
| const ozz::span<uint32_t>& _stream) { | ||
| assert((_stream.size() % 4) == 0 && "Input stream must be multiple of 4"); | ||
| // Check for minimum possible buffer size (5 bytes for 4 integers). | ||
| assert(_buffer.size_bytes() >= (_stream.size() + _stream.size() / 4) && | ||
| "Output buffer is too small"); | ||
|
|
||
| ozz::span<const ozz::byte> in = _buffer; | ||
| for (uint32_t* data = _stream.begin(); data < _stream.end(); data += 4) { | ||
| in = DecodeGV4(in, {data, 4}); | ||
| } | ||
|
|
||
| return in; | ||
| } | ||
|
|
||
| } // namespace ozz |