bitshuffle_hperf is an x86-only implementation - its entire vector body sits behind __i386__/__x86_64__, so on aarch64 every entry point compiles down to the scalar fallback. Measured against its own SIMD path that costs 8.3x on encode and 3.8x on decode, and it is the transform behind every compressed image the writer produces and every one the reader, preview and XDS plugin take apart again. The classic bitshuffle vendored beside it does have an aarch64 NEON path, and is already compiled into the same target, so this costs nothing new. BitShuffleBlock.h picks bshuf_trans_bit_elem/bshuf_untrans_bit_elem there and keeps bitshuf_encode_block / bitshuf_decode_block everywhere else, where hperf is about twice classic SSE2 and remains the better choice. The expected aarch64 gain is ~2.5x encode / ~1.7x decode: classic NEON is 128-bit and carries an extra pass, so it recovers part of the gap rather than all of it. The condition mirrors USEARMNEON in bitshuffle_core.c exactly, because with NEON off the classic scalar path is slower than hperf's and must not be selected. Swapping implementations is only safe while the two agree bit for bit - otherwise an ARM build would write files an x86 build could not read. They do: verified byte-identical output and mutual cross-decoding for elem_size 1/2/4/8 over block sizes from 8 to 65536 elements. Both are always compiled in, so the new test holds them to it on every architecture, not just the one that would notice. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
188 lines
8.8 KiB
C++
188 lines
8.8 KiB
C++
// SPDX-FileCopyrightText: 2024 Filip Leonarski, Paul Scherrer Institute <filip.leonarski@psi.ch>
|
|
// SPDX-License-Identifier: GPL-3.0-only
|
|
|
|
#pragma once
|
|
|
|
#include <vector>
|
|
#include <cstring>
|
|
|
|
#include <bitshuffle/bitshuffle.h>
|
|
#include <bitshuffle/bitshuffle_internals.h>
|
|
#include <lz4/lz4.h>
|
|
#include <zstd.h>
|
|
|
|
#include "BitShuffleBlock.h"
|
|
#include "../compression/CompressionAlgorithmEnum.h"
|
|
|
|
#include "../common/JFJochException.h"
|
|
#include "../common/CompressedImage.h"
|
|
|
|
extern "C" {
|
|
uint64_t bshuf_read_uint64_BE(const void* buf);
|
|
};
|
|
|
|
inline size_t JFJochDecompressHperfPtr(uint8_t *output,
|
|
CompressionAlgorithm algorithm,
|
|
const uint8_t *source,
|
|
size_t source_size,
|
|
size_t nelements,
|
|
size_t elem_size,
|
|
size_t block_size) {
|
|
if ((algorithm != CompressionAlgorithm::BSHUF_LZ4) &&
|
|
(algorithm != CompressionAlgorithm::BSHUF_ZSTD) &&
|
|
(algorithm != CompressionAlgorithm::BSHUF_ZSTD_RLE) &&
|
|
(algorithm != CompressionAlgorithm::BSHUF_ZSTD_RLE_HUFF))
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Algorithm not supported by hperf decompressor");
|
|
|
|
if ((block_size == 0) || ((block_size % BSHUF_BLOCKED_MULT) != 0))
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Invalid block size");
|
|
|
|
std::vector<char> decompressed_block(block_size * elem_size);
|
|
std::vector<char> scratch(block_size * elem_size);
|
|
|
|
const uint8_t *src_ptr = source;
|
|
const uint8_t *const source_end = source + source_size;
|
|
uint8_t *dst_ptr = output;
|
|
|
|
const size_t num_full_blocks = nelements / block_size;
|
|
const size_t reminder_size = nelements - num_full_blocks * block_size;
|
|
const size_t last_block_size = reminder_size - reminder_size % BSHUF_BLOCKED_MULT;
|
|
|
|
auto decode_block = [&](size_t current_nelements) {
|
|
// Both the block length and the block itself come from the stream, which for the CBOR path
|
|
// and for the XDS plugin is data we did not produce.
|
|
if (source_end - src_ptr < 4)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Truncated compressed block header");
|
|
|
|
const auto compressed_size = static_cast<size_t>(bshuf_read_uint32_BE(src_ptr));
|
|
src_ptr += 4;
|
|
|
|
if (static_cast<size_t>(source_end - src_ptr) < compressed_size)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Compressed block extends past the input buffer");
|
|
|
|
const size_t expected_size = current_nelements * elem_size;
|
|
size_t decompressed_size = 0;
|
|
|
|
switch (algorithm) {
|
|
case CompressionAlgorithm::BSHUF_LZ4: {
|
|
const int ret = LZ4_decompress_safe(reinterpret_cast<const char *>(src_ptr),
|
|
decompressed_block.data(),
|
|
static_cast<int>(compressed_size),
|
|
static_cast<int>(expected_size));
|
|
if (ret < 0 || static_cast<size_t>(ret) != expected_size)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "LZ4 decompression error");
|
|
decompressed_size = static_cast<size_t>(ret);
|
|
break;
|
|
}
|
|
case CompressionAlgorithm::BSHUF_ZSTD:
|
|
case CompressionAlgorithm::BSHUF_ZSTD_RLE:
|
|
case CompressionAlgorithm::BSHUF_ZSTD_RLE_HUFF: {
|
|
const size_t ret = ZSTD_decompress(decompressed_block.data(),
|
|
expected_size,
|
|
src_ptr,
|
|
compressed_size);
|
|
if (ZSTD_isError(ret) || ret != expected_size)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "ZSTD decompression error");
|
|
decompressed_size = ret;
|
|
break;
|
|
}
|
|
default:
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Algorithm not supported");
|
|
}
|
|
|
|
if (JFJochBitUnshuffleBlock(reinterpret_cast<char *>(dst_ptr),
|
|
decompressed_block.data(),
|
|
scratch.data(),
|
|
current_nelements,
|
|
elem_size) < 0)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "bitshuffle block decode error");
|
|
|
|
src_ptr += compressed_size;
|
|
dst_ptr += decompressed_size;
|
|
};
|
|
|
|
for (size_t i = 0; i < num_full_blocks; ++i)
|
|
decode_block(block_size);
|
|
|
|
if (last_block_size > 0)
|
|
decode_block(last_block_size);
|
|
|
|
const size_t leftover_bytes = (reminder_size % BSHUF_BLOCKED_MULT) * elem_size;
|
|
if (leftover_bytes > 0) {
|
|
if (static_cast<size_t>(source_end - src_ptr) < leftover_bytes)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Truncated trailing bytes");
|
|
memcpy(dst_ptr, src_ptr, leftover_bytes);
|
|
src_ptr += leftover_bytes;
|
|
}
|
|
|
|
return static_cast<size_t>(src_ptr - source);
|
|
}
|
|
|
|
inline void JFJochDecompressPtr(uint8_t *output,
|
|
CompressionAlgorithm algorithm,
|
|
const uint8_t *source,
|
|
size_t source_size,
|
|
size_t nelements,
|
|
size_t elem_size,
|
|
bool use_hperf = true) {
|
|
size_t block_size = 0;
|
|
if (algorithm != CompressionAlgorithm::NO_COMPRESSION) {
|
|
// The 12-byte bitshuffle header must be there before it can be read, and before
|
|
// source_size - 12 is handed to the decompressors below.
|
|
if (source_size < 12)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Buffer too short for the bitshuffle header");
|
|
if (bshuf_read_uint64_BE(const_cast<uint8_t *>(source)) != nelements * elem_size)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Mismatch in size");
|
|
auto tmp = bshuf_read_uint32_BE(source + 8);
|
|
block_size = tmp / elem_size;
|
|
}
|
|
|
|
switch (algorithm) {
|
|
case CompressionAlgorithm::NO_COMPRESSION:
|
|
if (source_size != nelements * elem_size)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Mismatch in size");
|
|
memcpy(output, source, source_size);
|
|
break;
|
|
case CompressionAlgorithm::BSHUF_LZ4:
|
|
if (use_hperf) {
|
|
if (JFJochDecompressHperfPtr(output, algorithm, source + 12, source_size - 12,
|
|
nelements, elem_size, block_size) != source_size - 12)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Decompression error");
|
|
} else {
|
|
if (bshuf_decompress_lz4(source + 12, output, nelements,
|
|
elem_size, block_size) != source_size - 12)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Decompression error");
|
|
}
|
|
break;
|
|
case CompressionAlgorithm::BSHUF_ZSTD_RLE:
|
|
case CompressionAlgorithm::BSHUF_ZSTD_RLE_HUFF:
|
|
case CompressionAlgorithm::BSHUF_ZSTD:
|
|
if (use_hperf) {
|
|
if (JFJochDecompressHperfPtr(output, algorithm, source + 12, source_size - 12,
|
|
nelements, elem_size, block_size) != source_size - 12)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Decompression error");
|
|
} else {
|
|
if (bshuf_decompress_zstd(source + 12, output, nelements,
|
|
elem_size, block_size) != source_size - 12)
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Decompression error");
|
|
}
|
|
break;
|
|
default:
|
|
throw JFJochException(JFJochExceptionCategory::Compression, "Not implemented algorithm");
|
|
}
|
|
}
|
|
|
|
template <class Td, class Ts>
|
|
void JFJochDecompress(std::vector<Td> &output, CompressionAlgorithm algorithm, const Ts *source_v, size_t source_size,
|
|
size_t nelements, bool use_hperf = true) {
|
|
output.resize(nelements);
|
|
JFJochDecompressPtr((uint8_t *) output.data(), algorithm, (uint8_t *) source_v, source_size,
|
|
nelements, sizeof(Td), use_hperf);
|
|
}
|
|
|
|
template <class Td, class Ts>
|
|
void JFJochDecompress(std::vector<Td> &output, CompressionAlgorithm algorithm, const std::vector<Ts> source_v,
|
|
size_t nelements, bool use_hperf = true) {
|
|
JFJochDecompress(output, algorithm, source_v.data(), source_v.size() * sizeof(Ts), nelements, use_hperf);
|
|
}
|