[Opt](compression) Opt gzip decompress by libdeflate on X86 and X86_6…

…4 platforms.

be/cmake/thirdparty.cmake

@@ -161,3 +161,8 @@ if (OS_MACOSX)
     add_thirdparty(iberty)
     add_thirdparty(intl)
 endif()
+
+# Only used on x86 or x86_64
+if ("${CMAKE_BUILD_TARGET_ARCH}" STREQUAL "x86" OR "${CMAKE_BUILD_TARGET_ARCH}" STREQUAL "x86_64")
+    add_thirdparty(deflate)
+endif()

be/src/util/block_compression.cpp

@@ -20,6 +20,11 @@
 #include <gen_cpp/parquet_types.h>
 #include <gen_cpp/segment_v2.pb.h>
 #include <glog/logging.h>
+// Only used on x86 or x86_64
+#if defined(__x86_64__) || defined(_M_X64) || defined(i386) || defined(__i386__) || \
+        defined(__i386) || defined(_M_IX86)
+#include <libdeflate.h>
+#endif
 #include <limits.h>
 #include <lz4/lz4.h>
 #include <lz4/lz4frame.h>
@@ -1006,6 +1011,39 @@ class GzipBlockCompression final : public ZlibBlockCompression {
     const static int MEM_LEVEL = 8;
 };
 
+// Only used on x86 or x86_64
+#if defined(__x86_64__) || defined(_M_X64) || defined(i386) || defined(__i386__) || \
+        defined(__i386) || defined(_M_IX86)
+class GzipBlockCompressionByLibdeflate final : public GzipBlockCompression {
+public:
+    GzipBlockCompressionByLibdeflate() : GzipBlockCompression() {}
+    static GzipBlockCompressionByLibdeflate* instance() {
+        static GzipBlockCompressionByLibdeflate s_instance;
+        return &s_instance;
+    }
+    ~GzipBlockCompressionByLibdeflate() override = default;
+
+    Status decompress(const Slice& input, Slice* output) override {
+        if (input.empty()) {
+            output->size = 0;
+            return Status::OK();
+        }
+        thread_local std::unique_ptr<libdeflate_decompressor, void (*)(libdeflate_decompressor*)>
+                decompressor {libdeflate_alloc_decompressor(), libdeflate_free_decompressor};
+        if (!decompressor) {
+            return Status::InternalError("libdeflate_alloc_decompressor error.");
+        }
+        std::size_t out_len;
+        auto result = libdeflate_gzip_decompress(decompressor.get(), input.data, input.size,
+                                                 output->data, output->size, &out_len);
+        if (result != LIBDEFLATE_SUCCESS) {
+            return Status::InternalError("libdeflate_gzip_decompress error, res={}, result);
+        }
+        return Status::OK();
+    }
+};
+#endif
+
 Status get_block_compression_codec(segment_v2::CompressionTypePB type,
                                    BlockCompressionCodec** codec) {
     switch (type) {
@@ -1054,8 +1092,13 @@ Status get_block_compression_codec(tparquet::CompressionCodec::type parquet_code
         *codec = ZstdBlockCompression::instance();
         break;
     case tparquet::CompressionCodec::GZIP:
+// Only used on x86 or x86_64
+#if defined(__x86_64__) || defined(_M_X64) || defined(i386) || defined(__i386__) || \
+        defined(__i386) || defined(_M_IX86)
+        *codec = GzipBlockCompressionByLibdeflate::instance();
+#else
         *codec = GzipBlockCompression::instance();
-        break;
+#endif
     default:
         return Status::InternalError("unknown compression type({})", parquet_codec);
     }

thirdparty/build-thirdparty.sh

@@ -1691,6 +1691,20 @@ build_avx2neon() {
     cp -r ./* "${TP_INSTALL_DIR}/include/avx2neon/"
 }
 
+# libdeflate
+build_libdeflate() {
+    check_if_source_exist "${LIBDEFLATE_SOURCE}"
+    cd "${TP_SOURCE_DIR}/${LIBDEFLATE_SOURCE}"
+
+    rm -rf "${BUILD_DIR}"
+    mkdir -p "${BUILD_DIR}"
+    cd "${BUILD_DIR}"
+
+    "${CMAKE_CMD}" -G "${GENERATOR}" -DCMAKE_INSTALL_PREFIX="${TP_INSTALL_DIR}" -DCMAKE_BUILD_TYPE=Release ..
+    "${BUILD_SYSTEM}" -j "${PARALLEL}"
+    "${BUILD_SYSTEM}" install
+}
+
 if [[ "${#packages[@]}" -eq 0 ]]; then
     packages=(
         libunixodbc
@@ -1754,6 +1768,7 @@ if [[ "${#packages[@]}" -eq 0 ]]; then
         libunwind
         dragonbox
         avx2neon
+        libdeflate
     )
     if [[ "$(uname -s)" == 'Darwin' ]]; then
         read -r -a packages <<<"binutils gettext ${packages[*]}"

thirdparty/vars.sh

@@ -478,6 +478,12 @@ AVX2NEON_NAME=v1.0.0.tar.gz
 AVX2NEON_SOURCE=AvxToNeon-1.0.0
 AVX2NEON_MD5SUM="692d0e0f8b885a86ebc5172a9d8ee8db"
 
+# libdeflate
+LIBDEFLATE_DOWNLOAD="https://github.com/ebiggers/libdeflate/archive/refs/tags/v1.19.zip"
+LIBDEFLATE_NAME="libdeflate-1.19.zip"
+LIBDEFLATE_SOURCE="libdeflate-1.19"
+LIBDEFLATE_MD5SUM="1ec42dfe7d777929ade295281560d750"
+
 # all thirdparties which need to be downloaded is set in array TP_ARCHIVES
 export TP_ARCHIVES=(
     'LIBEVENT'
@@ -548,6 +554,7 @@ export TP_ARCHIVES=(
     'HADOOP_LIBS'
     'DRAGONBOX'
     'AVX2NEON'
+    'LIBDEFLATE'
 )
 
 if [[ "$(uname -s)" == 'Darwin' ]]; then