rocksdb/CMakeLists.txt
Nathan Bronson 7d87f02799 support for concurrent adds to memtable
Summary:
This diff adds support for concurrent adds to the skiplist memtable
implementations.  Memory allocation is made thread-safe by the addition of
a spinlock, with small per-core buffers to avoid contention.  Concurrent
memtable writes are made via an additional method and don't impose a
performance overhead on the non-concurrent case, so parallelism can be
selected on a per-batch basis.

Write thread synchronization is an increasing bottleneck for higher levels
of concurrency, so this diff adds --enable_write_thread_adaptive_yield
(default off).  This feature causes threads joining a write batch
group to spin for a short time (default 100 usec) using sched_yield,
rather than going to sleep on a mutex.  If the timing of the yield calls
indicates that another thread has actually run during the yield then
spinning is avoided.  This option improves performance for concurrent
situations even without parallel adds, although it has the potential to
increase CPU usage (and the heuristic adaptation is not yet mature).

Parallel writes are not currently compatible with
inplace updates, update callbacks, or delete filtering.
Enable it with --allow_concurrent_memtable_write (and
--enable_write_thread_adaptive_yield).  Parallel memtable writes
are performance neutral when there is no actual parallelism, and in
my experiments (SSD server-class Linux and varying contention and key
sizes for fillrandom) they are always a performance win when there is
more than one thread.

Statistics are updated earlier in the write path, dropping the number
of DB mutex acquisitions from 2 to 1 for almost all cases.

This diff was motivated and inspired by Yahoo's cLSM work.  It is more
conservative than cLSM: RocksDB's write batch group leader role is
preserved (along with all of the existing flush and write throttling
logic) and concurrent writers are blocked until all memtable insertions
have completed and the sequence number has been advanced, to preserve
linearizability.

My test config is "db_bench -benchmarks=fillrandom -threads=$T
-batch_size=1 -memtablerep=skip_list -value_size=100 --num=1000000/$T
-level0_slowdown_writes_trigger=9999 -level0_stop_writes_trigger=9999
-disable_auto_compactions --max_write_buffer_number=8
-max_background_flushes=8 --disable_wal --write_buffer_size=160000000
--block_size=16384 --allow_concurrent_memtable_write" on a two-socket
Xeon E5-2660 @ 2.2Ghz with lots of memory and an SSD hard drive.  With 1
thread I get ~440Kops/sec.  Peak performance for 1 socket (numactl
-N1) is slightly more than 1Mops/sec, at 16 threads.  Peak performance
across both sockets happens at 30 threads, and is ~900Kops/sec, although
with fewer threads there is less performance loss when the system has
background work.

Test Plan:
1. concurrent stress tests for InlineSkipList and DynamicBloom
2. make clean; make check
3. make clean; DISABLE_JEMALLOC=1 make valgrind_check; valgrind db_bench
4. make clean; COMPILE_WITH_TSAN=1 make all check; db_bench
5. make clean; COMPILE_WITH_ASAN=1 make all check; db_bench
6. make clean; OPT=-DROCKSDB_LITE make check
7. verify no perf regressions when disabled

Reviewers: igor, sdong

Reviewed By: sdong

Subscribers: MarkCallaghan, IslamAbdelRahman, anthony, yhchiang, rven, sdong, guyg8, kradhakrishnan, dhruba

Differential Revision: https://reviews.facebook.net/D50589
2015-12-25 11:03:40 -08:00

451 lines
16 KiB
CMake

# This cmake build is for Windows 64-bit only.
#
# Prerequisites:
# You must have Visual Studio 2013 Update 4 installed. Start the Developer Command Prompt window that is a part of Visual Studio installation.
# Run the build commands from within the Developer Command Prompt window to have paths to the compiler and runtime libraries set.
# You must have git.exe in your %PATH% environment variable.
#
# To build Rocksdb for Windows is as easy as 1-2-3-4-5:
#
# 1. Update paths to third-party libraries in thirdparty.inc file
# 2. Create a new directory for build artifacts
# mkdir build
# cd build
# 3. Run cmake to generate project files for Windows, add more options to enable required third-party libraries.
# See thirdparty.inc for more information.
# sample command: cmake -G "Visual Studio 12 Win64" -DGFLAGS=1 -DSNAPPY=1 -DJEMALLOC=1 ..
# OR for VS Studio 15 cmake -G "Visual Studio 14 Win64" -DGFLAGS=1 -DSNAPPY=1 -DJEMALLOC=1 ..
# 4. Then build the project in debug mode (you may want to add /m[:<N>] flag to run msbuild in <N> parallel threads
# or simply /m ot use all avail cores)
# msbuild rocksdb.sln
#
# rocksdb.sln build features exclusions of test only code in Release. If you build ALL_BUILD then everything
# will be attempted but test only code does not build in Release mode.
#
# 5. And release mode (/m[:<N>] is also supported)
# msbuild rocksdb.sln /p:Configuration=Release
#
cmake_minimum_required(VERSION 2.6)
project(rocksdb)
include(${CMAKE_CURRENT_SOURCE_DIR}/thirdparty.inc)
execute_process(COMMAND $ENV{COMSPEC} " /C date /T" OUTPUT_VARIABLE DATE)
execute_process(COMMAND $ENV{COMSPEC} " /C time /T" OUTPUT_VARIABLE TIME)
string(REGEX REPLACE "(..)/(..)/..(..).*" "\\1/\\2/\\3" DATE ${DATE})
string(REGEX REPLACE "(..):(.....).*" " \\1:\\2" TIME ${TIME})
string(CONCAT GIT_DATE_TIME ${DATE} ${TIME})
find_package(Git)
if (GIT_FOUND AND EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/.git")
execute_process(COMMAND $ENV{COMSPEC} /C ${GIT_EXECUTABLE} -C ${CMAKE_CURRENT_SOURCE_DIR} rev-parse HEAD OUTPUT_VARIABLE GIT_SHA)
else()
set(GIT_SHA 0)
endif()
string(REGEX REPLACE "[^0-9a-f]+" "" GIT_SHA ${GIT_SHA})
set(BUILD_VERSION_CC ${CMAKE_CURRENT_SOURCE_DIR}/util/build_version.cc)
add_custom_command(OUTPUT ${BUILD_VERSION_CC}
COMMAND echo "#include \"build_version.h\"" > ${BUILD_VERSION_CC}
COMMAND echo "const char* rocksdb_build_git_sha = \"rocksdb_build_git_sha:${GIT_SHA}\";" >> ${BUILD_VERSION_CC}
COMMAND echo "const char* rocksdb_build_git_datetime = \"rocksdb_build_git_datetime:${GIT_DATE_TIME}\";" >> ${BUILD_VERSION_CC}
COMMAND echo const char* rocksdb_build_compile_date = __DATE__\; >> ${BUILD_VERSION_CC}
)
add_custom_target(GenerateBuildVersion DEPENDS ${BUILD_VERSION_CC})
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /Zi /nologo /EHsc /GS /Gd /GR /GF /fp:precise /Zc:wchar_t /Zc:forScope /errorReport:queue")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /FC /d2Zi+ /W3 /WX /wd4127 /wd4800 /wd4996")
# Used to run CI build and tests so we can run faster
set(OPTIMIZE_DEBUG_DEFAULT 0) # Debug build is unoptimized by default use -DOPTDBG=1 to optimize
if(DEFINED OPTDBG)
set(OPTIMIZE_DEBUG ${OPTDBG})
else()
set(OPTIMIZE_DEBUG ${OPTIMIZE_DEBUG_DEFAULT})
endif()
if((${OPTIMIZE_DEBUG} EQUAL 1))
message(STATUS "Debug optimization is enabled")
set(CMAKE_CXX_FLAGS_DEBUG "/Oxt /MDd")
else()
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Od /RTC1 /Gm /MDd")
endif()
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Oxt /Zp8 /Gm- /Gy /MD")
set(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} /DEBUG")
set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} /DEBUG")
add_definitions(-DWIN32 -DOS_WIN -D_MBCS -DWIN64)
include_directories(${PROJECT_SOURCE_DIR})
include_directories(${PROJECT_SOURCE_DIR}/include)
include_directories(${PROJECT_SOURCE_DIR}/third-party/gtest-1.7.0/fused-src)
set(ROCKSDB_LIBS rocksdblib${ARTIFACT_SUFFIX})
set(THIRDPARTY_LIBS ${THIRDPARTY_LIBS} gtest)
set(SYSTEM_LIBS Shlwapi.lib Rpcrt4.lib)
set(LIBS ${ROCKSDB_LIBS} ${THIRDPARTY_LIBS} ${SYSTEM_LIBS})
add_subdirectory(third-party/gtest-1.7.0/fused-src/gtest)
# Main library source code
set(SOURCES
db/builder.cc
db/c.cc
db/column_family.cc
db/compacted_db_impl.cc
db/compaction.cc
db/compaction_iterator.cc
db/compaction_job.cc
db/compaction_picker.cc
db/convenience.cc
db/dbformat.cc
db/db_filesnapshot.cc
db/db_impl.cc
db/db_impl_debug.cc
db/db_impl_experimental.cc
db/db_impl_readonly.cc
db/db_iter.cc
db/event_helpers.cc
db/experimental.cc
db/filename.cc
db/file_indexer.cc
db/flush_job.cc
db/flush_scheduler.cc
db/forward_iterator.cc
db/internal_stats.cc
db/log_reader.cc
db/log_writer.cc
db/managed_iterator.cc
db/memtable.cc
db/memtable_allocator.cc
db/memtable_list.cc
db/merge_helper.cc
db/merge_operator.cc
db/repair.cc
db/slice.cc
db/snapshot_impl.cc
db/table_cache.cc
db/table_properties_collector.cc
db/transaction_log_impl.cc
db/version_builder.cc
db/version_edit.cc
db/version_set.cc
db/wal_manager.cc
db/write_batch.cc
db/write_batch_base.cc
db/write_controller.cc
db/write_thread.cc
memtable/hash_cuckoo_rep.cc
memtable/hash_linklist_rep.cc
memtable/hash_skiplist_rep.cc
port/stack_trace.cc
port/win/env_win.cc
port/win/port_win.cc
port/win/win_logger.cc
table/adaptive_table_factory.cc
table/block.cc
table/block_based_filter_block.cc
table/block_based_table_builder.cc
table/block_based_table_factory.cc
table/block_based_table_reader.cc
table/block_builder.cc
table/block_hash_index.cc
table/block_prefix_index.cc
table/bloom_block.cc
table/cuckoo_table_builder.cc
table/cuckoo_table_factory.cc
table/cuckoo_table_reader.cc
table/flush_block_policy.cc
table/format.cc
table/full_filter_block.cc
table/get_context.cc
table/iterator.cc
table/merger.cc
table/sst_file_writer.cc
table/meta_blocks.cc
table/mock_table.cc
table/plain_table_builder.cc
table/plain_table_factory.cc
table/plain_table_index.cc
table/plain_table_key_coding.cc
table/plain_table_reader.cc
table/table_properties.cc
table/two_level_iterator.cc
tools/sst_dump_tool.cc
tools/dump/db_dump_tool.cc
util/arena.cc
util/auto_roll_logger.cc
util/bloom.cc
util/build_version.cc
util/cache.cc
util/coding.cc
util/compaction_job_stats_impl.cc
util/comparator.cc
util/concurrent_arena.cc
util/crc32c.cc
util/db_info_dumper.cc
util/delete_scheduler_impl.cc
util/dynamic_bloom.cc
util/env.cc
util/env_hdfs.cc
util/event_logger.cc
util/file_util.cc
util/file_reader_writer.cc
util/filter_policy.cc
util/hash.cc
util/histogram.cc
util/instrumented_mutex.cc
util/iostats_context.cc
tools/ldb_cmd.cc
tools/ldb_tool.cc
util/logging.cc
util/log_buffer.cc
util/memenv.cc
util/mock_env.cc
util/murmurhash.cc
util/mutable_cf_options.cc
util/options.cc
util/options_builder.cc
util/options_helper.cc
util/options_parser.cc
util/options_sanity_check.cc
util/perf_context.cc
util/perf_level.cc
util/random.cc
util/rate_limiter.cc
util/skiplistrep.cc
util/slice.cc
util/statistics.cc
util/status.cc
util/status_message.cc
util/string_util.cc
util/sync_point.cc
util/testharness.cc
util/testutil.cc
util/thread_local.cc
util/thread_status_impl.cc
util/thread_status_updater.cc
util/thread_status_util.cc
util/thread_status_util_debug.cc
util/vectorrep.cc
util/xfunc.cc
util/xxhash.cc
utilities/backupable/backupable_db.cc
utilities/checkpoint/checkpoint.cc
utilities/document/document_db.cc
utilities/document/json_document.cc
utilities/document/json_document_builder.cc
utilities/env_mirror.cc
utilities/flashcache/flashcache.cc
utilities/geodb/geodb_impl.cc
utilities/leveldb_options/leveldb_options.cc
utilities/memory/memory_util.cc
utilities/merge_operators/string_append/stringappend.cc
utilities/merge_operators/string_append/stringappend2.cc
utilities/merge_operators/put.cc
utilities/merge_operators/uint64add.cc
utilities/options/options_util.cc
utilities/redis/redis_lists.cc
utilities/spatialdb/spatial_db.cc
utilities/table_properties_collectors/compact_on_deletion_collector.cc
utilities/transactions/optimistic_transaction_impl.cc
utilities/transactions/optimistic_transaction_db_impl.cc
utilities/transactions/transaction_base.cc
utilities/transactions/transaction_impl.cc
utilities/transactions/transaction_db_impl.cc
utilities/transactions/transaction_db_mutex_impl.cc
utilities/transactions/transaction_lock_mgr.cc
utilities/transactions/transaction_util.cc
utilities/ttl/db_ttl_impl.cc
utilities/write_batch_with_index/write_batch_with_index.cc
utilities/write_batch_with_index/write_batch_with_index_internal.cc
)
# For test util library that is build only in DEBUG mode
# and linked to tests. Add test only code that is not #ifdefed for Release here.
set(TESTUTIL_SOURCE
db/db_test_util.cc
util/thread_status_updater_debug.cc
)
add_library(rocksdblib${ARTIFACT_SUFFIX} ${SOURCES})
set_target_properties(rocksdblib${ARTIFACT_SUFFIX} PROPERTIES COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/rocksdblib${ARTIFACT_SUFFIX}.pdb")
add_dependencies(rocksdblib${ARTIFACT_SUFFIX} GenerateBuildVersion)
add_library(rocksdb${ARTIFACT_SUFFIX} SHARED ${SOURCES})
set_target_properties(rocksdb${ARTIFACT_SUFFIX} PROPERTIES COMPILE_FLAGS "-DROCKSDB_DLL -DROCKSDB_LIBRARY_EXPORTS /Fd${CMAKE_CFG_INTDIR}/rocksdb${ARTIFACT_SUFFIX}.pdb")
add_dependencies(rocksdb${ARTIFACT_SUFFIX} GenerateBuildVersion)
target_link_libraries(rocksdb${ARTIFACT_SUFFIX} ${LIBS})
set(APPS
db/db_bench.cc
db/memtablerep_bench.cc
table/table_reader_bench.cc
tools/db_stress.cc
tools/write_stress.cc
tools/db_repl_stress.cc
tools/sst_dump.cc
tools/dump/rocksdb_dump.cc
tools/dump/rocksdb_undump.cc
util/cache_bench.cc
)
set(C_TESTS db/c_test.c)
set(TESTS
db/column_family_test.cc
db/compact_files_test.cc
db/compaction_iterator_test.cc
db/compaction_job_test.cc
db/compaction_job_stats_test.cc
db/compaction_picker_test.cc
db/comparator_db_test.cc
db/corruption_test.cc
db/cuckoo_table_db_test.cc
db/db_iter_test.cc
db/db_test.cc
db/db_compaction_filter_test.cc
db/db_compaction_test.cc
db/db_dynamic_level_test.cc
db/db_inplace_update_test.cc
db/db_log_iter_test.cc
db/db_universal_compaction_test.cc
db/db_wal_test.cc
db/db_tailing_iter_test.cc
db/db_table_properties_test.cc
db/dbformat_test.cc
db/deletefile_test.cc
db/fault_injection_test.cc
db/file_indexer_test.cc
db/filename_test.cc
db/flush_job_test.cc
db/inlineskiplist_test.cc
db/listener_test.cc
db/log_test.cc
db/manual_compaction_test.cc
db/memtable_list_test.cc
db/merge_test.cc
db/merge_helper_test.cc
db/options_file_test.cc
db/perf_context_test.cc
db/plain_table_db_test.cc
db/prefix_test.cc
db/skiplist_test.cc
db/table_properties_collector_test.cc
db/version_builder_test.cc
db/version_edit_test.cc
db/version_set_test.cc
db/wal_manager_test.cc
db/write_batch_test.cc
db/write_callback_test.cc
db/write_controller_test.cc
table/block_based_filter_block_test.cc
table/block_hash_index_test.cc
table/block_test.cc
table/cuckoo_table_builder_test.cc
table/cuckoo_table_reader_test.cc
table/full_filter_block_test.cc
table/merger_test.cc
table/table_test.cc
tools/db_sanity_test.cc
tools/ldb_cmd_test.cc
tools/reduce_levels_test.cc
tools/sst_dump_test.cc
util/arena_test.cc
util/autovector_test.cc
util/auto_roll_logger_test.cc
util/bloom_test.cc
util/cache_test.cc
util/coding_test.cc
util/crc32c_test.cc
util/delete_scheduler_test.cc
util/dynamic_bloom_test.cc
util/env_test.cc
util/event_logger_test.cc
util/filelock_test.cc
util/file_reader_writer_test.cc
util/heap_test.cc
util/histogram_test.cc
util/memenv_test.cc
util/mock_env_test.cc
util/options_test.cc
util/rate_limiter_test.cc
util/slice_transform_test.cc
util/thread_list_test.cc
util/thread_local_test.cc
utilities/backupable/backupable_db_test.cc
utilities/checkpoint/checkpoint_test.cc
utilities/document/document_db_test.cc
utilities/document/json_document_test.cc
utilities/geodb/geodb_test.cc
utilities/memory/memory_test.cc
utilities/merge_operators/string_append/stringappend_test.cc
utilities/options/options_util_test.cc
utilities/redis/redis_lists_test.cc
utilities/spatialdb/spatial_db_test.cc
utilities/table_properties_collectors/compact_on_deletion_collector_test.cc
utilities/transactions/optimistic_transaction_test.cc
utilities/transactions/transaction_test.cc
utilities/ttl/ttl_test.cc
utilities/write_batch_with_index/write_batch_with_index_test.cc
)
set(EXES ${APPS})
foreach(sourcefile ${EXES})
string(REPLACE ".cc" "" exename ${sourcefile})
string(REGEX REPLACE "^((.+)/)+" "" exename ${exename})
add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile})
target_link_libraries(${exename}${ARTIFACT_SUFFIX} ${LIBS})
endforeach(sourcefile ${EXES})
# test utilities are only build in debug
set(TESTUTILLIB testutillib${ARTIFACT_SUFFIX})
add_library(${TESTUTILLIB} STATIC ${TESTUTIL_SOURCE})
set_target_properties(${TESTUTILLIB} PROPERTIES COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/testutillib${ARTIFACT_SUFFIX}.pdb")
set_target_properties(${TESTUTILLIB}
PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
)
# Tests are excluded from Release builds
set(TEST_EXES ${TESTS})
foreach(sourcefile ${TEST_EXES})
string(REPLACE ".cc" "" exename ${sourcefile})
string(REGEX REPLACE "^((.+)/)+" "" exename ${exename})
add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile})
set_target_properties(${exename}${ARTIFACT_SUFFIX}
PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
)
target_link_libraries(${exename}${ARTIFACT_SUFFIX} ${LIBS} testutillib${ARTIFACT_SUFFIX})
endforeach(sourcefile ${TEST_EXES})
# C executables must link to a shared object
set(C_TEST_EXES ${C_TESTS})
foreach(sourcefile ${C_TEST_EXES})
string(REPLACE ".c" "" exename ${sourcefile})
string(REGEX REPLACE "^((.+)/)+" "" exename ${exename})
add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile})
set_target_properties(${exename}${ARTIFACT_SUFFIX}
PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
)
target_link_libraries(${exename}${ARTIFACT_SUFFIX} rocksdb${ARTIFACT_SUFFIX} testutillib${ARTIFACT_SUFFIX})
endforeach(sourcefile ${C_TEST_EXES})