dolysis/rocksdb
3
0
Fork 0
mirror of https://github.com/facebook/rocksdb.git synced 2024-11-26 07:30:54 +00:00
Code Issues Packages Projects Releases Wiki Activity
rocksdb/java/rocksjni
History
Changyu Bi cc23b46da1 Support using ZDICT_finalizeDictionary to generate zstd dictionary (#9857) 
Summary:
An untrained dictionary is currently simply the concatenation of several samples. The ZSTD API, ZDICT_finalizeDictionary(), can improve such a dictionary's effectiveness at low cost. This PR changes how dictionary is created by calling the ZSTD ZDICT_finalizeDictionary() API instead of creating raw content dictionary (when max_dict_buffer_bytes > 0), and pass in all buffered uncompressed data blocks as samples.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/9857

Test Plan:
#### db_bench test for cpu/memory of compression+decompression and space saving on synthetic data:
Set up: change the parameter [here](fb9a167a55/tools/db_bench_tool.cc (L1766)) to 16384 to make synthetic data more compressible.
```
# linked local ZSTD with version 1.5.2
# DEBUG_LEVEL=0 ROCKSDB_NO_FBCODE=1 ROCKSDB_DISABLE_ZSTD=1  EXTRA_CXXFLAGS="-DZSTD_STATIC_LINKING_ONLY -DZSTD -I/data/users/changyubi/install/include/" EXTRA_LDFLAGS="-L/data/users/changyubi/install/lib/ -l:libzstd.a" make -j32 db_bench

dict_bytes=16384
train_bytes=1048576
echo "========== No Dictionary =========="
TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=0 -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1
TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=0 -block_size=4096 2>&1 | grep elapsed
du -hc /dev/shm/dbbench/*sst | grep total

echo "========== Raw Content Dictionary =========="
TEST_TMPDIR=/dev/shm ./db_bench_main -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1
TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench_main -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -block_size=4096 2>&1 | grep elapsed
du -hc /dev/shm/dbbench/*sst | grep total

echo "========== FinalizeDictionary =========="
TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1
TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false -block_size=4096 2>&1 | grep elapsed
du -hc /dev/shm/dbbench/*sst | grep total

echo "========== TrainDictionary =========="
TEST_TMPDIR=/dev/shm ./db_bench -benchmarks=filluniquerandom,compact -num=10000000 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -block_size=4096 -max_background_jobs=24 -memtablerep=vector -allow_concurrent_memtable_write=false -disable_wal=true -max_write_buffer_number=8 >/dev/null 2>&1
TEST_TMPDIR=/dev/shm /usr/bin/time ./db_bench -use_existing_db=true -benchmarks=compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -block_size=4096 2>&1 | grep elapsed
du -hc /dev/shm/dbbench/*sst | grep total

# Result: TrainDictionary is much better on space saving, but FinalizeDictionary seems to use less memory.
# before compression data size: 1.2GB
dict_bytes=16384
max_dict_buffer_bytes =  1048576
                    space   cpu/memory
No Dictionary       468M    14.93user 1.00system 0:15.92elapsed 100%CPU (0avgtext+0avgdata 23904maxresident)k
Raw Dictionary      251M    15.81user 0.80system 0:16.56elapsed 100%CPU (0avgtext+0avgdata 156808maxresident)k
FinalizeDictionary  236M    11.93user 0.64system 0:12.56elapsed 100%CPU (0avgtext+0avgdata 89548maxresident)k
TrainDictionary     84M     7.29user 0.45system 0:07.75elapsed 100%CPU (0avgtext+0avgdata 97288maxresident)k
```

#### Benchmark on 10 sample SST files for spacing saving and CPU time on compression:
FinalizeDictionary is comparable to TrainDictionary in terms of space saving, and takes less time in compression.
```
dict_bytes=16384
train_bytes=1048576

for sst_file in `ls ../temp/myrock-sst/`
do
  echo "********** $sst_file **********"
  echo "========== No Dictionary =========="
  ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD

  echo "========== Raw Content Dictionary =========="
  ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD --compression_max_dict_bytes=$dict_bytes

  echo "========== FinalizeDictionary =========="
  ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD --compression_max_dict_bytes=$dict_bytes --compression_zstd_max_train_bytes=$train_bytes --compression_use_zstd_finalize_dict

  echo "========== TrainDictionary =========="
  ./sst_dump --file="../temp/myrock-sst/$sst_file" --command=recompress --compression_level_from=6 --compression_level_to=6 --compression_types=kZSTD --compression_max_dict_bytes=$dict_bytes --compression_zstd_max_train_bytes=$train_bytes
done

                         010240.sst (Size/Time) 011029.sst              013184.sst              021552.sst              185054.sst              185137.sst              191666.sst              7560381.sst             7604174.sst             7635312.sst
No Dictionary           28165569 / 2614419      32899411 / 2976832      32977848 / 3055542      31966329 / 2004590      33614351 / 1755877      33429029 / 1717042      33611933 / 1776936      33634045 / 2771417      33789721 / 2205414      33592194 / 388254
Raw Content Dictionary  28019950 / 2697961      33748665 / 3572422      33896373 / 3534701      26418431 / 2259658      28560825 / 1839168      28455030 / 1846039      28494319 / 1861349      32391599 / 3095649      33772142 / 2407843      33592230 / 474523
FinalizeDictionary      27896012 / 2650029      33763886 / 3719427      33904283 / 3552793      26008225 / 2198033      28111872 / 1869530      28014374 / 1789771      28047706 / 1848300      32296254 / 3204027      33698698 / 2381468      33592344 / 517433
TrainDictionary         28046089 / 2740037      33706480 / 3679019      33885741 / 3629351      25087123 / 2204558      27194353 / 1970207      27234229 / 1896811      27166710 / 1903119      32011041 / 3322315      32730692 / 2406146      33608631 / 570593
```

#### Decompression/Read test:
With FinalizeDictionary/TrainDictionary, some data structure used for decompression are in stored in dictionary, so they are expected to be faster in terms of decompression/reads.
```
dict_bytes=16384
train_bytes=1048576
echo "No Dictionary"
TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=0 > /dev/null 2>&1
TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=0 2>&1 | grep MB/s

echo "Raw Dictionary"
TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes > /dev/null 2>&1
TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd  -compression_max_dict_bytes=$dict_bytes 2>&1 | grep MB/s

echo "FinalizeDict"
TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false  > /dev/null 2>&1
TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes -compression_use_zstd_dict_trainer=false 2>&1 | grep MB/s

echo "Train Dictionary"
TEST_TMPDIR=/dev/shm/ ./db_bench -benchmarks=filluniquerandom,compact -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes > /dev/null 2>&1
TEST_TMPDIR=/dev/shm/ ./db_bench -use_existing_db=true -benchmarks=readrandom -cache_size=0 -compression_type=zstd -compression_max_dict_bytes=$dict_bytes -compression_zstd_max_train_bytes=$train_bytes 2>&1 | grep MB/s

No Dictionary
readrandom   :      12.183 micros/op 82082 ops/sec 12.183 seconds 1000000 operations;    9.1 MB/s (1000000 of 1000000 found)
Raw Dictionary
readrandom   :      12.314 micros/op 81205 ops/sec 12.314 seconds 1000000 operations;    9.0 MB/s (1000000 of 1000000 found)
FinalizeDict
readrandom   :       9.787 micros/op 102180 ops/sec 9.787 seconds 1000000 operations;   11.3 MB/s (1000000 of 1000000 found)
Train Dictionary
readrandom   :       9.698 micros/op 103108 ops/sec 9.699 seconds 1000000 operations;   11.4 MB/s (1000000 of 1000000 found)
```

Reviewed By: ajkr

Differential Revision: D35720026

Pulled By: cbi42

fbshipit-source-id: 24d230fdff0fd28a1bb650658798f00dfcfb2a1f
2022-05-20 12:09:09 -07:00
..
backup_engine_options.cc Fix remaining uses of "backupable" (#9792) 2022-04-05 09:52:33 -07:00
backupenginejni.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
cache.cc support getUsage and getPinnedUsage in JavaAPI for Cache (#7925) 2021-03-17 09:30:33 -07:00
cassandra_compactionfilterjni.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
cassandra_value_operator.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
checkpoint.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
clock_cache.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
columnfamilyhandle.cc Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
compact_range_options.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
compaction_filter.cc Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
compaction_filter_factory.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
compaction_filter_factory_jnicallback.cc Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
compaction_filter_factory_jnicallback.h Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
compaction_job_info.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
compaction_job_stats.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
compaction_options.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
compaction_options_fifo.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
compaction_options_universal.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
comparator.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
comparatorjnicallback.cc Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
comparatorjnicallback.h Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
compression_options.cc Support using ZDICT_finalizeDictionary to generate zstd dictionary (#9857) 2022-05-20 12:09:09 -07:00
concurrent_task_limiter.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
config_options.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
cplusplus_to_java_convert.h Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
env.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
env_options.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
event_listener.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
event_listener_jnicallback.cc Add (& fix) some simple source code checks (#8821) 2021-09-07 21:19:27 -07:00
event_listener_jnicallback.h Add (& fix) some simple source code checks (#8821) 2021-09-07 21:19:27 -07:00
filter.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
ingest_external_file_options.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
iterator.cc Extend Java RocksDB iterators to support indirect Byte Buffers (#9222) 2022-03-24 12:50:38 -07:00
jnicallback.cc Fix tabs and lint-ignores (#6734) 2020-04-20 11:39:31 -07:00
jnicallback.h Add event listeners to RocksJava (#7425) 2020-10-14 11:33:52 -07:00
loggerjnicallback.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
loggerjnicallback.h Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
lru_cache.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
memory_util.cc Small JNI improvements (#7371) 2020-10-14 22:23:56 -07:00
memtablejni.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
merge_operator.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
native_comparator_wrapper_test.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
optimistic_transaction_db.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
optimistic_transaction_options.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
options.cc NPE in Java_org_rocksdb_ColumnFamilyOptions_setSstPartitionerFactory (#9622) 2022-03-14 14:12:30 -07:00
options_util.cc Add a ConfigOptions for use in comparing objects and converting to/from strings (#6389) 2020-04-21 17:38:17 -07:00
persistent_cache.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
portal.h Multi file concurrency in MultiGet using coroutines and async IO (#9968) 2022-05-19 15:36:27 -07:00
ratelimiterjni.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
remove_emptyvalue_compactionfilterjni.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
restorejni.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
rocks_callback_object.cc Fix tabs and lint-ignores (#6734) 2020-04-20 11:39:31 -07:00
rocksdb_exception_test.cc Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
rocksjni.cc jni: uniformly use GetByteArrayRegion() to copy bytes (#9380) 2022-03-25 10:24:58 -07:00
slice.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
snapshot.cc Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
sst_file_manager.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
sst_file_reader_iterator.cc Extend Java RocksDB iterators to support indirect Byte Buffers (#9222) 2022-03-24 12:50:38 -07:00
sst_file_readerjni.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
sst_file_writerjni.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
sst_partitioner.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
statistics.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
statisticsjni.cc Fix tabs and lint-ignores (#6734) 2020-04-20 11:39:31 -07:00
statisticsjni.h Fix tabs and lint-ignores (#6734) 2020-04-20 11:39:31 -07:00
table.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
table_filter.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
table_filter_jnicallback.cc Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
table_filter_jnicallback.h Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
testable_event_listener.cc Update build files for java8 build (#9541) 2022-02-17 13:29:21 -08:00
thread_status.cc Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
trace_writer.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
trace_writer_jnicallback.cc Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
trace_writer_jnicallback.h Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
transaction.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
transaction_db.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
transaction_db_options.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
transaction_log.cc Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
transaction_notifier.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
transaction_notifier_jnicallback.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
transaction_notifier_jnicallback.h Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
transaction_options.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
ttl.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
wal_filter.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
wal_filter_jnicallback.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
wal_filter_jnicallback.h Replace namespace name "rocksdb" with ROCKSDB_NAMESPACE (#6433) 2020-02-20 12:09:57 -08:00
write_batch.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
write_batch_test.cc Remove own ToString() (#9955) 2022-05-06 13:03:58 -07:00
write_batch_with_index.cc jni: uniformly use GetByteArrayRegion() to copy bytes (#9380) 2022-03-25 10:24:58 -07:00
write_buffer_manager.cc Fix pointer to jlong conversion in 32 bit OS (#9396) 2022-03-01 09:02:15 -08:00
writebatchhandlerjnicallback.cc Add commit marker with timestamp (#9266) 2021-12-10 11:05:35 -08:00
writebatchhandlerjnicallback.h Add commit marker with timestamp (#9266) 2021-12-10 11:05:35 -08:00