Fix bad performance in debug mode

Summary:
See github issue 574: https://github.com/facebook/rocksdb/issues/574

Basically when we're running in DEBUG mode we're calling `usleep(0)` on
every mutex lock. I bisected the issue to
https://reviews.facebook.net/D36963. Instead of calling sleep(0), this
diff just avoids calling SleepForMicroseconds() when delay is not set.

Test Plan:
    bpl=10485760;overlap=10;mcz=2;del=300000000;levels=2;ctrig=10000000; delay=10000000; stop=10000000; wbn=30; mbc=20; mb=1073741824;wbs=268435456; dds=1; sync=0; r=100000; t=1; vs=800; bs=65536; cs=1048576; of=500000; si=1000000; ./db_bench --benchmarks=fillrandom --disable_seek_compaction=1 --mmap_read=0 --statistics=1 --histogram=1 --num=$r --threads=$t --value_size=$vs --block_size=$bs --cache_size=$cs --bloom_bits=10 --cache_numshardbits=4 --open_files=$of --verify_checksum=1 --db=/tmp/rdb10test --sync=$sync --disable_wal=1 --compression_type=snappy --stats_interval=$si --compression_ratio=0.5 --disable_data_sync=$dds --write_buffer_size=$wbs --target_file_size_base=$mb --max_write_buffer_number=$wbn --max_background_compactions=$mbc --level0_file_num_compaction_trigger=$ctrig --level0_slowdown_writes_trigger=$delay --level0_stop_writes_trigger=$stop --num_levels=$levels --delete_obsolete_files_period_micros=$del --min_level_to_compress=$mcz --max_grandparent_overlap_factor=$overlap --stats_per_interval=1 --max_bytes_for_level_base=$bpl --memtablerep=vector --use_existing_db=0 --disable_auto_compactions=1 --source_compaction_factor=10000000 | grep ops

Before:
fillrandom   :     117.525 micros/op 8508 ops/sec;    6.6 MB/s
After:
fillrandom   :       1.283 micros/op 779502 ops/sec;  606.6 MB/s

Reviewers: rven, yhchiang, sdong

Reviewed By: sdong

Subscribers: meyering, dhruba, leveldb

Differential Revision: https://reviews.facebook.net/D36963

diff --git a/util/thread_status_util_debug.cc b/util/thread_status_util_debug.cc
index b91bc4ee337675c5a024490dab47434b1430f900..94b19f3d2163ed0cbd0b491ab00caefce451cf27 100644 (file)
--- a/util/thread_status_util_debug.cc
+++ b/util/thread_status_util_debug.cc
@@ -20,10 +20,11 @@ void ThreadStatusUtil::TEST_SetStateDelay(
   states_delay[state].store(micro, std::memory_order_relaxed);
 }
 
-void ThreadStatusUtil::TEST_StateDelay(
-    const ThreadStatus::StateType state) {
-  Env::Default()->SleepForMicroseconds(
-      states_delay[state].load(std::memory_order_relaxed));
+void ThreadStatusUtil::TEST_StateDelay(const ThreadStatus::StateType state) {
+  auto delay = states_delay[state].load(std::memory_order_relaxed);
+  if (delay > 0) {
+    Env::Default()->SleepForMicroseconds(delay);
+  }
 }
 
 #endif  // !NDEBUG