There are methods like getThreadGroup() and ThreadGroupSwitcher class,
so seems that this is logical.
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
There are lots of thread pools and simple local-vs-global is not enough
already, it is good to know which one in particular uses threads.
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
* save format string for NetException
* format exceptions
* format exceptions 2
* format exceptions 3
* format exceptions 4
* format exceptions 5
* format exceptions 6
* fix
* format exceptions 7
* format exceptions 8
* Update MergeTreeIndexGin.cpp
* Update AggregateFunctionMap.cpp
* Update AggregateFunctionMap.cpp
* fix
* do not preallocate if max_size_to_preallocate_for_aggregation is too small
* skip optimisation for aggr without key
* increase default for max_size_to_preallocate_for_aggregation
- lots of static_cast
- add safe_cast
- types adjustments
- config
- IStorage::read/watch
- ...
- some TODO's (to convert types in future)
P.S. That was quite a journey...
v2: fixes after rebase
v3: fix conflicts after #42308 merged
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
* Uniformize disk reservation logs
* Remove log about destroying stuff that appears all the time
* More tweaks on disk reservation logs
* Reorder logs in hash join
* Remove log that provides little information
* Collapse part removal logs
Co-authored-by: Sergei Trifonov <sergei@clickhouse.com>
* impl
* stash
* clean up
* do not apply when HT is small
* make branch static
* also in merge
* do not hardcode look ahead value
* fix
* apply to methods with cheap key calculation
* more tests
* silence tidy
* fix build
* support HashMethodKeysFixed
* apply during merge only for cheap
* stash
* fixes
* rename method
* add feature flag
* cache prefetch threshold value
* fix
* fix
* Update HashMap.h
* fix typo
* 256KB as default l2 size
Co-authored-by: Alexey Milovidov <milovidov@clickhouse.com>
v2: fill AggregateColumnsConstData only for only_merge
(fixes 01291_aggregation_in_order and some other tests)
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
row_begin was wrong, and before this patch aggregator processing
{row_end, row_end} range, in other words, zero range.
Fixes: #9113 (cc @dimarub2000)
v2: add static_cast to fix UBSan
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
The check is currently *not* part of .clang-tidy. It complains about:
(1) "switch has multiple consecutive identical branches"
(2) "repeated branch in conditional chain"
About (1): Lots of findings in switches were about redundant
"[[fallthrough]]" in places where the compiler would not warn anyways. I
have cleaned these up.
About (2): In if-else_if-else chains, fixing the warning would usually
mean concatenating multiple if-conditions. As this would reduce
readability in most cases, I did not fix these places.
Because of (2), I also refrained from adding "bugprone-branch-clone" to
.clang-tidy.
Before it does lots of extra work, now, it will be significantly more
optimal (thousands of rows -> 1-2 million of rows).
v2: s/executeOnBlockSimple/executeOnBlockSmall/
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
This statistics significantly decrease performance of
optimize_aggregation_in_order with a prefix key.
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
* use AggregationMethod ctor with reserve
* add new settings
* add HashTablesStatistics
* support queries with limit
* support distributed and with external aggregation
* add new profile events
* add some tests
* add perf test
* export cache stats through AsynchronousMetrics
* rm redundant trace
* fix style
* fix 02122_parallel_formatting test
* review fixes
* fix 02122_parallel_formatting test
* apply also to two-level HTs
* try simpler strategy
* increase max_size_to_preallocate_for_aggregation for experiment
* fixes
* Revert "increase max_size_to_preallocate_for_aggregation for experiment"
This reverts commit 6cf6f75704.
* fix test
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Reproducer:
# NOTE: we need clickhouse from 33957 since right now LSan is broken due to getauxval().
$ url=https://s3.amazonaws.com/clickhouse-builds/33957/e04b862673644d313712607a0078f5d1c48b5377/package_asan/clickhouse
$ wget $url -o clickhouse-asan
$ chmod +x clickhouse-asan
$ ./clickhouse-asan server &
$ ./clickhouse-asan client
:) create table data (key Int, value String) engine=MergeTree() order by key
:) insert into data select number%5, toString(number) from numbers(10e6)
# usually it is enough one query, benchmark is just for stability of the results
# note, that if the exception was not happen from AggregatingInOrderTransform then add --continue_on_errors and wait
$ ./clickhouse-asan benchmark --query 'select key, uniqCombined64(value), groupArray(value) from data group by key' --optimize_aggregation_in_order=1 --memory_tracker_fault_probability=0.01, max_untracked_memory='2Mi'
LSan report:
==24595==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 3932160 byte(s) in 6 object(s) allocated from:
0 0xcadba93 in realloc ()
1 0xcc108d9 in Allocator<false, false>::realloc() obj-x86_64-linux-gnu/../src/Common/Allocator.h:134:30
2 0xde19eae in void DB::PODArrayBase<>::realloc<DB::Arena*&>(unsigned long, DB::Arena*&) obj-x86_64-linux-gnu/../src/Common/PODArray.h:161:25
3 0xde5f039 in void DB::PODArrayBase<>::reserveForNextSize<DB::Arena*&>(DB::Arena*&) obj-x86_64-linux-gnu/../src/Common/PODArray.h
4 0xde5f039 in void DB::PODArray<>::push_back<>(DB::GroupArrayNodeString*&, DB::Arena*&) obj-x86_64-linux-gnu/../src/Common/PODArray.h:432:19
5 0xde5f039 in DB::GroupArrayGeneralImpl<>::add() const obj-x86_64-linux-gnu/../src/AggregateFunctions/AggregateFunctionGroupArray.h:465:31
6 0xde5f039 in DB::IAggregateFunctionHelper<>::addBatchSinglePlaceFromInterval() const obj-x86_64-linux-gnu/../src/AggregateFunctions/IAggregateFunction.h:481:53
7 0x299df134 in DB::Aggregator::executeOnIntervalWithoutKeyImpl() obj-x86_64-linux-gnu/../src/Interpreters/Aggregator.cpp:869:31
8 0x2ca75f7d in DB::AggregatingInOrderTransform::consume() obj-x86_64-linux-gnu/../src/Processors/Transforms/AggregatingInOrderTransform.cpp:124:13
...
SUMMARY: AddressSanitizer: 4523184 byte(s) leaked in 12 allocation(s).
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
1. add LRUResourceCache for managing resource cache in lru policy
2. rollback LRUCache to the original version
3. add remove() in LRUCache
4. add unit tests for LRUResourceCache and LRUCache
