Split `ColumnOp` into one with row indices and one with `FieldName` & other enabled changes #1207

Centril · 2024-05-06T13:13:10Z

Description of Changes

The main goal of this PR

This PR has some main goals, while other changes are less important and are done as additional work (see commit descriptions). I recommend reviewing this PR commit by commit after having reviewed #1166 upon which this PR is based.

In this PR, we want to:

Let build_query and query execution work on a ColumnOp that stores ColId rather than FieldName (23e00d2). The previous version of ColumnOp is now called FieldOp and uses FieldName as before. This type is used by SQL compilation and query planning and is eventually compiled down to a ColumnOp. This means that Header is no longer used by ColumnOp nor build_query.
Making build_query itself (20340c8) and query execution infallible (3415702 and 986d80e). To make this possible, we must ensure that we only get bools where anything else does not make sense (977b5bb).
Split IndexSemiJoin into left and right versions (e0c12ce and 3968f4f). This enables us to make both versions less branchy and to significantly shrink the size of the right case (312 bytes vs 64).

Benchmarks

Benchmarks relative to based master on i7-7700K, 64GB RAM:

Benchmarking full-join: Collecting 100 samples in estimated 5.
full-join               time:   [270.49 µs 271.55 µs 273.56 µs]
                        change: [-11.546% -9.6360% -8.2237%] (p = 0.00 < 0.05)
                        Performance has improved.

Benchmarking incr-select: Collecting 100 samples in estimated 
incr-select             time:   [151.49 ns 152.82 ns 154.70 ns]
                        change: [-22.532% -22.144% -21.611%] (p = 0.00 < 0.05)
                        Performance has improved.

Benchmarking incr-join: Collecting 100 samples in estimated 5.
incr-join               time:   [684.71 ns 687.90 ns 692.34 ns]
                        change: [-16.301% -15.580% -14.816%] (p = 0.00 < 0.05)
                        Performance has improved.

Benchmarking query-indexes-multi: Collecting 100 samples in es
query-indexes-multi     time:   [591.14 ns 593.30 ns 596.36 ns]
                        change: [-10.874% -8.5941% -6.0559%] (p = 0.00 < 0.05)
                        Performance has improved.

API and ABI breaking changes

None

Expected complexity level and risk

2, large diff in terms of query planning and execution, but each change isn't as huge.

Testing

Existing tests have been tweaked to fit the changes.

cloutiertyler · 2024-05-10T01:40:28Z

benchmarks please

github-actions · 2024-05-10T01:40:43Z

Criterion benchmark results

Criterion benchmark report

YOU SHOULD PROBABLY IGNORE THESE RESULTS.

Criterion is a wall time based benchmarking system that is extremely noisy when run on CI. We collect these results for longitudinal analysis, but they are not reliable for comparing individual PRs.

Go look at the callgrind report instead.

empty

db	on disk	new latency	old latency	new throughput	old throughput
sqlite	💿	-	450.0±1.45ns	-	-
sqlite	🧠	-	447.7±1.55ns	-	-
stdb_raw	💿	714.5±0.73ns	745.7±1.66ns	-	-
stdb_raw	🧠	685.6±1.05ns	720.6±0.55ns	-	-

insert_1

db	on disk	schema	indices	preload	new latency	old latency	new throughput	old throughput

insert_bulk

db	on disk	schema	indices	preload	count	new latency	old latency	new throughput	old throughput
sqlite	💿	u32_u64_str	btree_each_column	2048	256	-	520.7±0.65µs	-	1920 tx/sec
sqlite	💿	u32_u64_str	unique_0	2048	256	-	137.0±0.41µs	-	7.1 Ktx/sec
sqlite	💿	u32_u64_u64	btree_each_column	2048	256	-	427.5±0.93µs	-	2.3 Ktx/sec
sqlite	💿	u32_u64_u64	unique_0	2048	256	-	127.5±0.40µs	-	7.7 Ktx/sec
sqlite	🧠	u32_u64_str	btree_each_column	2048	256	-	451.6±0.67µs	-	2.2 Ktx/sec
sqlite	🧠	u32_u64_str	unique_0	2048	256	-	123.9±0.92µs	-	7.9 Ktx/sec
sqlite	🧠	u32_u64_u64	btree_each_column	2048	256	-	372.0±0.58µs	-	2.6 Ktx/sec
sqlite	🧠	u32_u64_u64	unique_0	2048	256	-	108.5±0.51µs	-	9.0 Ktx/sec
stdb_raw	💿	u32_u64_str	btree_each_column	2048	256	521.6±14.03µs	511.6±13.12µs	1917 tx/sec	1954 tx/sec
stdb_raw	💿	u32_u64_str	unique_0	2048	256	515.6±22.78µs	424.9±16.16µs	1939 tx/sec	2.3 Ktx/sec
stdb_raw	💿	u32_u64_u64	btree_each_column	2048	256	396.9±13.49µs	358.8±6.84µs	2.5 Ktx/sec	2.7 Ktx/sec
stdb_raw	💿	u32_u64_u64	unique_0	2048	256	372.2±10.01µs	324.6±9.96µs	2.6 Ktx/sec	3.0 Ktx/sec
stdb_raw	🧠	u32_u64_str	btree_each_column	2048	256	341.7±0.24µs	338.4±0.46µs	2.9 Ktx/sec	2.9 Ktx/sec
stdb_raw	🧠	u32_u64_str	unique_0	2048	256	262.8±0.19µs	267.1±0.40µs	3.7 Ktx/sec	3.7 Ktx/sec
stdb_raw	🧠	u32_u64_u64	btree_each_column	2048	256	274.1±0.26µs	277.2±0.15µs	3.6 Ktx/sec	3.5 Ktx/sec
stdb_raw	🧠	u32_u64_u64	unique_0	2048	256	246.3±0.22µs	247.0±0.08µs	4.0 Ktx/sec	4.0 Ktx/sec

iterate

db	on disk	schema	indices	new latency	old latency	new throughput	old throughput
sqlite	💿	u32_u64_str	unique_0	-	21.8±0.12µs	-	44.8 Ktx/sec
sqlite	💿	u32_u64_u64	unique_0	-	20.0±0.11µs	-	48.8 Ktx/sec
sqlite	🧠	u32_u64_str	unique_0	-	20.5±0.06µs	-	47.7 Ktx/sec
sqlite	🧠	u32_u64_u64	unique_0	-	18.6±0.10µs	-	52.4 Ktx/sec
stdb_raw	💿	u32_u64_str	unique_0	4.7±0.00µs	4.7±0.00µs	208.4 Ktx/sec	207.8 Ktx/sec
stdb_raw	💿	u32_u64_u64	unique_0	4.6±0.00µs	4.6±0.00µs	214.2 Ktx/sec	212.8 Ktx/sec
stdb_raw	🧠	u32_u64_str	unique_0	4.7±0.00µs	4.7±0.00µs	209.7 Ktx/sec	209.0 Ktx/sec
stdb_raw	🧠	u32_u64_u64	unique_0	4.5±0.00µs	4.6±0.00µs	215.5 Ktx/sec	214.1 Ktx/sec

find_unique

db	on disk	key type	preload	new latency	old latency	new throughput	old throughput

filter

db	on disk	key type	index strategy	load	count	new latency	old latency	new throughput	old throughput
sqlite	💿	string	index	2048	256	-	66.4±0.15µs	-	14.7 Ktx/sec
sqlite	💿	u64	index	2048	256	-	64.2±0.31µs	-	15.2 Ktx/sec
sqlite	🧠	string	index	2048	256	-	64.6±0.06µs	-	15.1 Ktx/sec
sqlite	🧠	u64	index	2048	256	-	59.6±0.15µs	-	16.4 Ktx/sec
stdb_raw	💿	string	index	2048	256	5.1±0.00µs	5.2±0.00µs	190.2 Ktx/sec	189.4 Ktx/sec
stdb_raw	💿	u64	index	2048	256	5.1±0.00µs	5.1±0.00µs	193.2 Ktx/sec	192.1 Ktx/sec
stdb_raw	🧠	string	index	2048	256	5.1±0.00µs	5.1±0.00µs	191.3 Ktx/sec	190.3 Ktx/sec
stdb_raw	🧠	u64	index	2048	256	5.0±0.00µs	5.1±0.00µs	194.3 Ktx/sec	193.2 Ktx/sec

serialize

schema	format	count	new latency	old latency	new throughput	old throughput
u32_u64_str	bflatn_to_bsatn_fast_path	100	3.7±0.08µs	3.7±0.01µs	25.6 Mtx/sec	25.8 Mtx/sec
u32_u64_str	bflatn_to_bsatn_slow_path	100	3.5±0.02µs	3.4±0.01µs	27.5 Mtx/sec	27.7 Mtx/sec
u32_u64_str	bsatn	100	2.4±0.02µs	2.5±0.01µs	39.2 Mtx/sec	37.6 Mtx/sec
u32_u64_str	json	100	5.1±0.20µs	4.7±0.06µs	18.6 Mtx/sec	20.2 Mtx/sec
u32_u64_str	product_value	100	1013.2±0.40ns	1015.0±0.79ns	94.1 Mtx/sec	94.0 Mtx/sec
u32_u64_u64	bflatn_to_bsatn_fast_path	100	1455.4±39.41ns	1399.1±5.71ns	65.5 Mtx/sec	68.2 Mtx/sec
u32_u64_u64	bflatn_to_bsatn_slow_path	100	2.9±0.01µs	2.9±0.01µs	33.2 Mtx/sec	33.5 Mtx/sec
u32_u64_u64	bsatn	100	1693.2±20.30ns	1655.7±15.71ns	56.3 Mtx/sec	57.6 Mtx/sec
u32_u64_u64	json	100	3.2±0.02µs	3.1±0.03µs	30.0 Mtx/sec	30.8 Mtx/sec
u32_u64_u64	product_value	100	1009.7±2.75ns	1009.8±0.43ns	94.4 Mtx/sec	94.4 Mtx/sec
u64_u64_u32	bflatn_to_bsatn_fast_path	100	1081.2±3.84ns	1123.8±1.19ns	88.2 Mtx/sec	84.9 Mtx/sec
u64_u64_u32	bflatn_to_bsatn_slow_path	100	2.9±0.01µs	2.9±0.00µs	33.2 Mtx/sec	33.4 Mtx/sec
u64_u64_u32	bsatn	100	1686.0±37.85ns	1704.2±27.56ns	56.6 Mtx/sec	56.0 Mtx/sec
u64_u64_u32	json	100	3.2±0.02µs	3.3±0.04µs	29.6 Mtx/sec	29.2 Mtx/sec
u64_u64_u32	product_value	100	1010.1±0.73ns	1010.2±0.46ns	94.4 Mtx/sec	94.4 Mtx/sec

stdb_module_large_arguments

arg size	new latency	old latency	new throughput	old throughput
64KiB	88.6±13.92µs	91.7±4.61µs	-	-

stdb_module_print_bulk

line count	new latency	old latency	new throughput	old throughput
1	47.1±6.85µs	38.8±5.38µs	-	-
100	357.4±5.03µs	349.3±57.97µs	-	-
1000	2.3±0.53ms	2.3±0.48ms	-	-

remaining

name	new latency	old latency	new throughput	old throughput
sqlite/💿/update_bulk/u32_u64_str/unique_0/load=2048/count=256	-	47.8±0.19µs	-	20.4 Ktx/sec
sqlite/💿/update_bulk/u32_u64_u64/unique_0/load=2048/count=256	-	41.8±0.24µs	-	23.4 Ktx/sec
sqlite/🧠/update_bulk/u32_u64_str/unique_0/load=2048/count=256	-	40.7±0.05µs	-	24.0 Ktx/sec
sqlite/🧠/update_bulk/u32_u64_u64/unique_0/load=2048/count=256	-	36.2±0.28µs	-	27.0 Ktx/sec
stdb_module/💿/update_bulk/u32_u64_str/unique_0/load=2048/count=256	1379.8±7.34µs	1442.4±12.66µs	724 tx/sec	693 tx/sec
stdb_module/💿/update_bulk/u32_u64_u64/unique_0/load=2048/count=256	1092.4±12.21µs	1120.2±4.56µs	915 tx/sec	892 tx/sec
stdb_raw/💿/update_bulk/u32_u64_str/unique_0/load=2048/count=256	610.8±13.43µs	691.4±50.77µs	1637 tx/sec	1446 tx/sec
stdb_raw/💿/update_bulk/u32_u64_u64/unique_0/load=2048/count=256	551.1±9.34µs	542.0±20.72µs	1814 tx/sec	1845 tx/sec
stdb_raw/🧠/update_bulk/u32_u64_str/unique_0/load=2048/count=256	445.6±0.52µs	448.2±0.28µs	2.2 Ktx/sec	2.2 Ktx/sec
stdb_raw/🧠/update_bulk/u32_u64_u64/unique_0/load=2048/count=256	403.4±0.38µs	403.4±0.33µs	2.4 Ktx/sec	2.4 Ktx/sec

github-actions · 2024-05-10T01:41:47Z

Callgrind benchmark results

Callgrind Benchmark Report

These benchmarks were run using callgrind,
an instruction-level profiler. They allow comparisons between sqlite (sqlite), SpacetimeDB running through a module (stdb_module), and the underlying SpacetimeDB data storage engine (stdb_raw). Callgrind emulates a CPU to collect the below estimates.

Measurement changes larger than five percent are in bold.

In-memory benchmarks

callgrind: empty transaction

db	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
stdb_raw	6004	5999	0.08%	6856	6869	-0.19%
sqlite	5676	5686	-0.18%	6230	6150	1.30%

callgrind: filter

db	schema	indices	count	preload	_column	data_type	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
stdb_raw	u32_u64_str	no_index	64	128	2	string	120719	120710	0.01%	121493	121332	0.13%
stdb_raw	u32_u64_str	no_index	64	128	1	u64	78453	78444	0.01%	78979	78946	0.04%
stdb_raw	u32_u64_str	btree_each_column	64	128	2	string	25181	25176	0.02%	25675	25666	0.04%
stdb_raw	u32_u64_str	btree_each_column	64	128	1	u64	24141	24136	0.02%	24483	24522	-0.16%
sqlite	u32_u64_str	no_index	64	128	2	string	143658	143664	-0.00%	145268	145362	-0.06%
sqlite	u32_u64_str	no_index	64	128	1	u64	122999	123005	-0.00%	124309	124379	-0.06%
sqlite	u32_u64_str	btree_each_column	64	128	1	u64	130316	130316	0.00%	131892	131814	0.06%
sqlite	u32_u64_str	btree_each_column	64	128	2	string	133534	133521	0.01%	135294	135207	0.06%

callgrind: insert bulk

db	schema	indices	count	preload	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
stdb_raw	u32_u64_str	unique_0	64	128	943910	948257	-0.46%	963432	966417	-0.31%
stdb_raw	u32_u64_str	btree_each_column	64	128	1072824	1078622	-0.54%	1099084	1106696	-0.69%
sqlite	u32_u64_str	unique_0	64	128	398413	398423	-0.00%	418049	414803	0.78%
sqlite	u32_u64_str	btree_each_column	64	128	971486	971492	-0.00%	1012186	1008818	0.33%

callgrind: iterate

db	schema	indices	count	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
stdb_raw	u32_u64_str	unique_0	1024	147937	147932	0.00%	147975	148036	-0.04%
stdb_raw	u32_u64_str	unique_0	64	15760	15746	0.09%	15798	15850	-0.33%
sqlite	u32_u64_str	unique_0	1024	1046910	1046911	-0.00%	1050362	1050381	-0.00%
sqlite	u32_u64_str	unique_0	64	75041	75051	-0.01%	76213	76149	0.08%

callgrind: serialize_product_value

count	format	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
64	bsatn	25717	25717	0.00%	27961	28063	-0.36%
64	json	47438	47438	0.00%	50022	50022	0.00%
16	bsatn	8118	8118	0.00%	9410	9512	-1.07%
16	json	12142	12142	0.00%	13978	13978	0.00%

callgrind: update bulk

db	schema	indices	count	preload	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
stdb_raw	u32_u64_str	unique_0	1024	1024	22432484	22471962	-0.18%	22993300	23078216	-0.37%
stdb_raw	u32_u64_str	unique_0	64	128	1422971	1425382	-0.17%	1494659	1500696	-0.40%
sqlite	u32_u64_str	unique_0	1024	1024	1802006	1802012	-0.00%	1811302	1811094	0.01%
sqlite	u32_u64_str	unique_0	64	128	128542	128548	-0.00%	131494	131420	0.06%

On-disk benchmarks

callgrind: empty transaction

db	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
stdb_raw	6364	6365	-0.02%	7208	7259	-0.70%
sqlite	5718	5718	0.00%	6298	6208	1.45%

callgrind: filter

db	schema	indices	count	preload	_column	data_type	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
stdb_raw	u32_u64_str	no_index	64	128	2	string	121079	121076	0.00%	121933	121922	0.01%
stdb_raw	u32_u64_str	no_index	64	128	1	u64	78813	78810	0.00%	79331	79376	-0.06%
stdb_raw	u32_u64_str	btree_each_column	64	128	1	u64	24501	24502	-0.00%	24867	25052	-0.74%
stdb_raw	u32_u64_str	btree_each_column	64	128	2	string	25738	25540	0.78%	26320	26118	0.77%
sqlite	u32_u64_str	no_index	64	128	2	string	145579	145579	0.00%	147537	147369	0.11%
sqlite	u32_u64_str	no_index	64	128	1	u64	124935	124920	0.01%	126637	126502	0.11%
sqlite	u32_u64_str	btree_each_column	64	128	2	string	135571	135587	-0.01%	137649	137571	0.06%
sqlite	u32_u64_str	btree_each_column	64	128	1	u64	132412	132412	0.00%	134254	134284	-0.02%

callgrind: insert bulk

db	schema	indices	count	preload	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
stdb_raw	u32_u64_str	unique_0	64	128	895678	899093	-0.38%	945732	947465	-0.18%
stdb_raw	u32_u64_str	btree_each_column	64	128	1023147	1026673	-0.34%	1079977	1083013	-0.28%
sqlite	u32_u64_str	unique_0	64	128	415961	415961	0.00%	434619	431889	0.63%
sqlite	u32_u64_str	btree_each_column	64	128	1022061	1022071	-0.00%	1061043	1059193	0.17%

callgrind: iterate

db	schema	indices	count	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
stdb_raw	u32_u64_str	unique_0	1024	148297	148298	-0.00%	148355	148490	-0.09%
stdb_raw	u32_u64_str	unique_0	64	16120	16112	0.05%	16178	16304	-0.77%
sqlite	u32_u64_str	unique_0	1024	1049963	1049963	0.00%	1053699	1053685	0.00%
sqlite	u32_u64_str	unique_0	64	76813	76813	0.00%	78137	78103	0.04%

callgrind: serialize_product_value

count	format	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
64	bsatn	25717	25717	0.00%	27961	28063	-0.36%
64	json	47438	47438	0.00%	50022	50022	0.00%
16	bsatn	8118	8118	0.00%	9410	9512	-1.07%
16	json	12142	12142	0.00%	13978	13978	0.00%

callgrind: update bulk

db	schema	indices	count	preload	total reads + writes	old total reads + writes	Δrw	estimated cycles	old estimated cycles	Δcycles
stdb_raw	u32_u64_str	unique_0	1024	1024	21390132	21423310	-0.15%	22027740	22112670	-0.38%
stdb_raw	u32_u64_str	unique_0	64	128	1378537	1380648	-0.15%	1448577	1454450	-0.40%
sqlite	u32_u64_str	unique_0	1024	1024	1809802	1809802	0.00%	1818486	1818450	0.00%
sqlite	u32_u64_str	unique_0	64	128	132690	132690	0.00%	135806	135576	0.17%

2. Shrink SqlAst to 80 bytes, so it can be passed in registers 3. Store end-result Header in IndexSemiJoin 4. Remove operational use of Header in ColumnOp & build_query 5. Simplify RowRef::{get, project, project_owned}

1. Make IndexSemiJoin::filter infallible. 2. Make ColumnOp::compare and friends infallible. 3. Make RowRef::{get, project, project_owned} infallible.

2. Document RelValue::{get, read_or_take_column, project_owned} 3. Refactor optimize_select 4. Ensure in optimize_select that conditions are merged with preceding selects

…eless

Centril force-pushed the centril/columnop-with-colid branch from 2ca64d3 to 754654c Compare May 6, 2024 16:50

Centril marked this pull request as ready for review May 6, 2024 16:50

This was referenced May 7, 2024

From.{join => joins} + remove Option layer + Remove RelOps::try_fold #1208

Merged

subscription benches: reorder footprint fields #1209

Merged

Centril force-pushed the centril/columnop-with-colid branch 2 times, most recently from 27c19aa to e4f8dd9 Compare May 7, 2024 21:11

Centril changed the base branch from master to centril/index-join-inner-colid May 7, 2024 21:11

Centril force-pushed the centril/columnop-with-colid branch 5 times, most recently from 340516e to 7aaa0ef Compare May 8, 2024 13:33

Centril changed the title ~~[WIP] Split ColumnOp into one with row indices and one with FieldName~~ Split ColumnOp into one with row indices and one with FieldName & other enabled changes May 8, 2024

Centril force-pushed the centril/index-join-inner-colid branch 2 times, most recently from cf8e041 to c1d8311 Compare May 13, 2024 08:27

Centril force-pushed the centril/columnop-with-colid branch from 7aaa0ef to 3ccfeac Compare May 13, 2024 09:12

Centril force-pushed the centril/index-join-inner-colid branch 2 times, most recently from 5bc3f77 to 8ca55bf Compare May 13, 2024 09:17

Centril force-pushed the centril/columnop-with-colid branch from 3ccfeac to 3968f4f Compare May 13, 2024 09:18

Centril requested a review from joshua-spacetime May 13, 2024 10:26

Centril mentioned this pull request May 13, 2024

perf(query eval): change ColumnOp to use row positions instead of field names + allocate less in the common case #1113

Open

bfops added the release-any To be landed in any release window label May 13, 2024

Centril mentioned this pull request May 14, 2024

Flatten ColumnOp, avoiding allocation in the common case #1234

Open

joshua-spacetime approved these changes May 24, 2024

View reviewed changes

Centril force-pushed the centril/index-join-inner-colid branch from 8ca55bf to a53bb88 Compare May 24, 2024 15:37

Centril force-pushed the centril/columnop-with-colid branch 2 times, most recently from 74f913e to 13eca30 Compare May 28, 2024 07:07

Centril force-pushed the centril/index-join-inner-colid branch from a53bb88 to 72b74e4 Compare May 30, 2024 23:48

Centril added 12 commits May 31, 2024 01:49

1. Split ColumnOp into ColumnOp & FieldOp, former storing ColId

aa6b78e

2. Shrink SqlAst to 80 bytes, so it can be passed in registers 3. Store end-result Header in IndexSemiJoin 4. Remove operational use of Header in ColumnOp & build_query 5. Simplify RowRef::{get, project, project_owned}

Make parts of build_query actually infallible.

d2970d9

1. Make IndexSemiJoin::filter infallible. 2. Make ColumnOp::compare and friends infallible. 3. Make RowRef::{get, project, project_owned} infallible.

Make RelOps::next itself infallible

fff5cca

1. with_select{_cmp}: ensure type safety o query exec cannot panic

1054672

2. Document RelValue::{get, read_or_take_column, project_owned} 3. Refactor optimize_select 4. Ensure in optimize_select that conditions are merged with preceding selects

remove RelOps::{head, row_count}; head is redundant & row_count is us…

6f377ef

…eless

remove Relation trait; it does not carry its weight

9d77bd6

make build_query infallible

716b04e

simplify IndexSemiJoin, make it slightly less branchy

e5e750f

simplify try_index_join

3fd4037

split IndexSemiJoin into Left & Right parts

cdc6ddb

move get_field_pos to test code

162d4b5

move test version of build_query to test code

9227167

Centril force-pushed the centril/columnop-with-colid branch from 13eca30 to 9227167 Compare May 30, 2024 23:56

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Split `ColumnOp` into one with row indices and one with `FieldName` & other enabled changes #1207

Split `ColumnOp` into one with row indices and one with `FieldName` & other enabled changes #1207

Centril commented May 6, 2024 •

edited

cloutiertyler commented May 10, 2024

github-actions bot commented May 10, 2024 •

edited

Criterion benchmark report

empty

insert_1

insert_bulk

iterate

find_unique

filter

serialize

stdb_module_large_arguments

stdb_module_print_bulk

remaining

github-actions bot commented May 10, 2024 •

edited

Callgrind Benchmark Report

callgrind: empty transaction

callgrind: filter

callgrind: insert bulk

callgrind: iterate

callgrind: serialize_product_value

callgrind: update bulk

callgrind: empty transaction

callgrind: filter

callgrind: insert bulk

callgrind: iterate

callgrind: serialize_product_value

callgrind: update bulk

Split ColumnOp into one with row indices and one with FieldName & other enabled changes #1207

Are you sure you want to change the base?

Split ColumnOp into one with row indices and one with FieldName & other enabled changes #1207

Conversation

Centril commented May 6, 2024 • edited

Description of Changes

The main goal of this PR

Benchmarks

API and ABI breaking changes

Expected complexity level and risk

Testing

cloutiertyler commented May 10, 2024

github-actions bot commented May 10, 2024 • edited

Criterion benchmark report

empty

insert_1

insert_bulk

iterate

find_unique

filter

serialize

stdb_module_large_arguments

stdb_module_print_bulk

remaining

github-actions bot commented May 10, 2024 • edited

Callgrind Benchmark Report

callgrind: empty transaction

callgrind: filter

callgrind: insert bulk

callgrind: iterate

callgrind: serialize_product_value

callgrind: update bulk

callgrind: empty transaction

callgrind: filter

callgrind: insert bulk

callgrind: iterate

callgrind: serialize_product_value

callgrind: update bulk

Split `ColumnOp` into one with row indices and one with `FieldName` & other enabled changes #1207

Split `ColumnOp` into one with row indices and one with `FieldName` & other enabled changes #1207

Centril commented May 6, 2024 •

edited

github-actions bot commented May 10, 2024 •

edited

github-actions bot commented May 10, 2024 •

edited