issues
search
h2oai
/
db-benchmark
reproducible benchmark of database-like ops
https://h2oai.github.io/db-benchmark
Mozilla Public License 2.0
321
stars
85
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
check if cuDF join materializes answer
#155
jangorecki
closed
4 years ago
1
join data could be better distributed
#154
jangorecki
closed
3 years ago
1
Data generation for join queries fails for 1e9 rows
#153
amyskov
closed
4 years ago
2
dplyr cannot compute groupby on 1e9 rows data anymore
#152
jangorecki
opened
4 years ago
0
clickhouse should measure timings for cache=TRUE
#151
jangorecki
closed
4 years ago
1
Report step does not run
#150
maartenbreddels
closed
4 years ago
5
maintain clean output of main run.sh script
#149
jangorecki
closed
4 years ago
0
pending issues on cuDF
#148
jangorecki
closed
3 years ago
1
Julia: fix missing ( and ) bug
#147
bkamins
closed
4 years ago
9
julia upgrade post improvements, benchplot dict
#146
jangorecki
closed
4 years ago
2
update Julia benchmark
#145
bkamins
closed
4 years ago
15
Dask benchmark improvements
#144
ravwojdyla
closed
4 years ago
37
ability to pause benchmark
#143
jangorecki
closed
4 years ago
1
pandas not able to solve q10 anymore
#142
jangorecki
closed
4 years ago
1
join questions should ensure answers are materialized
#141
jangorecki
closed
4 years ago
1
maintenance doc
#140
jangorecki
closed
4 years ago
0
new task: timeseries
#139
jangorecki
opened
4 years ago
1
better performance regression tracking
#138
jangorecki
closed
3 years ago
3
ClickHouse join task
#137
jangorecki
closed
3 years ago
2
dask rounding different between different runs
#136
jangorecki
closed
4 years ago
0
pandas could use dt fread on csv
#135
jangorecki
closed
4 years ago
1
solution launcher
#134
jangorecki
closed
4 years ago
0
clickhouse segfault during groupby q10, sorted 1e9
#133
jangorecki
closed
4 years ago
1
clickhouse segfault causes benchmark launcher script to fail
#132
jangorecki
closed
4 years ago
1
new task: read
#131
jangorecki
opened
4 years ago
2
Make command-line script for running the benchmark easily
#130
st-pasha
closed
4 years ago
4
cudf should spil to main memory when running out of gpu memory
#129
jangorecki
closed
3 years ago
9
preview spark results using `tail` method
#128
jangorecki
closed
4 years ago
1
Timeout has to be more flexible
#127
jangorecki
closed
4 years ago
1
use on-disk data storage if OOM happens
#126
jangorecki
closed
4 years ago
5
benchplot rewritten
#125
jangorecki
closed
4 years ago
0
devel/stable releases of solutions
#124
jangorecki
closed
4 years ago
2
add support for multiple task in scripts launcher
#123
jangorecki
closed
4 years ago
1
pydatatable groupby q6 and q9
#122
jangorecki
closed
4 years ago
1
cudf groupby q6 median
#121
jangorecki
closed
4 years ago
4
dask groupby q9 correlation
#120
jangorecki
closed
3 years ago
1
Provide types explicitly when parsing CSV for juliadf
#119
nalimilan
closed
4 years ago
1
benchplot groupby count of groups for q8 incorrect
#118
jangorecki
closed
4 years ago
1
review groupby advanced questions
#117
jangorecki
closed
4 years ago
7
use dask-cudf to utilize multiple GPUs
#116
jangorecki
closed
3 years ago
8
sort-ops tasks
#115
jangorecki
opened
4 years ago
0
data.table should use all logical CPU in join script
#114
jangorecki
closed
4 years ago
0
more validation in report.R
#113
jangorecki
closed
4 years ago
1
ClickHouse server crash during groupby 1e9 k=2
#112
jangorecki
closed
4 years ago
3
not enough memory to read 1e9 data
#111
jangorecki
opened
4 years ago
4
R gc is choking during groupby 1e9 k=2
#110
jangorecki
opened
4 years ago
2
Incorrectly displayed input table information
#109
ismirsehregal
closed
4 years ago
2
benchplot legend tweaks
#108
jangorecki
closed
4 years ago
0
Add Rust's DataFusion (arrow)
#107
andygrove
opened
4 years ago
9
join task data generation
#106
jangorecki
closed
4 years ago
1
Previous
Next