quasiben

NVTX Push-Pop Range Statistics:

 Time(%)  Total Time (ns)  Instances     Average      Minimum      Maximum                   Range
 -------  ---------------  ---------  -------------  ----------  -----------  ------------------------------------
    17.2   64,995,321,276    394,497      164,754.9         625  901,797,038  ucxpy:UCXPY_PROGRESS
    14.5   54,752,809,329        486  112,660,101.5     891,800  252,248,846  libcudf:inner_join
    10.9   41,264,369,784        243  169,812,221.3   1,515,855  193,479,633  cudf_python:SORT_INDEX
    10.6   40,038,984,207        648   61,788,555.9  59,485,360   66,858,388  cudf_python:SCATTER_BY_MAP
    10.3   39,047,108,085        243  160,687,687.6     341,675  181,847,452  libcudf:sorted_order
    10.0   37,966,864,680        648   58,590,840.6  53,434,145   65,389,102  libcudf:partition

quasiben

#!/bin/bash -eu

# Environment variables to enable GPUs, InfiniBand, NVLink
# These are read by the scheduler and client script
module load cuda/11.0.3
export PYTHON=/gpfs/fs1/bzaitlen/miniconda3/envs/20201008/bin/python
export DASK_UCX__CUDA_COPY=True
export DASK_UCX__TCP=True
export DASK_UCX__NVLINK=True
export DASK_UCX__INFINIBAND=True

quasiben

quasiben

name: deconvolve
channels:
  - rapidsai-nightly
  - nvidia
  - conda-forge
  - defaults
dependencies:
  - _libgcc_mutex=0.1=conda_forge
  - _openmp_mutex=4.5=1_llvm
  - abseil-cpp=20200225.2=he1b5a44_0

quasiben

quasiben

"""
conda install -c conda-forge postgresql psycopg2 sqlalchemy

initdb -D /tmp/dask_db
postgresql -D /tmp/dask_db
pg_ctl -D /tmp/dask_db -l logfile start
createuser --encrypted --pwprompt dask
createdb --owner=dask dask_db
"""

quasiben

Query 01
Find top 100 products that are sold together frequently in given stores. Only products in certain categories
sold in specific stores are considered, and "sold together frequently" means at least 50 customers bought
these productstogether in a transaction.
Query 02
Find the top 30 products that are mostly viewed together with a given product in online store. Note that
the order of products viewed does not matter, and "viewed together" relates to a web_clickstreams
click_session of a known user with a session timeout of 60min.If the duration between two click of a user is
greater then the session timeout, a new session begins. With a session timeout of 60min.
Query 03

quasiben

quasiben

quasiben