blob: b2d03fd4da54447b1f5f4dd8a27a892e84edf379 [file] [log] [blame]
.. Licensed to the Apache Software Foundation (ASF) under one
.. or more contributor license agreements. See the NOTICE file
.. distributed with this work for additional information
.. regarding copyright ownership. The ASF licenses this file
.. to you under the Apache License, Version 2.0 (the
.. "License"); you may not use this file except in compliance
.. with the License. You may obtain a copy of the License at
.. http://www.apache.org/licenses/LICENSE-2.0
.. Unless required by applicable law or agreed to in writing,
.. software distributed under the License is distributed on an
.. "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
.. KIND, either express or implied. See the License for the
.. specific language governing permissions and limitations
.. under the License.
.. WARNING
.. This is an auto-generated file. Please do not edit.
.. To reproduce, please run :code:`./make_data_model_rst.sh`.
.. (This requires you have the
.. `psql client <https://www.postgresql.org/download/>`_
.. and have started the docker containers using
.. :code:`docker-compose up`).
.. _benchmark-data-model:
Benchmark data model
====================
.. graphviz:: data_model.dot
.. _benchmark-ingestion:
Benchmark ingestion helper functions
====================================
ingest_benchmark_run_view
-------------------------
:code:`ingest_benchmark_run_view(from_jsonb jsonb)`
The argument is a JSON object. NOTE: key names must be entirely
lowercase, or the insert will fail. Extra key-value pairs are ignored.
Example::
[
{
"benchmark_name": "Benchmark 2",
"benchmark_version": "version 0",
"parameter_values": {"arg0": 100, "arg1": 5},
"value": 2.5,
"git_commit_timestamp": "2019-02-08 22:35:53 +0100",
"git_hash": "324d3cf198444a",
"val_min": 1,
"val_q1": 2,
"val_q3": 3,
"val_max": 4,
"std_dev": 1.41,
"n_obs": 8,
"run_timestamp": "2019-02-14 03:00:05 -0600",
"mac_address": "08:00:2b:01:02:03",
"benchmark_language": "Python",
"language_implementation_version": "CPython 2.7",
"dependencies": {"six": "", "numpy": "1.14", "other_lib": "1.0"}
},
{
"benchmark_name": "Benchmark 2",
"benchmark_version": "version 0",
"parameter_values": {"arg0": 1000, "arg1": 5},
"value": 5,
"git_commit_timestamp": "2019-02-08 22:35:53 +0100",
"git_hash": "324d3cf198444a",
"std_dev": 3.14,
"n_obs": 8,
"run_timestamp": "2019-02-14 03:00:10 -0600",
"mac_address": "08:00:2b:01:02:03",
"benchmark_language": "Python",
"language_implementation_version": "CPython 2.7",
"dependencies": {"six": "", "numpy": "1.14", "other_lib": "1.0"}
}
]
To identify which columns in "benchmark_run_view" are required,
please see the view documentation in :ref:`benchmark-data-model`.
back to `Benchmark data model <benchmark-data-model>`_
ingest_benchmark_view
---------------------
:code:`ingest_benchmark_view(from_jsonb jsonb)`
The argument is a JSON object. NOTE: key names must be entirely
lowercase, or the insert will fail. Extra key-value pairs are ignored.
Example::
[
{
"benchmark_name": "Benchmark 1",
"parameter_names": ["arg0", "arg1", "arg2"],
"benchmark_description": "First benchmark",
"benchmark_type": "Time",
"units": "miliseconds",
"lessisbetter": true,
"benchmark_version": "second version",
"benchmark_language": "Python"
},
{
"benchmark_name": "Benchmark 2",
"parameter_names": ["arg0", "arg1"],
"benchmark_description": "Description 2.",
"benchmark_type": "Time",
"units": "nanoseconds",
"lessisbetter": true,
"benchmark_version": "second version",
"benchmark_language": "Python"
}
]
To identify which columns in "benchmark_view" are required,
please see the view documentation in :ref:`benchmark-data-model`.
back to `Benchmark data model <benchmark-data-model>`_
ingest_benchmark_runs_with_context
----------------------------------
:code:`ingest_benchmark_runs_with_context(from_jsonb jsonb)`
The argument is a JSON object. NOTE: key names must be entirely
lowercase, or the insert will fail. Extra key-value pairs are ignored.
The object contains three key-value pairs::
{"context": {
"mac_address": "08:00:2b:01:02:03",
"benchmark_language": "Python",
"language_implementation_version": "CPython 3.6",
"dependencies": {"six": "", "numpy": "1.14", "other_lib": "1.0"},
"git_commit_timestamp": "2019-02-14 22:42:22 +0100",
"git_hash": "123456789abcde",
"run_timestamp": "2019-02-14 03:00:40 -0600",
"extra stuff": "does not hurt anything and will not be added."
},
"benchmark_version": {
"Benchmark Name 1": "Any string can be a version.",
"Benchmark Name 2": "A git hash can be a version.",
"An Unused Benchmark Name": "Will be ignored."
},
"benchmarks": [
{
"benchmark_name": "Benchmark Name 1",
"parameter_values": {"argument1": 1, "argument2": "value2"},
"value": 42,
"val_min": 41.2,
"val_q1": 41.5,
"val_q3": 42.5,
"val_max": 42.8,
"std_dev": 0.5,
"n_obs": 100,
"run_metadata": {"any": "key-value pairs"},
"run_notes": "Any relevant notes."
},
{
"benchmark_name": "Benchmark Name 2",
"parameter_values": {"not nullable": "Use {} if no params."},
"value": 8,
"std_dev": 1,
"n_obs": 2,
}
]
}
- The entry for "context" contains the machine, environment, and timestamp
information common to all of the runs
- The entry for "benchmark_version" maps benchmark
names to their version strings. (Which can be a git hash,
the entire code string, a number, or any other string of your choice.)
- The entry for "benchmarks" is a list of benchmark run data
for the given context and benchmark versions. The first example
benchmark run entry contains all possible values, even
nullable ones, and the second entry omits all nullable values.
back to `Benchmark data model <benchmark-data-model>`_
ingest_machine_view
-------------------
:code:`ingest_machine_view(from_jsonb jsonb)`
The argument is a JSON object. NOTE: key names must be entirely
lowercase, or the insert will fail. Extra key-value pairs are ignored.
Example::
{
"mac_address": "0a:00:2d:01:02:03",
"machine_name": "Yet-Another-Machine-Name",
"memory_bytes": 8589934592,
"cpu_actual_frequency_hz": 2300000000,
"os_name": "OSX",
"architecture_name": "x86_64",
"kernel_name": "18.2.0",
"cpu_model_name": "Intel(R) Core(TM) i5-7360U CPU @ 2.30GHz",
"cpu_core_count": 2,
"cpu_thread_count": 4,
"cpu_frequency_max_hz": 2300000000,
"cpu_frequency_min_hz": 2300000000,
"cpu_l1d_cache_bytes": 32768,
"cpu_l1i_cache_bytes": 32768,
"cpu_l2_cache_bytes": 262144,
"cpu_l3_cache_bytes": 4194304,
"machine_other_attributes": {"just": "an example"},
"gpu_information": "",
"gpu_part_number": "",
"gpu_product_name": ""
}
To identify which columns in "machine_view" are required,
please see the view documentation in :ref:`benchmark-data-model`.
back to `Benchmark data model <benchmark-data-model>`_
.. _benchmark-views:
Benchmark views
===============
benchmark_run_view
------------------
Each benchmark run.
- Each entry is unique on the machine, environment, benchmark,
and git commit timestamp.
=============================== =========== ======== =========== ===========
Column Type Nullable Default Description
=============================== =========== ======== =========== ===========
benchmark_run_id int8 not null serial primary key
benchmark_name citext not null unique
benchmark_version citext not null unique
parameter_values jsonb not null '{}'::jsonb unique
value numeric not null
git_commit_timestamp timestamptz not null unique
git_hash text not null
val_min numeric
val_q1 numeric
val_q3 numeric
val_max numeric
std_dev numeric not null
n_obs int4 not null
run_timestamp timestamptz not null unique
run_metadata jsonb
run_notes text
mac_address macaddr not null unique
benchmark_language citext not null unique
language_implementation_version citext not null ''::citext unique
dependencies jsonb not null '{}'::jsonb unique
=============================== =========== ======== =========== ===========
back to `Benchmark data model <benchmark-data-model>`_
benchmark_view
--------------
The details about a particular benchmark.
- "benchmark_name" is unique for a given "benchmark_language"
- Each entry is unique on
("benchmark_language", "benchmark_name", "benchmark_version")
===================== ====== ======== ======= ===========
Column Type Nullable Default Description
===================== ====== ======== ======= ===========
benchmark_id int4 not null serial primary key
benchmark_name citext not null unique
parameter_names _text
benchmark_description text not null
benchmark_type citext not null unique
units citext not null unique
lessisbetter bool not null
benchmark_version citext not null unique
benchmark_language citext not null unique
===================== ====== ======== ======= ===========
back to `Benchmark data model <benchmark-data-model>`_
environment_view
----------------
The build environment used for a reported benchmark run.
(Will be inferred from each "benchmark_run" if not expicitly added).
- Each entry is unique on
("benchmark_language", "language_implementation_version", "dependencies")
- "benchmark_language" is unique in the "benchmark_language" table
- "benchmark_language" plus "language_implementation_version" is unique in
the "language_implementation_version" table
- "dependencies" is unique in the "dependencies" table
=============================== ====== ======== =========== ===========
Column Type Nullable Default Description
=============================== ====== ======== =========== ===========
environment_id int4 not null serial primary key
benchmark_language citext not null unique
language_implementation_version citext not null ''::citext unique
dependencies jsonb not null '{}'::jsonb unique
=============================== ====== ======== =========== ===========
back to `Benchmark data model <benchmark-data-model>`_
machine_view
------------
The machine environment (CPU, GPU, OS) used for each benchmark run.
- "mac_address" is unique in the "machine" table
- "gpu_part_number" is unique in the "gpu" (graphics processing unit) table
Empty string (''), not null, is used for machines that won't use the GPU
- "cpu_model_name" is unique in the "cpu" (central processing unit) table
- "os_name", "os_architecture_name", and "os_kernel_name"
are unique in the "os" (operating system) table
- "machine_other_attributes" is a key-value store for any other relevant
data, e.g. '{"hard_disk_type": "solid state"}'
======================== ======= ======== ========== ===========
Column Type Nullable Default Description
======================== ======= ======== ========== ===========
machine_id int4 not null serial primary key
mac_address macaddr not null unique
machine_name citext not null
memory_bytes int8 not null
cpu_actual_frequency_hz int8 not null
os_name citext not null unique
architecture_name citext not null unique
kernel_name citext not null ''::citext unique
cpu_model_name citext not null unique
cpu_core_count int4 not null
cpu_thread_count int4 not null
cpu_frequency_max_hz int8 not null
cpu_frequency_min_hz int8 not null
cpu_l1d_cache_bytes int4 not null
cpu_l1i_cache_bytes int4 not null
cpu_l2_cache_bytes int4 not null
cpu_l3_cache_bytes int4 not null
gpu_information citext not null ''::citext unique
gpu_part_number citext not null ''::citext
gpu_product_name citext not null ''::citext
machine_other_attributes jsonb
======================== ======= ======== ========== ===========
back to `Benchmark data model <benchmark-data-model>`_