-
Notifications
You must be signed in to change notification settings - Fork 0
/
figures.R
172 lines (150 loc) · 5.74 KB
/
figures.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
# Copyright (c) 2022, Voltron Data.
#
# Licensed 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.
# Packages needs ----------------------------------------------------------
library(arrow)
library(dplyr)
library(conbenchcoms) ## remotes::install_github("conbench/conbenchcoms", dependencies = TRUE)
library(ggplot2)
library(lubridate)
library(tidyr)
library(forcats)
library(geomtextpath)
library(gh)
# Monitoring Performance Changes in Arrow ---------------------------------
## change plot
Sys.unsetenv("CONBENCH_URL")
## history of query 22
commits_gh_raw <- gh::gh(
"/repos/apache/arrow/commits",
since = "2022-04-15T00:00:00",
until = "2022-08-19T00:00:00",
.limit = Inf
)
## release dates
release_dates <- data.frame(
release_date = c(as.Date("2022-05-02"), as.Date("2022-08-10")),
version = c("Arrow 8.0.0", "Arrow 9.0.0"),
hjust_num = c(0.1, 0.9)
)
cpp_commits_gh <- data.frame(
commits_gh_sha = vapply(commits_gh_raw, "[[", "", "sha"),
commits_gh_datetime = vapply(seq_along(commits_gh_raw), \(x) commits_gh_raw[[x]]$commit$author$date, character(1)),
commit_message = vapply(seq_along(commits_gh_raw), \(x) commits_gh_raw[[x]]$commit$message, character(1))
) %>%
mutate(commits_gh_date = ymd_hms(commits_gh_datetime)) %>%
filter(grepl("\\[C\\+\\+\\]", commit_message)) %>%
filter(!grepl("^MINOR", commit_message)) %>%
group_by(commits_gh_date = as.Date(commits_gh_datetime)) %>%
filter(commits_gh_datetime == max(commits_gh_datetime)) %>% ## last commit of the day
as_tibble()
## on just one piece of hardware
history_join_runs <- runs(cpp_commits_gh$commits_gh_sha) %>%
filter(hardware.name == "ursa-i9-9960x") %>%
mutate(commit.timestamp = ymd_hms(commit.timestamp))
## this step will take some time
history_join_benchmarks <- benchmarks(history_join_runs$id)
history_join_benchmarks %>%
filter(
tags.query_id == "TPCH-22",
tags.format %in% "parquet",
tags.scale_factor == 10
) %>%
left_join(history_join_runs, by = c("run_id" = "id")) %>%
mutate(commit.timestamp = as.Date(commit.timestamp)) %>%
ggplot(aes(x = commit.timestamp, y = stats.mean)) +
geom_line(colour = "#005050", size = 1.1) +
geom_point(colour = "#005050", alpha = 0.8, size = 1.1) +
geom_textvline(data = release_dates,
aes(
xintercept = release_date,
label = version,
hjust = hjust_num),
size = 6, linetype = 2, family = "Work Sans"
) +
scale_x_date() +
labs(
x = "Benchmark Date",
y = "Time to Complete the Query (s)",
title = "Query 22 Timings - Arrow release dates marked by vertical lines",
) +
theme_minimal(base_family = "Work Sans") +
theme(
plot.title.position = "plot",
plot.background = element_rect(fill = "white"),
axis.text = element_text(size = 14),
axis.title = element_text(size = 14)
)
ggsave("history.png", width = 2560, height = 1707, units = "px")
ggsave("history.jpeg", width = 2560, height = 1707, units = "px")
## Let's look at all the queries
## before and after where benchmarks ran
shas <- c(
"old" = "0024962ff761d1d5f3a63013e67886334f1e57ca",
"new" = "ee2e9448c8565820ba38a2df9e44ab6055e5df1d"
)
join_runs <- runs(shas) %>%
filter(hardware.name == "ursa-i9-9960x", !has_errors) %>%
mutate(commit.timestamp = ymd_hms(commit.timestamp))
join_benchmarks <- benchmarks(join_runs$id) %>%
filter(
tags.format == "parquet",
tags.scale_factor == 10
) %>%
left_join(join_runs, by = c("run_id" = "id")) %>%
mutate(state = case_when(
commit.sha == "0024962ff761d1d5f3a63013e67886334f1e57ca" ~ "old",
commit.sha == "ee2e9448c8565820ba38a2df9e44ab6055e5df1d" ~ "new",
TRUE ~ NA_character_
))
query_with_joins <- c(
"TPCH-02", "TPCH-03", "TPCH-04", "TPCH-05",
"TPCH-07", "TPCH-08", "TPCH-09", "TPCH-10",
"TPCH-11", "TPCH-12", "TPCH-13", "TPCH-14",
"TPCH-15", "TPCH-16", "TPCH-17", "TPCH-18",
"TPCH-19", "TPCH-20", "TPCH-21", "TPCH-22"
)
## percent change
join_benchmarks %>%
select(tags.query_id, state, stats.mean) %>%
pivot_wider(names_from = state, values_from = stats.mean) %>%
mutate(per_change = old / new) %>%
mutate(query_with_joins = ifelse(tags.query_id %in% query_with_joins, "Query with joins", "Query without a join")) %>%
mutate(tags.query_id = gsub("TPCH-", "", tags.query_id)) %>%
ggplot(aes(y = fct_reorder(tags.query_id, per_change), x = per_change)) +
labs(
x = "Percent Change", y = "Query",
title = "Percent Change in Query Performance after Hash Join Improvement"
) +
scale_x_continuous(labels = ~ scales::percent(.x, big.mark = ",")) +
scale_fill_manual(values = c("#005050", "#3BD9FF")) +
geom_col(aes(fill = query_with_joins), alpha = 0.8) +
geom_text(
aes(
x = per_change - 0.18,
label = scales::percent(per_change, accuracy = 1, big.mark = ",")
),
size = 4,
colour = "white",
family = "Work Sans") +
theme_minimal(base_family = "Work Sans") +
theme(
plot.title.position = "plot",
plot.background = element_rect(fill = "white"),
legend.title = element_blank(),
legend.position = "bottom",
axis.text = element_text(size = 12),
axis.title = element_text(size = 12)
)
ggsave("all_queries.png", width = 2560, height = 1707, units = "px")
ggsave("all_queries.jpeg", width = 2560, height = 1707, units = "px")