Project

General

Profile

Pipeline Optimization » History » Version 23

Bryan Cosca, 04/21/2016 05:33 PM

1 1 Bryan Cosca
h1. Pipeline Optimization
2
3 17 Bryan Cosca
h2. Overview
4 1 Bryan Cosca
5 19 Bryan Cosca
This wiki page is designed to help users make their Arvados pipelines cost and compute efficient for production level data. It will teach users how to make the most out of their pipeline to save money or compute time on their clusters.
6 17 Bryan Cosca
7
h2. Pipeline design
8
9
h3. Choosing the right number of jobs
10 1 Bryan Cosca
11 19 Bryan Cosca
Currently, your pipeline may have one script that runs everything from alignment to variant calling. If something unexpected happens and your pipeline fails, you don't want to start from the beginning again, you want to resume from where you left off!
12 1 Bryan Cosca
13 19 Bryan Cosca
This is where the notion of choosing which jobs to create comes into play. The right number of jobs depends on how versatile you want your pipeline to be. Specifically, how many steps do you want your pipeline to have?
14
15
Questions you may ask yourself are: 
16
17 18 Bryan Cosca
Do I want to output all your intermediate files to keep? 
18 1 Bryan Cosca
How many checkpoints do I want my pipeline to have?
19 22 Bryan Cosca
Do I want to do alignment and variant calling in one step? Should I separate them? (usually, yes!)
20 1 Bryan Cosca
21 19 Bryan Cosca
Each job must output a collection of files. This is how you can checkpoint your pipeline. The outputs of jobs are always available, so you can reuse them when the next job fails.
22 18 Bryan Cosca
23 23 Bryan Cosca
If you don't want to have a lot of jobs or a lot of outputs, you should combine multiple commands in a job. You can do all your bam realignment/recalibration in one step, and you don't need to output all the intermediate bam files to keep. Keep in mind, you always choose what to upload to keep, so if you don't need files later on, its best to leave it on the compute node. One way to do so is to use a temporary directory or "tmpdir":https://dev.arvados.org/projects/arvados/wiki/Writing_a_Script_Calling_a_Third_Party_Tool to store all your intermediate files, and then only upload what you want to keep.
24 19 Bryan Cosca
25
If you choose to do multiple computations in a job, you should try piping them together. Creating pipes between tools has shown to sometimes be faster than writing/reading from disk. Feel free to pipe your tools together, for example using subprocess.PIPE in the "python subprocess module":https://docs.python.org/2/library/subprocess.html. Sometimes piping is faster, sometimes it's not. You'll have to try for yourself. 
26
27 1 Bryan Cosca
If you want a lot of checkpoints you should have a job for each command. You'll be able to resume/restart work easily if any unexpected interruption happens. Also, you can choose different node types for each command. For example, BWA-mem can scale a lot better than fastqc or varscan, so having a 16 core node for something that doesn't have native multi-threading would be wasteful.
28
29 23 Bryan Cosca
One more thing to keep in mind is that jobs don't necessarily depend on each other (unless you set "output_of":http://doc.arvados.org/user/tutorials/running-external-program.html in a parameter). If you have multiple jobs depending on one input, then they will all run at the same time. For example, if you have multiple variant callers, they will all run in parallel against one bam file.
30
31 19 Bryan Cosca
h3. Writing to keep
32
33 1 Bryan Cosca
Once you've chosen how to create jobs, you'll need a way to write your outputs to keep.
34
35 23 Bryan Cosca
There are two ways to write your output collection to keep. "Writing straight to keep":https://dev.arvados.org/projects/arvados/wiki/Writing_a_Script_Calling_a_Third_Party_Tool and staging a file in a temporary directory and then uploading to keep.
36 19 Bryan Cosca
37
In general, writing straight to keep will reap more benefits. TaskOutputDir acts like a pipe, so you never have to spend node time on uploading data.
38
One problem though is if your job is dependent on using your output directory as a temporary space for files. If your job uses its output directory for computation, then your job will be trying to compute over a network and could become very slow. That being said, it's very safe for a job to write to a temporary directory then spending compute time uploading to keep. If you have time, it's worth trying both and seeing how much time you save by doing both. Most of the time, writing straight to keep using TaskOutputDir will be the right option, but using a tmpdir is always the safe alternative.
39
40 1 Bryan Cosca
h3. Choosing the right number of tasks
41
42 19 Bryan Cosca
Once you have your job architecture set up, you need to start looking at the core level of the compute node. Specifically, how much compute are you using and how can you take advantage of a machine.
43 1 Bryan Cosca
44
If you know how much RAM you want your job to have, you can set min_ram_mb_per_node in runtime_constraints to ensure that you get a compute node big enough for your computation. Most tools scale up with more RAM, so feel free to test your job on different nodetypes to see what it works on the best.
45 19 Bryan Cosca
46 21 Sarah Guthrie
If you have a job that's very i/o intensive, it's worth it to change keep_cache_mb_per_task in runtime_constraints. You can do more computation if you have more data readily available! The default is 256 MiB, which is equal to 4 keep blocks. Try 512MiB or 1024MiB to see if you get better performance.
47 17 Bryan Cosca
48 21 Sarah Guthrie
If you have a job is natively multithreaded, you should set num_cores_per_node in runtime_constraints to the number of threads you want. Most clouds have machines up to 16 threads, so it would be worth testing on multiple nodetypes to see what works best for you.
49
50
If you want multiple computations (tasks) on a node, setting max_tasks_per_node in "runtime_constraints":http://doc.arvados.org/api/schema/Job.html allows you to choose how many tasks you would like to run on a machine. For example, if you have a lot of small tasks that use 1 core/1GB ram, you can put multiple of those on a bigger machine. For example, 8 tasks on an 8 core machine. If you want to utilize machines better for cost savings, you should use crunchstat-summary to find out the maximum memory/cpu usage for one task, and see if you can fit more than 1 of those on a machine. One warning, however is if you do run out of RAM (some compute nodes can't swap) your process will die.
51 17 Bryan Cosca
52 19 Bryan Cosca
h3. How to optimize the number of tasks when you don't have native multithreading
53 17 Bryan Cosca
54 19 Bryan Cosca
Tools like GATK have native multi-threading where if you pass a -t, it will use the correct number of cores on the node. You usually want take advantage of this, and choose the min_cores_per_node that equals your threading parameter. You can use any number of min_tasks_per_node making sure that your tool-threading*min_tasks_per_node is <= min_cores_per_node. Also making sure that your node has enough RAM to allocate to all the tasks.
55
56
Tools like varscan/freebayes don't have native multi-threading so you need to find a workaround. Generally, these tools have a -L/--intervals to pass in certain loci to work on. If you have a bed file you can split reads on, then you can create a new task per interval. Then, have a job merge the outputs together.
57
58 2 Bryan Cosca
h2. Crunchstat Summary
59 16 Bryan Cosca
60 3 Bryan Cosca
Crunchstat-summary is an arvados tool to help choose optimal configurations for arvados jobs and pipeline instances. It helps you choose "runtime_constraints":http://doc.arvados.org/api/schema/Job.html specified in the pipeline template under each job by graphing job statistics. For example: CPU usage, RAM, and Keep network traffic over time.
61 1 Bryan Cosca
62 2 Bryan Cosca
h3. How to install crunchstat-summary
63 3 Bryan Cosca
64
<pre>
65
$ git clone https://github.com/curoverse/arvados.git
66
$ cd arvados/tools/crunchstat-summary/
67
$ python setup.py build
68
$ python setup.py install --user
69
</pre>
70 1 Bryan Cosca
71
h3. How to use crunchstat-summary
72 3 Bryan Cosca
73
<pre>
74
$ ./bin/crunchstat-summary --help
75
usage: crunchstat-summary [-h]
76
                          [--job UUID | --pipeline-instance UUID | --log-file LOG_FILE]
77
                          [--skip-child-jobs] [--format {html,text}]
78
                          [--verbose]
79
80
Summarize resource usage of an Arvados Crunch job
81
82
optional arguments:
83
  -h, --help            show this help message and exit
84
  --job UUID            Look up the specified job and read its log data from
85
                        Keep (or from the Arvados event log, if the job is
86
                        still running)
87 1 Bryan Cosca
  --pipeline-instance UUID
88 3 Bryan Cosca
                        Summarize each component of the given pipeline
89
                        instance
90
  --log-file LOG_FILE   Read log data from a regular file
91
  --skip-child-jobs     Do not include stats from child jobs
92
  --format {html,text}  Report format
93
  --verbose, -v         Log more information (once for progress, twice for
94
                        debug)
95
</pre>
96 1 Bryan Cosca
97 16 Bryan Cosca
There are two ways of using crunchstat-summary: a text view for an overall view of a job or an html page, which graphs usage over time.
98 14 Bryan Cosca
99 16 Bryan Cosca
Case 1: A job that does bwa-aln mapping and converts to bam using samtools.
100 8 Bryan Cosca
101
<pre>
102 19 Bryan Cosca
$ ~/arvados/tools/crunchstat-summary/bin/crunchstat-summary --format text --job qr1hi-8i9sb-bzn6hzttfu9cetv
103 8 Bryan Cosca
category        metric  task_max        task_max_rate   job_total
104
blkio:202:0     read    310334464       -       913853440
105
blkio:202:0     write   2567127040      -       7693406208
106
blkio:202:16    read    8036201472      155118884.01    4538585088
107
blkio:202:16    write   55502038016     0       0
108
blkio:202:32    read    2756608 100760.59       6717440
109
blkio:202:32    write   53570560        0       99514368
110
cpu     cpus    8       -       -
111
cpu     sys     1592.34 1.17    805.32
112
cpu     user    11061.28        7.98    4620.17
113
cpu     user+sys        12653.62        8.00    5425.49
114
mem     cache   7454289920      -       -
115
mem     pgmajfault      1859    -       830
116
mem     rss     7965265920      -       -
117
mem     swap    5537792 -       -
118
net:docker0     rx      2023609029      -       2093089079
119
net:docker0     tx      21404100070     -       49909181906
120
net:docker0     tx+rx   23427709099     -       52002270985
121
net:eth0        rx      44750669842     67466325.07     14233805360
122
net:eth0        tx      2126085781      20171074.09     3670464917
123 1 Bryan Cosca
net:eth0        tx+rx   46876755623     67673532.73     17904270277
124
time    elapsed 949     -       1899
125 8 Bryan Cosca
# Number of tasks: 3
126
# Max CPU time spent by a single task: 12653.62s
127
# Max CPU usage in a single interval: 799.88%
128
# Overall CPU usage: 285.70%
129
# Max memory used by a single task: 7.97GB
130
# Max network traffic in a single task: 46.88GB
131
# Max network speed in a single interval: 67.67MB/s
132
# Keep cache miss rate 0.00%
133
# Keep cache utilization 0.00%
134 1 Bryan Cosca
#!! qr1hi-8i9sb-bzn6hzttfu9cetv max CPU usage was 800% -- try runtime_constraints "min_cores_per_node":8
135 8 Bryan Cosca
#!! qr1hi-8i9sb-bzn6hzttfu9cetv max RSS was 7597 MiB -- try runtime_constraints "min_ram_mb_per_node":7782
136
</pre>
137
138 19 Bryan Cosca
$ ~/arvados/tools/crunchstat-summary/bin/crunchstat-summary --format html --job qr1hi-8i9sb-bzn6hzttfu9cetv > qr1hi-8i9sb-bzn6hzttfu9cetv.html
139
140 1 Bryan Cosca
!86538baca4ecef099d9fad76ad9c7180.png!
141
142 16 Bryan Cosca
Here, you can see the distinct computation steps between the bwa-aln and the samtools step. Since there is a noticeable plateau on CPU usage for both computations, it would be worth trying to run the job on a bigger node. For example, a 16 core node to see if the computation can scale higher than 8 cores. 
143 1 Bryan Cosca
144 16 Bryan Cosca
Another thing to note is you can also see the runtime_constraints recommendations. These recommendations are for you to set to ensure the job will be able to call the right node type and run reliably when reproduced.
145
146 12 Bryan Cosca
Case study 2: FastQC
147
148
<pre>
149 19 Bryan Cosca
$ ~/arvados/tools/crunchstat-summary/bin/crunchstat-summary --format text --job qr1hi-8i9sb-nxqqxravvapt10h
150 12 Bryan Cosca
category	metric	task_max	task_max_rate	job_total
151
blkio:0:0	read	174349211138	65352499.20	174349211138
152
blkio:0:0	write	0	0	0
153
cpu	cpus	8	-	-
154
cpu	sys	364.95	0.17	364.95
155
cpu	user	17589.59	6.59	17589.59
156 1 Bryan Cosca
cpu	user+sys	17954.54	6.72	17954.54
157
fuseops	read	1330241	498.40	1330241
158 16 Bryan Cosca
fuseops	write	0	0	0
159 1 Bryan Cosca
keepcache	hit	2655806	1038.00	2655806
160
keepcache	miss	2633	1.60	2633
161
keepcalls	get	2658439	1039.00	2658439
162 16 Bryan Cosca
keepcalls	put	0	0	0
163 1 Bryan Cosca
mem	cache	19836608512	-	-
164 16 Bryan Cosca
mem	pgmajfault	19	-	19
165 1 Bryan Cosca
mem	rss	1481367552	-	-
166 16 Bryan Cosca
net:eth0	rx	178321	17798.40	178321
167
net:eth0	tx	7156	685.00	7156
168 11 Bryan Cosca
net:eth0	tx+rx	185477	18483.40	185477
169 16 Bryan Cosca
net:keep0	rx	175959092914	107337311.20	175959092914
170 1 Bryan Cosca
net:keep0	tx	0	0	0
171
net:keep0	tx+rx	175959092914	107337311.20	175959092914
172
time	elapsed	3301	-	3301
173 16 Bryan Cosca
# Number of tasks: 1
174 1 Bryan Cosca
# Max CPU time spent by a single task: 17954.54s
175 16 Bryan Cosca
# Max CPU usage in a single interval: 672.01%
176 1 Bryan Cosca
# Overall CPU usage: 543.91%
177 16 Bryan Cosca
# Max memory used by a single task: 1.48GB
178 1 Bryan Cosca
# Max network traffic in a single task: 175.96GB
179 16 Bryan Cosca
# Max network speed in a single interval: 107.36MB/s
180
# Keep cache miss rate 0.10%
181 11 Bryan Cosca
# Keep cache utilization 99.09%
182
#!! qr1hi-8i9sb-nxqqxravvapt10h max CPU usage was 673% -- try runtime_constraints "min_cores_per_node":7
183 5 Bryan Cosca
#!! qr1hi-8i9sb-nxqqxravvapt10h max RSS was 1413 MiB -- try runtime_constraints "min_ram_mb_per_node":1945
184
</pre>
185 19 Bryan Cosca
186
~/arvados/tools/crunchstat-summary/bin/crunchstat-summary --format html --job qr1hi-8i9sb-nxqqxravvapt10h > qr1hi-8i9sb-nxqqxravvapt10h.html
187 16 Bryan Cosca
188 1 Bryan Cosca
!62222dc72a51c18c15836796e91f3bc7.png!
189 16 Bryan Cosca
190 5 Bryan Cosca
One thing to point out here is "keep_cache utilization":http://doc.arvados.org/api/schema/Job.html, which can be changed using 'keep_cache_mb_per_task'. You can see keep cache utilization at 99.09%, which means its at a good point. You can try increasing this since it is almost at 100%, but it may not yield significant gains.
191 11 Bryan Cosca
192 1 Bryan Cosca
Another thing to note is to look at the CPU usage and keep transfer rate graphs. You should look to see if they ever mirror each other, which is a sign of a cpu bound job, or an i/o bound job. For example, if keep transfer is low but CPU usage is high, then your job is highly dependent on CPU, which means you should upgrade to a higher core node. If CPU usage is low and keep transfer is high, then you may want to increase the keep_cache_mb_per_task in order to be able to compute on more data.