GSoC 2018: Sorting Algorithm and Benchmarking

Started by Kefan Yangalmost 8 years ago8 messageshackers
Jump to latest
#1Kefan Yang
starordust@gmail.com

Hello,

My name is Kefan Yang. I am so excited that my proposal ‘Sorting Algorithm and Benchmarking 2018’ has been accepted. I see on the Wiki page that you can mentor this proposal. This is what I’ve done in these few days:
1. I’ve carefully read through Submit a Patch and have a basic understanding of the submission process.
2. I’ve set up the environment and ready to code. The sorting routine is kind of separated from other modules so I don’t need much time to get familiar with the code base. If things go smoothly, I can start coding ahead of schedule
Now I have some questions:
1. If I understand it correctly, the sorting benchmark should be an executable under the src/bin/ folder just like pgbench?
2. Do you, or other community members, have any suggestions about the final version of proposal?
3. Are there any specific thing you expect me to do during this community bonding period?
Thanks for your time!

Regards,
Kefan

Attachments:

proposal_gsoc2018.pdfapplication/pdf; name=proposal_gsoc2018.pdfDownload+2-2
#2Robert Haas
robertmhaas@gmail.com
In reply to: Kefan Yang (#1)
Re: GSoC 2018: Sorting Algorithm and Benchmarking

On Wed, Apr 25, 2018 at 3:12 PM, Kefan Yang <starordust@gmail.com> wrote:

If I understand it correctly, the sorting benchmark should be an executable
under the src/bin/ folder just like pgbench?

What would this executable do, exactly?

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

#3Andrey Borodin
amborodin@acm.org
In reply to: Kefan Yang (#1)
Re: GSoC 2018: Sorting Algorithm and Benchmarking

Hi!

26 апр. 2018 г., в 0:12, Kefan Yang <starordust@gmail.com> написал(а):

My name is Kefan Yang. I am so excited that my proposal ‘Sorting Algorithm and Benchmarking 2018’ has been accepted.

Welcome! I'm also glad you've chosen this project.

I see on the Wiki page that you can mentor this proposal.

Yes, we with Atri Sharma will mentor this project.

This is what I’ve done in these few days:
• I’ve carefully read through Submit a Patch and have a basic understanding of the submission process.

Cool!

• I’ve set up the environment and ready to code. The sorting routine is kind of separated from other modules so I don’t need much time to get familiar with the code base. If things go smoothly, I can start coding ahead of schedule

Great!

Now I have some questions:
• If I understand it correctly, the sorting benchmark should be an executable under the src/bin/ folder just like pgbench?

Well, I think it is up to you how to make reproducible, precise and correct benchmarking :)
You can make exec (we will not have to merge it into PG, we need just results), or you can make Postgres extension (it is much easier) which will be executed from psql.

• Do you, or other community members, have any suggestions about the final version of proposal?

As far as I know, proposal cannot be modified. Actual work can be slightly adjusted though.

• Are there any specific thing you expect me to do during this community bonding period?

Let's establish communication. You can also build PostgreSQL from git and try pgbench (tool for benchmarking overall performance). Read something about TPC-(A,B,C) and YCSB, and may be even read something about sysbench.
I can think of following mediums of communications:
1. Slack channel
2. Telegram chat
3. Skype chat
4. E-mail

Atri, Kefar, how do you think, which of these will do best for us?

Best re

#4Kefan Yang
starordust@gmail.com
In reply to: Robert Haas (#2)
RE: GSoC 2018: Sorting Algorithm and Benchmarking

Hey,

This executable should read a config file (path to sorting routines, test data size etc.), perform tests on different sorting routines, and output the result.

Kefan

From: Robert Haas
Sent: April 26, 2018 11:53 AM
To: Kefan Yang
Cc: Andrey Borodin; PostgreSQL Hackers
Subject: Re: GSoC 2018: Sorting Algorithm and Benchmarking

On Wed, Apr 25, 2018 at 3:12 PM, Kefan Yang <starordust@gmail.com> wrote:

If I understand it correctly, the sorting benchmark should be an executable
under the src/bin/ folder just like pgbench?

What would this executable do, exactly?

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

#5Kefan Yang
starordust@gmail.com
In reply to: Andrey Borodin (#3)
RE: GSoC 2018: Sorting Algorithm and Benchmarking

Thank you for the feedback!

I think Slack would be great for our communication

From: Andrey Borodin
Sent: April 26, 2018 11:38 PM
To: Kefan Yang
Cc: PostgreSQL Hackers; Atri Sharma
Subject: Re: GSoC 2018: Sorting Algorithm and Benchmarking

Hi!

26 апр. 2018 г., в 0:12, Kefan Yang <starordust@gmail.com> написал(а):

My name is Kefan Yang. I am so excited that my proposal ‘Sorting Algorithm and Benchmarking 2018’ has been accepted.

Welcome! I'm also glad you've chosen this project.

I see on the Wiki page that you can mentor this proposal.

Yes, we with Atri Sharma will mentor this project.

This is what I’ve done in these few days:
• I’ve carefully read through Submit a Patch and have a basic understanding of the submission process.

Cool!

• I’ve set up the environment and ready to code. The sorting routine is kind of separated from other modules so I don’t need much time to get familiar with the code base. If things go smoothly, I can start coding ahead of schedule

Great!

Now I have some questions:
• If I understand it correctly, the sorting benchmark should be an executable under the src/bin/ folder just like pgbench?

Well, I think it is up to you how to make reproducible, precise and correct benchmarking :)
You can make exec (we will not have to merge it into PG, we need just results), or you can make Postgres extension (it is much easier) which will be executed from psql.

• Do you, or other community members, have any suggestions about the final version of proposal?

As far as I know, proposal cannot be modified. Actual work can be slightly adjusted though.

• Are there any specific thing you expect me to do during this community bonding period?

Let's establish communication. You can also build PostgreSQL from git and try pgbench (tool for benchmarking overall performance). Read something about TPC-(A,B,C) and YCSB, and may be even read something about sysbench.
I can think of following mediums of communications:
1. Slack channel
2. Telegram chat
3. Skype chat
4. E-mail

Atri, Kefar, how do you think, which of these will do best for us?

Best re

#6Andrey Borodin
amborodin@acm.org
In reply to: Kefan Yang (#5)
Re: GSoC 2018: Sorting Algorithm and Benchmarking

27 апр. 2018 г., в 23:23, Kefan Yang <starordust@gmail.com> написал(а):

I think Slack would be great for our communication

Here's Slack invite app http://postgres-slack.herokuapp.com/
Should we create a channel for this project?

Best regards, Andrey Borodin.

#7Atri Sharma
atri.jiit@gmail.com
In reply to: Andrey Borodin (#6)
Re: GSoC 2018: Sorting Algorithm and Benchmarking

+1 For Slack

On Sat, 28 Apr 2018, 09:52 Andrey Borodin, <x4mmm@yandex-team.ru> wrote:

Show quoted text

27 апр. 2018 г., в 23:23, Kefan Yang <starordust@gmail.com> написал(а):

I think Slack would be great for our communication

Here's Slack invite app http://postgres-slack.herokuapp.com/
Should we create a channel for this project?

Best regards, Andrey Borodin.

#8Kefan Yang
starordust@gmail.com
In reply to: Andrey Borodin (#6)
RE: GSoC 2018: Sorting Algorithm and Benchmarking

Hey,

I’ve joined the workspace. The username is Kefan Yang
I agree we can create a channel for this project

From: Andrey Borodin
Sent: April 27, 2018 9:22 PM
To: Kefan Yang
Cc: PostgreSQL Hackers; Atri Sharma
Subject: Re: GSoC 2018: Sorting Algorithm and Benchmarking

27 апр. 2018 г., в 23:23, Kefan Yang <starordust@gmail.com> написал(а):

I think Slack would be great for our communication

Here's Slack invite app http://postgres-slack.herokuapp.com/
Should we create a channel for this project?

Best regards, Andrey Borodin.