Emacs benchmarking suite

Emacs benchmarking suite

[Dhavan Vaidya]

[-- Attachment #1: Type: text/plain, Size: 1201 bytes --]

Hello everyone!

This is my first every message to a mailing list so I have no idea how
this works. Pointing out mistakes is welcome, I will learn.

On the topic, Sacha interviewed Johnw recently in which he mentioned
the need of having a benchmarking/profiling system for emacs. He says
the performance has been decreasing over time.

Upon listening to this, I contacted John and showed my interest in
putting such a system in place. He happily agreed to help me with it.

During this time, I switched job and migrated to a different place.
This took 2 months for things settle back to normal.

Yesterday, I talked with John again on the subject on #emacs and he
said nothing has happened for this so far.

I am interested in this project, and would love to contribute.

There are a few issues however. First is, I have never learned any of
the Lisp language. I have just purchased SICP and going to start it.
Moreover, I have never worked with profiling systems either and know
nothing about them.

I want to learn and contribute to the dearest editor/interpreter.

Share your thoughts on this, give me guidelines and help would be much
appreciated!

-- 
- D Vaidya.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: Emacs benchmarking suite

[Marcin Borkowski]

On 2016-03-19, at 11:53, Dhavan Vaidya <quark@codingquark.com> wrote:

> Hello everyone!

Hi!

> There are a few issues however. First is, I have never learned any of
> the Lisp language. I have just purchased SICP and going to start it.

I'd strongly recommend Robert J. Chassell's book, "Introduction to
Programming in Emacs Lisp", which you can read in Emacs.  Type

C-h i d m emacs lisp intro RET

or evaluate the form

(info "(eintr) Top")

Hth,

-- 
Marcin Borkowski
http://octd.wmi.amu.edu.pl/en/Marcin_Borkowski
Faculty of Mathematics and Computer Science
Adam Mickiewicz University

Re: Emacs benchmarking suite

[Dhavan Vaidya]

[-- Attachment #1: Type: text/plain, Size: 634 bytes --]

On Sat, 2016-03-19 at 12:57 +0100, Marcin Borkowski wrote:
> I'd strongly recommend Robert J. Chassell's book, "Introduction to
> Programming in Emacs Lisp", which you can read in Emacs.  Type
> 
> C-h i d m emacs lisp intro RET
> 
> or evaluate the form
> 
> (info "(eintr) Top")

Well, I thought it'll be better to learn Lisp itself first, understand
the paradigm behind it properly and then go to eintr. I have read some
pages of eintr before, and was not exactly following it. It used to be
weird - I wouldn't get it while reading it, and after a few hours, all
of a sudden it made sense.

 
-- 
- D Vaidya.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

RE: Emacs benchmarking suite

[Drew Adams]

> It used to be weird - I wouldn't get it while reading it, and after
> a few hours, all of a sudden it made sense.

Funny about that.  That's the way something new hits us sometimes.

We rarely start from a blank slate, and our initial understanding
can only be modified, built upon, and adapted, never really ignored
or escaped.

The result can be that when a quite-different view is grasped, that
seems to happen in a quantum leap and does not seem to be the result
of a logical or historical progression.  Paradigm shift; cf. ye olde
quantity-leaps-into-quality thingie.  We often cannot see any bridge
that got us from there to here.  We see only a discontinuity.

Recursion, higher-order functions, abstract data types, infinite
data structures, unification,... can be just such weird eye-openers.
The first time I saw a declarative (e.g. Lisp/Prolog) definition of
`append' I'm sure I stared at it like a deer in the headlights for
a few minutes, before it registered.  Or maybe, as you describe, it
suddenly hit me later.

Like staring at a SIRDS for the first time, trying hard to see
something 3D in it to no avail, and then drifting off, paying less
attention, staring past it vaguely, and suddenly having a 3D
dinosaur pop out of the page unexpectedly.

Enjoy.

---

"It will flame out, like shining from shook foil"
- G.M. Hopkins, 1877 http://www.poetryfoundation.org/poem/173660

Re: Emacs benchmarking suite

[John Wiegley]

>>>>> Dhavan Vaidya <quark@codingquark.com> writes:

> On the topic, Sacha interviewed Johnw recently in which he mentioned the
> need of having a benchmarking/profiling system for emacs. He says the
> performance has been decreasing over time.

Hi Dhavan!

I should clarify: John has the perception that it has been decreasing, but
this may not at all be the case. Hence the need for measurement. :)

> Share your thoughts on this, give me guidelines and help would be much
> appreciated!

Investing some time in learning a bit of Lisp is a good idea; even better will
be to find a buddy/collaborator who can help you to architect a benchmarking
framework, ideally based on work that has already been done for this sort of
thing in other language.

We want to be able to measure resource usage of Lisp code, to compare this
usage against previous runs (say, build to build), and a way of charting the
progression over time. This way, a dev who builds Emacs frequently can watch
for trends over time.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2