defun function call count

[Tatsuya Tsuda]

I have counted all of the defun function call of Shen-C from program start to test suite end.

Below is the result.

https://gist.github.com/otabat/0ffb06fb7517fcd11906086fcc511bee

This was done by natively injecting counting codes at the beginning of the defun body.

You can use this data and measure the function call time to improve your Shen implementation.

Tatsuya Tsuda

[deech]

Mother of god! I replaced `element?` with a native implementation and the elisp test suite went from 45 -> 34 seconds! Thank you!

[Neal Alexander]

Looks to be mostly shen Prolog stuff.

I suspect my port would benefit if Shen had a separate mechanism for handling very large vectors manually via mmap/mremap.

[Bruno Deferrari]

On Thu, Apr 27, 2017 at 4:23 PM, deech <aditya...@gmail.com> wrote:

Mother of god! I replaced `element?` with a native implementation and the elisp test suite went from 45 -> 34 seconds! Thank you!

Be careful, if the underlying's platform equality operator does not behave in the same way as Shen's `=` function you may be changing behavior in some (rare?) cases. It is easy to solve anyway, all you have to do is to take a fast path and use the native version when possible, and fall back to Shen's `element?` implementation otherwise.

Added a link to that gist in the wikii: https://github.com/Shen-Language/wiki/wiki/PortDeveloperNotes#amount-of-time-system-functions-are-called

 

On Thursday, April 27, 2017 at 2:03:42 PM UTC-5, Tatsuya Tsuda wrote:

I have counted all of the defun function call of Shen-C from program start to test suite end.

Below is the result.

https://gist.github.com/otabat/0ffb06fb7517fcd11906086fcc511bee

This was done by natively injecting counting codes at the beginning of the defun body.

You can use this data and measure the function call time to improve your Shen implementation.

Tatsuya Tsuda

--
You received this message because you are subscribed to the Google Groups "Shen" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qilang+unsubscribe@googlegroups.com.
To post to this group, send email to qil...@googlegroups.com.
Visit this group at https://groups.google.com/group/qilang.
For more options, visit https://groups.google.com/d/optout.

--

BD

[deech]

Accounted for that. Still getting the perf. increase!

On Thursday, April 27, 2017 at 2:36:19 PM UTC-5, Bruno Deferrari wrote:

On Thu, Apr 27, 2017 at 4:23 PM, deech <aditya...@gmail.com> wrote:

Mother of god! I replaced `element?` with a native implementation and the elisp test suite went from 45 -> 34 seconds! Thank you!

Be careful, if the underlying's platform equality operator does not behave in the same way as Shen's `=` function you may be changing behavior in some (rare?) cases. It is easy to solve anyway, all you have to do is to take a fast path and use the native version when possible, and fall back to Shen's `element?` implementation otherwise.

Added a link to that gist in the wikii: https://github.com/Shen-Language/wiki/wiki/PortDeveloperNotes#amount-of-time-system-functions-are-called

 

On Thursday, April 27, 2017 at 2:03:42 PM UTC-5, Tatsuya Tsuda wrote:

I have counted all of the defun function call of Shen-C from program start to test suite end.

Below is the result.

https://gist.github.com/otabat/0ffb06fb7517fcd11906086fcc511bee

This was done by natively injecting counting codes at the beginning of the defun body.

You can use this data and measure the function call time to improve your Shen implementation.

Tatsuya Tsuda

--
You received this message because you are subscribed to the Google Groups "Shen" group.

To unsubscribe from this group and stop receiving emails from it, send an email to qilang+un...@googlegroups.com.

To post to this group, send email to qil...@googlegroups.com.
Visit this group at https://groups.google.com/group/qilang.
For more options, visit https://groups.google.com/d/optout.

--

BD