Re: Performance penalties of in-Perl docn vs compiled CGIs.

I'm wondering what performance penalty I'm paying for

This will depend to some extent on your environment and therefore the only way to know for sure is to benchmark it.

An obvious solution would be to compile the Perl into e.g. C executables

I might humbly suggest that an even more obvious solution would be to enable back-end persistence via mod_perl or FCGI or some other proxy.

The other easy option is to split the POD from the code and store it elsewhere.

🦛

Comment on Re: Performance penalties of in-Perl docn vs compiled CGIs.

Replies are listed 'Best First'.
Re^2: Performance penalties of in-Perl docn vs compiled CGIs. by phirun (Novice) on Feb 02, 2021 at 11:08 UTC
> The other easy option is to split the POD from the code and store it elsewhere. Yes, that's something I've often pondered, but you end up with two versions and the ensuing hassle. Software is just a side-dabble for me, so my knowledge of real-world issues is limited, esp when a team is involved. The popularity of other interpreted languages for CGI - PHP for example - and apparent popularity of attempted Perl compilers some years back has always made me curious as to any real-world benefits of such measures. In the modern world of huge RAM and lightning-fast CPUs there's probably none, but a knowledgable comment or two from someone with experience in that area would slake decades of merely wondering.	[reply]

Replies are listed 'Best First'.

Re^2: Performance penalties of in-Perl docn vs compiled CGIs.
by phirun (Novice) on Feb 02, 2021 at 11:08 UTC

> The other easy option is to split the POD from the code and store it elsewhere. Yes, that's something I've often pondered, but you end up with two versions and the ensuing hassle. Software is just a side-dabble for me, so my knowledge of real-world issues is limited, esp when a team is involved. The popularity of other interpreted languages for CGI - PHP for example - and apparent popularity of attempted Perl compilers some years back has always made me curious as to any real-world benefits of such measures. In the modern world of huge RAM and lightning-fast CPUs there's probably none, but a knowledgable comment or two from someone with experience in that area would slake decades of merely wondering.

[reply]