«

Hello,

Just curious as to whether or not anyone here knows which is faster between
SSI and PHP? My business partner and I are thinking of using either of them
to include files (ie. menus, headers, footers) as part of templates. My mate
wants to use PHP and I'm leaning towards SSI. I've searched google groups
for an answer but found conflicting responses... perhaps it depends on the
application?

I thought perhaps SSI would be faster because it's a local Apache module and
PHP is a third party interpreter... any comments?

Thanks in advance,

Nick.

e n | c k m a wrote:
> Just curious as to whether or not anyone here knows which is faster between
> SSI and PHP?

Just tested a simple file with one include in PHP and SSI over 10,000
requests with ApacheBench. The mean average for time per request was
1.207 milloseconds for PHP, and 0.861 for SSI. Of course, if your
application needs to do things SSI can't, you'll want to use something
more complex like PHP.

"Leif K-Brooks" <> wrote in message
news:...
> e n | c k m a wrote:
> > Just curious as to whether or not anyone here knows which is faster
between
> > SSI and PHP?
>
> Just tested a simple file with one include in PHP and SSI over 10,000
> requests with ApacheBench. The mean average for time per request was
> 1.207 milloseconds for PHP, and 0.861 for SSI. Of course, if your
> application needs to do things SSI can't, you'll want to use something
> more complex like PHP.

So, SSI is .35 milliseconds faster. Compare this to the tens or even
hundreds of milliseconds it takes to get the results back to the client. I
think the difference is irrelevant

--
Cheers
Richard.

rf wrote:
> So, SSI is .35 milliseconds faster. Compare this to the tens or even
> hundreds of milliseconds it takes to get the results back to the client. I
> think the difference is irrelevant

Yeah, but some of can't sleep at night because our application is .01
milliseconds too slow.

rf wrote:

> "Leif K-Brooks" <> wrote in message
> news:...
>
>>e n | c k m a wrote:
>>
>>>Just curious as to whether or not anyone here knows which is faster
>
> between
>
>>>SSI and PHP?
>>
>>Just tested a simple file with one include in PHP and SSI over 10,000
>>requests with ApacheBench. The mean average for time per request was
>>1.207 milloseconds for PHP, and 0.861 for SSI. Of course, if your
>>application needs to do things SSI can't, you'll want to use something
>>more complex like PHP.
>
>
> So, SSI is .35 milliseconds faster. Compare this to the tens or even
> hundreds of milliseconds it takes to get the results back to the client. I
> think the difference is irrelevant
>

10,000 times .35 is 3500 seconds of server CPU time. That can be
important in some circumstances. My opinion is also that SSI will be
faster than PHP at a raw level. However, the above test is
inconclusive because does not explain whether or not server-side
includes processing was turned off when testing PHP, and if PHP was
turned off when testing SSI. That could make a difference. Also,
there are engines, particularly Zend, which cache PHP output files.
So the particular configuration of your server will highly affect your
results. In any case, switching your include statements between SSI
and PHP is a simple find/replace if you do it right. So just pick
one, and once you build your site, you can evaluate which is faster.
I think processing time for include files is not going to be your
biggest concern.

Shailesh Humbad wrote:
> rf wrote:

> > So, SSI is .35 milliseconds faster. Compare this to the tens or even
> > hundreds of milliseconds it takes to get the results back to the client.
I
> > think the difference is irrelevant
> >
>
> 10,000 times .35 is 3500 seconds of server CPU time.

Last time I looked 10,000 times .35 *milliseconds* is 3.5 seconds

Insignificant, given that even with a 100MHz nic, and assuming 47000(*)
bytes per hit (including images), it would take 47 seconds just to ship the
data out, ignoring all other network considerations.

(*) plug in your average page size here.

There are threads happenning at the moment that suggest that doing any PHP
(or SSI) will defeat browser/proxy caching, unless carefull attention is
taken with the headers. Eliminating server side processing for just 10% of
your pages (thus enabling caching) would have a far more dramatic effect on
the server than 3.5 seconds of CPU time.

--
Cheers
Richard.

rf wrote
> Shailesh Humbad wrote:
> > rf wrote:
>
> > > So, SSI is .35 milliseconds faster. Compare this to the tens or even
> > > hundreds of milliseconds it takes to get the results back to the
client.
> I
> > > think the difference is irrelevant
> > >
> >
> > 10,000 times .35 is 3500 seconds of server CPU time.
>
> Last time I looked 10,000 times .35 *milliseconds* is 3.5 seconds
>
> Insignificant, given that even with a 100MHz nic, and assuming 47000(*)
> bytes per hit (including images), it would take 47 seconds just to ship
the
> data out, ignoring all other network considerations.
>
> (*) plug in your average page size here.

Damn, forgot to add this bit:

Consider a chain. You should be worrying real hard about the weakest link in
that chain. You should ignore the bits that have little or no impact on the
performance of the chain.

I once found one of my progammers had spent half a day optimizing a piece of
code. Did well, got it down from one second to half a second CPU time. The
trouble was this peice of code executed once in the initialization section
of a program that ran for over 5 hours. An overall saving of .0027%. We had
a conversation about this

--
Cheers
Richard.

rf <rf@.invalid> wrote:

> There are threads happenning at the moment that suggest that doing any PHP
> (or SSI) will defeat browser/proxy caching, unless carefull attention is
> taken with the headers.

To ensure that you don't break caching, use XBitHack and a little RTFM
about setting permissions. Works just fine.

--
Art Sackett,
Patron Saint of Drunken Fornication

e n | c k m a wrote:
> Hello,
>
> Just curious as to whether or not anyone here knows which is faster between
> SSI and PHP? My business partner and I are thinking of using either of them
> to include files (ie. menus, headers, footers) as part of templates. My mate
> wants to use PHP and I'm leaning towards SSI. I've searched google groups
> for an answer but found conflicting responses... perhaps it depends on the
> application?
>
> I thought perhaps SSI would be faster because it's a local Apache module and
> PHP is a third party interpreter... any comments?


SSI is part of PHP

Weyoun the Dancing Borg

> SSI is part of PHP

Er, what?

What leads you to this conclusion? SSI was around way before PHP was
written.

--
Cheers
Richard.