<?php $hms = gmdate('H:i:s',$seconds); ?>

Pretty sweet, no? The trick here is to tell PHP your date is in reference to the Unix epoch, and make sure you do it in GMT (a la gmdate) so it doesn’t try to take your time zone into account.  This example works for up to 24 hours, so if you need more than that, consider the following two examples:

Allowing the function to display days:

<?php $hms = gmdate('d:H:i:s',$seconds); ?>

Allowing the function to display more than 24 hours:

<?php $hms = floor($seconds/3600).gmdate(':i:s',$seconds); ?>

You can change the format as well, per the normal php date() function’s format.  Enjoy!

While tooling around on Gizmodo, I came across a  large banner ad for Internet Explorer that boasted its HTML5 performance over its able-bodied counterparts.  From my extensive experience, I knew there was something a bit FishIE right off the bat.  For once, I clicked on the ad and headed to their site for further investigation.  What resulted angered me and sent me on a day-long tangent to expose MS for their dishonesty and their continuing vendetta against every user of the internet.

The site in question was a set of Microsoft-designed examples designed to prove why HTML5 and IE9 beta were a beautiful pairing, and coaxed me to download the beta and “become a part of the future”.  Sounds great, I’m game.  After the 30-minute download, install, restart, upgrade, upgrade, and two restarts, I was finally presented with the most–WTF?? Microsoft’s version of Chrome?  Surely they jest….however that’s not the point of this article.

What really upset me was the performance of IE9 versus my dear Chrome 8 Stable in the HTML5 tests.  As we’ve come to accept almost unequivocably, is that every facet of Chrome’s performance exceeds IE9′s, and generally by a large margin. The numbers don’t lie–I compiled the stats for Microsoft’s tests into the following table (see system specs below):

Now an advantage to IE9 is not to be unexpected, since these pages were specifically designed to offer IE an advantage in speed and stability. The problem, however, is that Chrome not only didn’t win, it was the absolute slowest across the board.  Not only is IE faster in this example, it’s between 4 and 150 times faster.  Where the test really started to look dubious to me was in the following metrics for the Speed Reading Frame Rates and Redraws:

The redraws, as annotated by the example, show that Chrome and Firefox 4, despite their lower framerates, are rendering over 18 times as many frames as IE.  Veeeeery suspicious.  In addition to this, I captured the following screenshot of the test while it was running:

IE9 64-bit Speed Read HTML5 TestClick to enlarge.

Now no browser that I have ever seen has had the ability to actually fade/motion blur different image elements into each other unless it’s specifically programmed to do so, and this behavior was most definitely not exhibited by any of the other browsers for this test.  I’m not sure if some sort of logic was included to enable this feature, but in two very large ways, Internet Explorer is behaving differently from both Chrome and Firefox.  If anyone can offer some insight into this, I would appreciate hearing about it.

Moving on.  For the sake of academics, I ran a few common compliance tests and benchmarks found around the internet:

Ah, now that’s more like it… The only thing surprising here is that IE9 64-bit appears to be slower than its x86 counterpart…

EDIT: as it was pointed out, Microsoft, for some reason has only implemented a native 32-bit JS interpreter, which is kiiiinda stupid.  DOM parsing/rendering is not so fast that JS performance becomes insignificant, so I’m not sure what they were thinking there.

Regardless, Chrome is between 100% and 300% faster in the performance-based benchmarks.  All tests were performed on the same Intel Core2 Quad Q6600 @ 2.70GHz, 4GB DDR2 RAM, ATi Radeon 2600XT with a maximized window running at 1280×1024, so there was no system advantage across tests.

What’s my point??  It seems a bit strange how all Microsofts HTML5 speed tests seem to be cases of selectively inefficient scripting.  If they were done properly, simple examples like the Santa’s helper test wouldn’t be so incredibly dissonant.  All that IE9 seems to have going for it is purely the hardware acceleration for heavy canvas graphics crunching.  It clearly has the advantage with alpha-blending, but this really isn’t a fair fight, considering it’s using hardware acceleration while the other two are not.  In practice though, Chrome and FF still blow IE9 away as far is performance is concerned.  They also have that unique advantage of being able to render webpages correctly, which probably shouldn’t be ignored.

I plan on looking into the scripting to find out why the results are so incredibly different, so stay tuned for a follow-up post.