First time here? You are looking at the most recent posts. You may also want to check out older archives or the tag cloud. Please leave a comment, ask a question and consider subscribing to the latest posts via RSS. Thank you for visiting! (hide this)

The answer is: it depends. Flash/Flex is faster when it comes to graphic rendering, but Silverlight is faster when it comes to pure computation speed.

There are two speed benchmarks for RIA application, and they both focus on UI rendering:

  • BubbleMark: it displays many bubble floating around the screen as fastest as possible, and use the resulting frame per second as result of the test
  • GUIMark: similar to the BubbleMark, but with more effects and animations used: vectorial images, alpha transparencies, text reflow and so on.

Then another benchmark has been developed, CountingPrimes, which focuses only on pure computational speed (and measures the time taken to compute 10000000 prime numbers).

And here are the results of the test run by the authors of the benchmarks (green the fastest, red the slowest):

UPDATE: Removed the table with the results as it will be an infringement of Silverlight license (not allowed to publish benchmarks of beta software)

From a quick look at the figures it seems that:

I was not able to run all the tests myself, and the version on Silverlight 2 is implemented with Beta 1 and doesn’t run with the beta2 plugin.

Anyone has the time/will to try and optimize the code used in the GuiMark?

kick it on DotNetKicks.com

Technorati Tags: ,,,
posted on Sunday, August 31, 2008 10:14 PM

Comments on this entry:

# re: Which is faster? Flash or Silverlight?

Left by Lorenzo Barbieri at 8/31/2008 10:49 PM

I'm not sure that the beta license allows to publish benchmark results, but perhaps I'm wrong.

# re: Which is faster? Flash or Silverlight?

Left by Thomas Hansen at 9/1/2008 1:45 AM

@Lorenzo
In fact by installing Service Pack 2 for Windows you cannot publish benchmarks for any .Net piece of code. And since Silverlight obviously is the CLR this is probably a breaking of the license terms. However for MSFT to sue this poor guy for publishing the results would be imensily stupid of them and basically would generate an uproar so big it would probably be very damaging to them ;)

# re: Which is faster? Flash or Silverlight?

Left by Simone at 9/1/2008 8:49 AM

@Lorenzo, I'll have a look at the license.
Anyway, I'm just reporting the results published on the respective benchmark's sites.

# re: Which is faster? Flash or Silverlight?

Left by James Curran at 9/2/2008 11:21 PM

MSFT's fussiness about benchmarks for Betas goes back to the earlier 90's. They had just release the first beta of NT (They may still have been calling it OS/2 v3.0 at that time). Anyway, it was a very early release, and what they now refer to as a "checked" build -- lot's & lot's of asserts, trace messages, and a non-optimized compile.

Naturally, it ran rather slow. Which got the trade press all excited, 'cuz they could run articles comparing how slow it was to Windows 3.0.

After that, MS learned their lesson -- they create an optimized build specifically for each beta release, and include the embargo on benchmarks on all beta EULAs.

# re: Which is faster? Flash or Silverlight?

Left by Matthew Fabb at 9/3/2008 7:27 PM

I wouldn't necessary say that Silverlight is 4 times faster than Flash when it comes to pure computational power, just that it's 4 times faster when it comes to counting primes. Because I imagine that different type of math equations would result in different speed differences.

A better benchmark would test different type of equations, as well as other functions and operations. Example, here's link to an old benchmark of JavaScript versus Java versus ActionScript (www.oddhammer.com/actionscriptperformance/set4/). A benchmark like that for Flash versus Silverlight would be great. You could then average it out for a speed difference or just say y is x times faster than y when it comes to say rounding numbers.

For Flash there's even optimizations (http://osflash.org/as3_speed_optimizations) where speed changes depending how an equation is set up. Example multiplying by 0.5 is faster than diving by 2, despite the fact you end up with the same results. These examples I think makes it impossible to prove speed by just one type of equation.

Also as others have mentioned, the results will mean more with the RTM version of Silverlight 2.0, which will likely include a lot more optimizations. By that time, the tests can also be done against Flash Player 10, instead of Flash 9, as it looks like 10 will be out shortly (a month or two?).

# re: Which is faster? Flash or Silverlight?

Left by Pete at 9/18/2008 3:55 PM

Speed is one issue. Will the person running the browser download silverlight run package, is a better question. Sure we are all geeks ready to try the new software; however, most of the people that hit my site are fearful of downloading a virus, so many would not download silverlight. So I removed the silverlight and replaced it jquery.

Comments have been closed on this topic.