LWJGL 2.5 Released

[Luy]

I just found the following paper: http://algo2.iti.kit.edu/wassenberg/timing/timing_pitfalls.pdf
If I get it right, there is a reliable timer called High-Precision Event Timer (HPET) on current hardware and QPC uses that from Vista on.

[Matzon]

I just found the following paper: http://algo2.iti.kit.edu/wassenberg/timing/timing_pitfalls.pdf
If I get it right, there is a reliable timer called High-Precision Event Timer (HPET) on current hardware and QPC uses that from Vista on.

http://en.wikipedia.org/wiki/High_Precision_Event_Timer:

The following operating systems are known not to be able to use HPET: Windows XP,[note 2] Windows Server 2003, and earlier Windows versions, Linux kernels prior to 2.6.[note 3]

The following operating systems are known to be able to use HPET: Windows Vista, Windows 2008, Windows 7, x86 based versions of Mac OS X, Linux operating systems using the 2.6 kernel, FreeBSD and OpenSolaris.

Before we can implement a workaround, we need a reliable way of determining if a timer is "wrong" ... ?

[Luy]

Before we can implement a workaround, we need a reliable way of determining if a timer is "wrong" ... ?

Sure. I'm no expert, but some googling revealed that on linux information about the used timers can be found in /sys/devices/system/clocksource/clocksource0/ or you can try to directly access the HPET as described here: http://blog.fpmurphy.com/2009/07/linux-hpet-support.html

On Windows Vista/7 a device driver is installed if HPET is available, so you could probably check that. See e.g. http://forum.scottmueller.com/viewtopic.php?f=1&t=1284

[theagentd]

In NVFramebufferMultisampleCoverage, the variable GL_MULTISAMPLE_COVERAGE_MODES_NV is spelled GL_MULITSAMPLE_COVERAGE_MODES_NV. Unless mulITsampling is a new function, it's spelled incorrectly. xd

[jouvieje]

ok changes made.

If you have any other remarks do point them out, always trying to make it better

Thanks  

One thing I could suggest as an improvement. It could be great if the applet loader could download and extract in parallel, instead of extracting everything at the end. This could probably speed a little more the loading (of the first run), which is always great.

[wondersonic]

Hi,
Again a great release!   

• General: Added support for disabling runtime function checks, buffer checks and state tracking. Activated with -Dorg.lwjgl.util.NoChecks=true

I've dreamt of this point since a long time.

Notice: We'd like to remind people to include the copyright, conditions and disclaimer statement for LWJGL in their products, as required by the license. Though we are not about to claim foul in any way, it would be nice to see a link back to lwjgl.org in the credits or documentation at the very minimum.

Coming to my mind another idea: is there a 3D version of the LWJGL logo available somewhere? 

Again thanks!
WS

[Simon Felix]

Hi,
Again a great release!   

• General: Added support for disabling runtime function checks, buffer checks and state tracking. Activated with -Dorg.lwjgl.util.NoChecks=true

I've dreamt of this point since a long time.

Yeah, me too. I tried disabling the checks but found it made no difference... Could you measure the performance impact of the runtime checks?

[wondersonic]

Hi,
Again a great release!   

• General: Added support for disabling runtime function checks, buffer checks and state tracking. Activated with -Dorg.lwjgl.util.NoChecks=true

I've dreamt of this point since a long time.

Yeah, me too. I tried disabling the checks but found it made no difference... Could you measure the performance impact of the runtime checks?

I guess I didn't make performance tests since 3 months but at the time these checks costed around 10% of the time.