Performance Mode is a Windows option that allows one to tune the platform for power vs performance on the graphics. The assumption by Microsoft is that using the GPU is higher power, but in most cases the enegery required by the slower handling of graphics by the CPU vs the GPU actually is higher than leaving the GPU acceleration enabled. That said, if something is able to be h/w accelerated, and you want that to occur, you should leave the driver in Performance Mode. There is some accelerations around compression, and if the ones you are using are listed either in the User Guide (in the appendix) or in the TPS if you have IBL access.
Hope this helps.
Thank you Kirk but i didn't understand your answer Probably translating problem.. Were you trying to explain this?
Performance mode ON: Some elements hardware accelerated
off: some elements software acc.(cpu decoding=more power consumption)
I Read Appendix in user guide but there isn't nothing about texture compression (if it possible to enable it). It's only described,really quickly, Opengl and DX api :\
Yes I was explaining that the "Performance Mode" is feature of Windows that we support per the way Microsoft put it in the O/S as a way to disable hardware acceleration for whatever reason.
We accelerate much of the OGL and DX stacks but there are some things that are not and therefore always run in the CPU (slower and sometimes higher power required to do becuase it takes longer AND the CPU is running at high usage in those cases.
Not at the driver level.
There are LOTS of things you can do to tune the application, and it is possible there are options "above" the driver at the OS level that allows Windows to change how it handles it but that is more of a question for Microsoft as we do not control that functionality. The driver implements a hardware accelerated interface and it is up to the OS and application to decide if it makes sense to use it.
Generally speaking, there are cases where the overhead of setting up the hardware to do the acceleration will be higher than if you had just done the function in software. We see that a LOT in benchmarks where they try to measure a complex operation using a simple use case and do not show the :"real world" benefits of using the hardware. 3DMark is a GOOD example of a real world benchmark. Other benchmarks we find lack any resemblance to actual work loads.
Hope this helps.
Kirk thanks for all but what do you mean with a lot of things to accelerate 3D apps by OS? Try to make a little list for me and i will test to myself. With us15 the only tips that i use are,in os, no paging file, fsb overclock,hyperthreading,high performance mode,performance mode on,edit 3d app properties and stop. All increase the gpu benchmark score but i belive that i can do something more
Hi, Razak. Our apologies for a delay in response time. The holidays always seem to get in the way of work, don't they? Thanks for your participation in the Embedded Community and we hope to hear from you throughout 2014. Let me see what I can do about getting an answer for you. Thanks! LynnZ.
Since I do not know your application, a list would not make any sense- the one you provide is a good start.
Because the GMA500 software support is EOL, I am having trouble finding documentation to determine is S3TC is supposed to be supported. I see in the XP/Linux EMGD driver that it is NOT supported so that would seem to indicate that it is not supported in the GMA EMGD for Win7. If you do a capabilites query, doe it indicate supported or not??
Sorry we cannot be more help.