Hello there
Non-direct buffers exist within the java heap space, and thus may be shuffled around by the garbage collector to reduce fragmentation. OpenGL may try and use data from the buffers that you pass in at any time - if the buffer has been moved since you passed its address to OpenGL then you'll be reading data from an invalid location. Direct buffers are allocated outside of the java heap, and thus do not move around. For opengl use, always allocate buffers like so
FloatBuffer fb = ByteBuffer.allocateDirect( size ).order( ByteBuffer.nativeOrder() ).asFloatBuffer();
Also be aware of the performance issue in http://code.google.com/p/android/issues/detail?id=11078
Ryan
On Jan 6, 5:13 am, kong <zfr... @gmail.com> wrote:
- 显示引用的文字 -举报垃圾内容
Jimmy Jam 查看个人资料 翻译成中文（简体） 更多选项 7月27日, 上午4时19分
I ended up using the syntax specified by RyanMcNally, using "ByteOrder.nativeOrder()" instead of ByteBuffer.nativeOrder()
e.g. sphereVertex = ByteBuffer.allocateDirect( 400000 ).order( ByteOrder.nativeOrder() ).asFloatBuffer();
This was after I found a workable syntax elsewhere with the statements broken out like this (sphereVertex was previously defined as a FloatBuffer):
ByteBuffer vbb = ByteBuffer.allocateDirect(400000); vbb.order(ByteOrder.nativeOrder()); sphereVertex = vbb.asFloatBuffer();
When that worked I took a closer look at the condensed syntax and made the one change to ByteOrder and that worked too.
