Please note: We are aware of an issue affecting replies on the Arm Community forums, which may not be loading as expected.
We apologize for any inconvenience and appreciate your patience while we investigate and work to resolve the issue.
Thank you for your understanding.
I find this in wiki en.wikipedia.org/.../Half-precision_floating-point_format
"ARM alternative half-precision[edit]
ARM processors support (via a floating point control register bit) an "alternative half-precision" format, which does away with the special case for an exponent value of 31 (111112).[9] It is almost identical to the IEEE format, but there is no encoding for infinity or NaNs; instead, an exponent of 31 encodes normalized numbers in the range 65536 to 131008.
"
I want to know that this arm alternative half precision standard also true for mali GPU.More specificlly, does mali gpu half precision have inf and NaN encoded or not.Cause I found in mali G76 GPU where a shader program encounter half precision problem and was suppose to get NaN eventually(the render texture i use is rgba half format),but instead the out put is not NaN.
I will be really greatful if anybody can help.
Mali GPUs have nothing to do with the alternative floating point format you mention here; that only applies to (some) Arm CPUs.
The detail depends on the API, but in general GPUs are not required to conform to IEEE floating point. In particular, NaN generation is allowed, but not usually required. E.g. see this for OpenGL ES:
www.khronos.org/.../GLSL_ES_Specification_3.20.html