I’m having a bit of a problem when using a single channel 32bit texture, more exactly: EPixelFormat::PF_R32_UINT
It seems to act very weird: when the values are half of a 32bit max value, it’ goes all black in the material, when going higher it goes and over over bright, are all the 32bits used to build a color, or is there something im missing?
Did you write some code to expose it? since AFAIK there is no exposed 32 bit INT format for assets, only float 16bit.
how are you trying to use the values? You cant really display a 32 bit value as output as most of the output formats won’t support it. you can force many to be 32bit using ‘gbufferformat 5’ at the console to see if it has any effect. Its possible that some intermediate format is reading a bit pattern from the 32bit as NaNs or something since the 32bit mantissa will cover the whole exponent of the 16 bit values (making it easy to encode all 0’s which the GPU may flag as nan).
You can do calculations in 32 bit but shouldn’t try to return a value that large.
What exactly are you trying to do, just store int values? If so maybe bit shifting is your friend but there are some limitations there.