jit.gl.pix float64 precision?

Regis Planque's icon

Regis Planque

9月 19 2024 | 8:00 午後

Is there a way to force jit.gl.pix to work with float64 instead of float32?

Like jit.pix @precision float64

TFL's icon

TFL

9月 19 2024 | 9:29 午後

@type float64
But maybe you want to set that attribute beforehand, on whatever object you use to feed your jit.gl.pix.

Regis Planque's icon

Regis Planque

9月 19 2024 | 9:43 午後

According to jit.gl.pix Reference, float64 is not in the possible values of @type:

@type [symbol]

The output texture data type (default = auto). When type is auto the type is set by the input texture or matrix type. Possible values:

'auto' ( Adapt to input data type )
'char' ( 8bit unsigned byte )
'long' ( 32bit integer )
'half' ( 16bit float )
'float' ( 32bit float )
'float16' ( 16bit float )
'float32' ( 32bit float )

TFL's icon

TFL

9月 20 2024 | 7:15 午前

Oh my bad you're right!

This said, according to the openGL doc, double (other name for float64) precision is only available since OpenGL 4.0, while Max is based on 3.

Regis Planque's icon

Regis Planque

9月 20 2024 | 7:20 午前

Okay.... Too bad.

Thanks