I have series of compute operations on middle of render frame.
Command buffer 1 - G render encoder
Command buffer 2 - compute encoder with 3 computes <- here is my compute in seq.
Command buffer 3 - another encoder and final pass
On discret GPU compute takes ~4ms and on Intel ~18ms. Is it ok?
Can I do something with it to run faster? Or it's bug?
thank you!
