On Sunday, January 15, 2017 12:19:36 PM PST Jordan Justen wrote:
> Maybe you should drop 'TCS' from the subject? I think TCS is the
> reason for the change, but it doesn't only affect TCS barrier
> messages, right?
Well, scalar TCS and CS are already in align1 mode, so even though
this is common c
Maybe you should drop 'TCS' from the subject? I think TCS is the
reason for the change, but it doesn't only affect TCS barrier
messages, right?
On 2017-01-15 01:45:45, Kenneth Graunke wrote:
> In commit 7428e6f86ab5 we switched the barrier SEND message's
> destination type to UW to avoid problems
In commit 7428e6f86ab5 we switched the barrier SEND message's
destination type to UW to avoid problems in SIMD16 compute shaders.
Tessellation control shaders also use barriers, and in vec4 mode, we
were emitting them in align16 mode. The simulator warns that only UD,
D, F, and DF are valid desti
