On Thu, 2017-01-12 at 15:14 -0800, Matt Turner wrote:
> On Thu, Jan 5, 2017 at 5:07 AM, Samuel Iglesias Gonsálvez
> <sigles...@igalia.com> wrote:
> > From: "Juan A. Suarez Romero" <jasua...@igalia.com>
> >
> > The execution data size is the biggest type size of any instruction
> > operand.
> >
> > We will use it to know if the instruction deals with DF, because in
> > Ivy
> > we need to duplicate the execution size and regioning parameters.
> > ---
> > src/mesa/drivers/dri/i965/brw_fs.cpp | 19 ++++++++++++++-----
> > src/mesa/drivers/dri/i965/brw_ir_fs.h | 1 +
> > 2 files changed, 15 insertions(+), 5 deletions(-)
> >
> > diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp
> > b/src/mesa/drivers/dri/i965/brw_fs.cpp
> > index c8a0693..eb3b4aa 100644
> > --- a/src/mesa/drivers/dri/i965/brw_fs.cpp
> > +++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
> > @@ -340,6 +340,19 @@ fs_inst::has_source_and_destination_hazard()
> > const
> > }
> > }
> >
> > +unsigned
> > +fs_inst::exec_data_size() const
> > +{
> > + unsigned exec_data_size = 0;
> > +
> > + for (int i = 0; i < this->sources; i++) {
> > + if (this->src[i].type != BAD_FILE)
> > + exec_data_size = MAX2(exec_data_size, type_sz(this-
> > >src[i].type));
>
> Indentation is wrong (looks like 2 space, but should be 3 space).
> Also
> please use braces in nested control flow.
>
OK
> > + }
> > +
> > + return exec_data_size;
> > +}
> > +
> > bool
> > fs_inst::is_copy_payload(const brw::simple_allocator &grf_alloc)
> > const
> > {
> > @@ -4577,11 +4590,7 @@ get_fpu_lowered_simd_width(const struct
> > gen_device_info *devinfo,
> > !inst->force_writemask_all) {
> > const unsigned channels_per_grf = inst->exec_size /
> > DIV_ROUND_UP(inst->size_written, REG_SIZE);
> > - unsigned exec_type_size = 0;
> > - for (int i = 0; i < inst->sources; i++) {
> > - if (inst->src[i].file != BAD_FILE)
> > - exec_type_size = MAX2(exec_type_size, type_sz(inst-
> > >src[i].type));
> > - }
> > + unsigned exec_type_size = inst->exec_data_size();
> > assert(exec_type_size);
> >
> > /* The hardware shifts exactly 8 channels per compressed
> > half of the
> > diff --git a/src/mesa/drivers/dri/i965/brw_ir_fs.h
> > b/src/mesa/drivers/dri/i965/brw_ir_fs.h
> > index cad3712..9875f2d 100644
> > --- a/src/mesa/drivers/dri/i965/brw_ir_fs.h
> > +++ b/src/mesa/drivers/dri/i965/brw_ir_fs.h
> > @@ -349,6 +349,7 @@ public:
> > bool can_change_types() const;
> > bool has_side_effects() const;
> > bool has_source_and_destination_hazard() const;
> > + unsigned exec_data_size() const;
>
> In the reworked patch in your new branch, this still exists even
> though the function is no longer a member.
>
Right, I forgot to delete it.
Thanks!
Sam
signature.asc
Description: This is a digitally signed message part
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
