[AMDGPU] Fix AddedComplexity for s_buffer_load patterns. NFCI.

[foad]

Summary

We set AddedComplexity = 100 for s_load patterns to prefer them over
global loads, but for s_buffer_load patterns there is no need to do
this and it was quietly overriding the AddedComplexity of each
individual GCNPat that is defined inside SMLoad_Pattern (but in practice
that did not appear to make any difference).

[arsenm]

llvm/lib/Target/AMDGPU/SMInstructions.td

904

Can you just move this up instead of indenting all the patterns?

[foad]

Update

llvm/lib/Target/AMDGPU/SMInstructions.td

 foreach vt = SReg_256.RegTypes in {
 defm : SMRD_Pattern <"S_LOAD_DWORDX8", vt>;
 }
 
 foreach vt = SReg_512.RegTypes in {
 defm : SMRD_Pattern <"S_LOAD_DWORDX16", vt>;
 }
 
+} // End let AddedComplexity = 100
+
 defm : SMLoad_Pattern <"S_BUFFER_LOAD_DWORD",     i32>;
 defm : SMLoad_Pattern <"S_BUFFER_LOAD_DWORDX2",   v2i32>;
 defm : SMLoad_Pattern <"S_BUFFER_LOAD_DWORDX4",   v4i32>;
 defm : SMLoad_Pattern <"S_BUFFER_LOAD_DWORDX8",   v8i32>;
 defm : SMLoad_Pattern <"S_BUFFER_LOAD_DWORDX16",  v16i32>;
 
 defm : SMLoad_Pattern <"S_BUFFER_LOAD_DWORD",     f32>;
 defm : SMLoad_Pattern <"S_BUFFER_LOAD_DWORDX2",   v2f32>;
 defm : SMLoad_Pattern <"S_BUFFER_LOAD_DWORDX4",   v4f32>;
 defm : SMLoad_Pattern <"S_BUFFER_LOAD_DWORDX8",   v8f32>;
 defm : SMLoad_Pattern <"S_BUFFER_LOAD_DWORDX16",  v16f32>;
-} // End let AddedComplexity = 100
 
 let OtherPredicates = [HasSMemTimeInst] in {
 def : GCNPat <
   (i64 (readcyclecounter)),
   (S_MEMTIME)
 >;
 } // let OtherPredicates = [HasSMemTimeInst]