[llvm] 4c178d8 - [AMDGPU] Pre-commit 2 new saddr load tests. NFC.

[Stanislav Mekhanoshin via llvm-commits]

Author: Stanislav Mekhanoshin
Date: 2021-05-05T09:16:11-07:00
New Revision: 4c178d809b1df3216de251d5345b8ecc9cc3990e

URL: https://github.com/llvm/llvm-project/commit/4c178d809b1df3216de251d5345b8ecc9cc3990e
DIFF: https://github.com/llvm/llvm-project/commit/4c178d809b1df3216de251d5345b8ecc9cc3990e.diff

LOG: [AMDGPU] Pre-commit 2 new saddr load tests. NFC.

Added: 
    

Modified: 
    llvm/test/CodeGen/AMDGPU/global-saddr-load.ll

Removed: 
    


################################################################################
diff  --git a/llvm/test/CodeGen/AMDGPU/global-saddr-load.ll b/llvm/test/CodeGen/AMDGPU/global-saddr-load.ll
index b79b29231b8d..1db97999cb9d 100644
--- a/llvm/test/CodeGen/AMDGPU/global-saddr-load.ll
+++ b/llvm/test/CodeGen/AMDGPU/global-saddr-load.ll
@@ -2383,5 +2383,121 @@ define amdgpu_ps float @global_load_saddr_i8_offset_or_i64_imm_offset_4160(i8 ad
   ret float %to.vgpr
 }
 
+; --------------------------------------------------------------------------------
+; Full 64-bit scalar add.
+; --------------------------------------------------------------------------------
+
+define amdgpu_ps void @global_addr_64bit_lsr_iv(float addrspace(1)* inreg %arg) {
+; GFX9-LABEL: global_addr_64bit_lsr_iv:
+; GFX9:       ; %bb.0: ; %bb
+; GFX9-NEXT:    s_mov_b64 s[0:1], 0
+; GFX9-NEXT:  BB128_1: ; %bb3
+; GFX9-NEXT:    ; =>This Inner Loop Header: Depth=1
+; GFX9-NEXT:    s_add_u32 s4, s2, s0
+; GFX9-NEXT:    s_addc_u32 s5, s3, s1
+; GFX9-NEXT:    v_mov_b32_e32 v0, s4
+; GFX9-NEXT:    v_mov_b32_e32 v1, s5
+; GFX9-NEXT:    global_load_dword v0, v[0:1], off glc
+; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    s_add_u32 s0, s0, 4
+; GFX9-NEXT:    s_addc_u32 s1, s1, 0
+; GFX9-NEXT:    s_cmpk_eq_i32 s0, 0x400
+; GFX9-NEXT:    s_cbranch_scc0 BB128_1
+; GFX9-NEXT:  ; %bb.2: ; %bb2
+; GFX9-NEXT:    s_endpgm
+;
+; GFX10-LABEL: global_addr_64bit_lsr_iv:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_mov_b64 s[0:1], 0
+; GFX10-NEXT:  BB128_1: ; %bb3
+; GFX10-NEXT:    ; =>This Inner Loop Header: Depth=1
+; GFX10-NEXT:    s_add_u32 s4, s2, s0
+; GFX10-NEXT:    s_addc_u32 s5, s3, s1
+; GFX10-NEXT:    v_mov_b32_e32 v0, s4
+; GFX10-NEXT:    v_mov_b32_e32 v1, s5
+; GFX10-NEXT:    s_add_u32 s0, s0, 4
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_cmpk_eq_i32 s0, 0x400
+; GFX10-NEXT:    global_load_dword v0, v[0:1], off glc dlc
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    s_cbranch_scc0 BB128_1
+; GFX10-NEXT:  ; %bb.2: ; %bb2
+; GFX10-NEXT:    s_endpgm
+bb:
+  br label %bb3
+
+bb2:                                              ; preds = %bb3
+  ret void
+
+bb3:                                              ; preds = %bb3, %bb
+  %i = phi i32 [ 0, %bb ], [ %i8, %bb3 ]
+  %i4 = zext i32 %i to i64
+  %i5 = getelementptr inbounds float, float addrspace(1)* %arg, i64 %i4
+  %i6 = load volatile float, float addrspace(1)* %i5, align 4
+  %i8 = add nuw nsw i32 %i, 1
+  %i9 = icmp eq i32 %i8, 256
+  br i1 %i9, label %bb2, label %bb3
+}
+
+define amdgpu_ps void @global_addr_64bit_lsr_iv_multiload(float addrspace(1)* inreg %arg, float addrspace(1)* inreg %arg.1) {
+; GFX9-LABEL: global_addr_64bit_lsr_iv_multiload:
+; GFX9:       ; %bb.0: ; %bb
+; GFX9-NEXT:    s_mov_b64 s[0:1], 0
+; GFX9-NEXT:  BB129_1: ; %bb3
+; GFX9-NEXT:    ; =>This Inner Loop Header: Depth=1
+; GFX9-NEXT:    s_add_u32 s4, s2, s0
+; GFX9-NEXT:    s_addc_u32 s5, s3, s1
+; GFX9-NEXT:    v_mov_b32_e32 v0, s4
+; GFX9-NEXT:    v_mov_b32_e32 v1, s5
+; GFX9-NEXT:    global_load_dword v2, v[0:1], off glc
+; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    global_load_dword v2, v[0:1], off glc
+; GFX9-NEXT:    s_waitcnt vmcnt(0)
+; GFX9-NEXT:    s_add_u32 s0, s0, 4
+; GFX9-NEXT:    s_addc_u32 s1, s1, 0
+; GFX9-NEXT:    s_cmpk_eq_i32 s0, 0x400
+; GFX9-NEXT:    ; kill: killed $vgpr0_vgpr1
+; GFX9-NEXT:    s_cbranch_scc0 BB129_1
+; GFX9-NEXT:  ; %bb.2: ; %bb2
+; GFX9-NEXT:    s_endpgm
+;
+; GFX10-LABEL: global_addr_64bit_lsr_iv_multiload:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_mov_b64 s[0:1], 0
+; GFX10-NEXT:  BB129_1: ; %bb3
+; GFX10-NEXT:    ; =>This Inner Loop Header: Depth=1
+; GFX10-NEXT:    s_add_u32 s4, s2, s0
+; GFX10-NEXT:    s_addc_u32 s5, s3, s1
+; GFX10-NEXT:    v_mov_b32_e32 v0, s4
+; GFX10-NEXT:    v_mov_b32_e32 v1, s5
+; GFX10-NEXT:    s_add_u32 s0, s0, 4
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_cmpk_eq_i32 s0, 0x400
+; GFX10-NEXT:    ; kill: killed $vgpr0_vgpr1
+; GFX10-NEXT:    global_load_dword v2, v[0:1], off glc dlc
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    global_load_dword v2, v[0:1], off glc dlc
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    s_cbranch_scc0 BB129_1
+; GFX10-NEXT:  ; %bb.2: ; %bb2
+; GFX10-NEXT:    s_endpgm
+bb:
+  br label %bb3
+
+bb2:                                              ; preds = %bb3
+  ret void
+
+bb3:                                              ; preds = %bb3, %bb
+  %i = phi i32 [ 0, %bb ], [ %i8, %bb3 ]
+  %i4 = zext i32 %i to i64
+  %i5 = getelementptr inbounds float, float addrspace(1)* %arg, i64 %i4
+  %i6 = load volatile float, float addrspace(1)* %i5, align 4
+  %i5.1 = getelementptr inbounds float, float addrspace(1)* %arg.1, i64 %i4
+  %i6.1 = load volatile float, float addrspace(1)* %i5, align 4
+  %i8 = add nuw nsw i32 %i, 1
+  %i9 = icmp eq i32 %i8, 256
+  br i1 %i9, label %bb2, label %bb3
+}
+
 !0 = !{i32 0, i32 1073741824} ; (1 << 30)
 !1 = !{i32 0, i32 1073741825} ; (1 << 30) + 1