; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt -cost-model -analyze -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx900 -mattr=+half-rate-64-ops < %s | FileCheck -check-prefixes=ALL,FASTF64 %s
-; RUN: opt -cost-model -analyze -mtriple=amdgcn-unknown-amdhsa -mattr=-half-rate-64-ops < %s | FileCheck -check-prefixes=ALL,SLOWF64 %s
-; RUN: opt -cost-model -analyze -cost-kind=code-size -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx900 -mattr=+half-rate-64-ops < %s | FileCheck -check-prefixes=ALL-SIZE,FASTF64-SIZE %s
-; RUN: opt -cost-model -analyze -cost-kind=code-size -mtriple=amdgcn-unknown-amdhsa -mattr=-half-rate-64-ops < %s | FileCheck -check-prefixes=ALL-SIZE,SLOW64-SIZE %s
+; RUN: opt -cost-model -analyze -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx90a -mattr=+half-rate-64-ops < %s | FileCheck -check-prefixes=FASTF16,GFX90A-FASTF64 %s
+; RUN: opt -cost-model -analyze -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx900 -mattr=+half-rate-64-ops < %s | FileCheck -check-prefixes=NOPACKEDF32,FASTF16,FASTF64 %s
+; RUN: opt -cost-model -analyze -mtriple=amdgcn-unknown-amdhsa -mattr=-half-rate-64-ops < %s | FileCheck -check-prefixes=NOPACKEDF32,SLOWF64 %s
+; RUN: opt -cost-model -cost-kind=code-size -analyze -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx90a -mattr=+half-rate-64-ops < %s | FileCheck -check-prefixes=FASTF16-SIZE,GFX90A-FASTF64-SIZE %s
+; RUN: opt -cost-model -cost-kind=code-size -analyze -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx900 -mattr=+half-rate-64-ops < %s | FileCheck -check-prefixes=NOPACKEDF32-SIZE,FASTF16-SIZE %s
+; RUN: opt -cost-model -cost-kind=code-size -analyze -mtriple=amdgcn-unknown-amdhsa -mattr=-half-rate-64-ops < %s | FileCheck -check-prefixes=NOPACKEDF32-SIZE,SLOWF64-SIZE %s
; END.
define amdgpu_kernel void @fsub_f32() #0 {
-; ALL-LABEL: 'fsub_f32'
-; ALL-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f32 = fsub float undef, undef
-; ALL-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v2f32 = fsub <2 x float> undef, undef
-; ALL-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v3f32 = fsub <3 x float> undef, undef
-; ALL-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %v5f32 = fsub <5 x float> undef, undef
-; ALL-NEXT: Cost Model: Found an estimated cost of 10 for instruction: ret void
+; GFX90A-FASTF64-LABEL: 'fsub_f32'
+; GFX90A-FASTF64-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f32 = fsub float undef, undef
+; GFX90A-FASTF64-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v2f32 = fsub <2 x float> undef, undef
+; GFX90A-FASTF64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v3f32 = fsub <3 x float> undef, undef
+; GFX90A-FASTF64-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v5f32 = fsub <5 x float> undef, undef
+; GFX90A-FASTF64-NEXT: Cost Model: Found an estimated cost of 10 for instruction: ret void
;
-; ALL-SIZE-LABEL: 'fsub_f32'
-; ALL-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f32 = fsub float undef, undef
-; ALL-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v2f32 = fsub <2 x float> undef, undef
-; ALL-SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v3f32 = fsub <3 x float> undef, undef
-; ALL-SIZE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %v5f32 = fsub <5 x float> undef, undef
-; ALL-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
+; NOPACKEDF32-LABEL: 'fsub_f32'
+; NOPACKEDF32-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f32 = fsub float undef, undef
+; NOPACKEDF32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v2f32 = fsub <2 x float> undef, undef
+; NOPACKEDF32-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v3f32 = fsub <3 x float> undef, undef
+; NOPACKEDF32-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %v5f32 = fsub <5 x float> undef, undef
+; NOPACKEDF32-NEXT: Cost Model: Found an estimated cost of 10 for instruction: ret void
+;
+; GFX90A-FASTF64-SIZE-LABEL: 'fsub_f32'
+; GFX90A-FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f32 = fsub float undef, undef
+; GFX90A-FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v2f32 = fsub <2 x float> undef, undef
+; GFX90A-FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v3f32 = fsub <3 x float> undef, undef
+; GFX90A-FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v5f32 = fsub <5 x float> undef, undef
+; GFX90A-FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; NOPACKEDF32-SIZE-LABEL: 'fsub_f32'
+; NOPACKEDF32-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f32 = fsub float undef, undef
+; NOPACKEDF32-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v2f32 = fsub <2 x float> undef, undef
+; NOPACKEDF32-SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v3f32 = fsub <3 x float> undef, undef
+; NOPACKEDF32-SIZE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %v5f32 = fsub <5 x float> undef, undef
+; NOPACKEDF32-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
;
%f32 = fsub float undef, undef
%v2f32 = fsub <2 x float> undef, undef
}
define amdgpu_kernel void @fsub_f64() #0 {
+; GFX90A-FASTF64-LABEL: 'fsub_f64'
+; GFX90A-FASTF64-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f64 = fsub double undef, undef
+; GFX90A-FASTF64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v2f64 = fsub <2 x double> undef, undef
+; GFX90A-FASTF64-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v3f64 = fsub <3 x double> undef, undef
+; GFX90A-FASTF64-NEXT: Cost Model: Found an estimated cost of 10 for instruction: ret void
+;
; FASTF64-LABEL: 'fsub_f64'
; FASTF64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %f64 = fsub double undef, undef
; FASTF64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v2f64 = fsub <2 x double> undef, undef
; SLOWF64-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %v3f64 = fsub <3 x double> undef, undef
; SLOWF64-NEXT: Cost Model: Found an estimated cost of 10 for instruction: ret void
;
-; ALL-SIZE-LABEL: 'fsub_f64'
-; ALL-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %f64 = fsub double undef, undef
-; ALL-SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v2f64 = fsub <2 x double> undef, undef
-; ALL-SIZE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %v3f64 = fsub <3 x double> undef, undef
-; ALL-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
+; GFX90A-FASTF64-SIZE-LABEL: 'fsub_f64'
+; GFX90A-FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f64 = fsub double undef, undef
+; GFX90A-FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v2f64 = fsub <2 x double> undef, undef
+; GFX90A-FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v3f64 = fsub <3 x double> undef, undef
+; GFX90A-FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
+;
+; NOPACKEDF32-SIZE-LABEL: 'fsub_f64'
+; NOPACKEDF32-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %f64 = fsub double undef, undef
+; NOPACKEDF32-SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v2f64 = fsub <2 x double> undef, undef
+; NOPACKEDF32-SIZE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %v3f64 = fsub <3 x double> undef, undef
+; NOPACKEDF32-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
;
%f64 = fsub double undef, undef
%v2f64 = fsub <2 x double> undef, undef
}
define amdgpu_kernel void @fsub_f16() #0 {
-; FASTF64-LABEL: 'fsub_f16'
-; FASTF64-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f16 = fsub half undef, undef
-; FASTF64-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v2f16 = fsub <2 x half> undef, undef
-; FASTF64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v3f16 = fsub <3 x half> undef, undef
-; FASTF64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v4f16 = fsub <4 x half> undef, undef
-; FASTF64-NEXT: Cost Model: Found an estimated cost of 10 for instruction: ret void
+; FASTF16-LABEL: 'fsub_f16'
+; FASTF16-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f16 = fsub half undef, undef
+; FASTF16-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v2f16 = fsub <2 x half> undef, undef
+; FASTF16-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v3f16 = fsub <3 x half> undef, undef
+; FASTF16-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v4f16 = fsub <4 x half> undef, undef
+; FASTF16-NEXT: Cost Model: Found an estimated cost of 10 for instruction: ret void
;
; SLOWF64-LABEL: 'fsub_f16'
; SLOWF64-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f16 = fsub half undef, undef
; SLOWF64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v4f16 = fsub <4 x half> undef, undef
; SLOWF64-NEXT: Cost Model: Found an estimated cost of 10 for instruction: ret void
;
-; FASTF64-SIZE-LABEL: 'fsub_f16'
-; FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f16 = fsub half undef, undef
-; FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v2f16 = fsub <2 x half> undef, undef
-; FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v3f16 = fsub <3 x half> undef, undef
-; FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v4f16 = fsub <4 x half> undef, undef
-; FASTF64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
+; FASTF16-SIZE-LABEL: 'fsub_f16'
+; FASTF16-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f16 = fsub half undef, undef
+; FASTF16-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v2f16 = fsub <2 x half> undef, undef
+; FASTF16-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v3f16 = fsub <3 x half> undef, undef
+; FASTF16-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v4f16 = fsub <4 x half> undef, undef
+; FASTF16-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
;
-; SLOW64-SIZE-LABEL: 'fsub_f16'
-; SLOW64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f16 = fsub half undef, undef
-; SLOW64-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v2f16 = fsub <2 x half> undef, undef
-; SLOW64-SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v3f16 = fsub <3 x half> undef, undef
-; SLOW64-SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v4f16 = fsub <4 x half> undef, undef
-; SLOW64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
+; SLOWF64-SIZE-LABEL: 'fsub_f16'
+; SLOWF64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %f16 = fsub half undef, undef
+; SLOWF64-SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v2f16 = fsub <2 x half> undef, undef
+; SLOWF64-SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v3f16 = fsub <3 x half> undef, undef
+; SLOWF64-SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v4f16 = fsub <4 x half> undef, undef
+; SLOWF64-SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
;
%f16 = fsub half undef, undef
%v2f16 = fsub <2 x half> undef, undef