aboutsummaryrefslogtreecommitdiffstats
path: root/lib/Target/R600/AMDGPUInstructions.td
diff options
context:
space:
mode:
Diffstat (limited to 'lib/Target/R600/AMDGPUInstructions.td')
-rw-r--r--lib/Target/R600/AMDGPUInstructions.td125
1 files changed, 120 insertions, 5 deletions
diff --git a/lib/Target/R600/AMDGPUInstructions.td b/lib/Target/R600/AMDGPUInstructions.td
index b86b781..c215865 100644
--- a/lib/Target/R600/AMDGPUInstructions.td
+++ b/lib/Target/R600/AMDGPUInstructions.td
@@ -23,6 +23,8 @@ class AMDGPUInst <dag outs, dag ins, string asm, list<dag> pattern> : Instructio
let Pattern = pattern;
let Itinerary = NullALU;
+ let isCodeGenOnly = 1;
+
let TSFlags{63} = isRegisterLoad;
let TSFlags{62} = isRegisterStore;
}
@@ -34,9 +36,15 @@ class AMDGPUShaderInst <dag outs, dag ins, string asm, list<dag> pattern>
}
+def FP32Denormals : Predicate<"Subtarget.hasFP32Denormals()">;
+def FP64Denormals : Predicate<"Subtarget.hasFP64Denormals()">;
+def UnsafeFPMath : Predicate<"TM.Options.UnsafeFPMath">;
+
def InstFlag : OperandWithDefaultOps <i32, (ops (i32 0))>;
def ADDRIndirect : ComplexPattern<iPTR, 2, "SelectADDRIndirect", [], []>;
+let OperandType = "OPERAND_IMMEDIATE" in {
+
def u32imm : Operand<i32> {
let PrintMethod = "printU32ImmOperand";
}
@@ -49,6 +57,8 @@ def u8imm : Operand<i8> {
let PrintMethod = "printU8ImmOperand";
}
+} // End OperandType = "OPERAND_IMMEDIATE"
+
//===--------------------------------------------------------------------===//
// Custom Operands
//===--------------------------------------------------------------------===//
@@ -125,13 +135,35 @@ def COND_NE : PatLeaf <
def COND_NULL : PatLeaf <
(cond),
- [{return false;}]
+ [{(void)N; return false;}]
>;
//===----------------------------------------------------------------------===//
// Load/Store Pattern Fragments
//===----------------------------------------------------------------------===//
+class PrivateMemOp <dag ops, dag frag> : PatFrag <ops, frag, [{
+ return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::PRIVATE_ADDRESS;
+}]>;
+
+class PrivateLoad <SDPatternOperator op> : PrivateMemOp <
+ (ops node:$ptr), (op node:$ptr)
+>;
+
+class PrivateStore <SDPatternOperator op> : PrivateMemOp <
+ (ops node:$value, node:$ptr), (op node:$value, node:$ptr)
+>;
+
+def extloadi8_private : PrivateLoad <extloadi8>;
+def sextloadi8_private : PrivateLoad <sextloadi8>;
+def extloadi16_private : PrivateLoad <extloadi16>;
+def sextloadi16_private : PrivateLoad <sextloadi16>;
+def load_private : PrivateLoad <load>;
+
+def truncstorei8_private : PrivateStore <truncstorei8>;
+def truncstorei16_private : PrivateStore <truncstorei16>;
+def store_private : PrivateStore <store>;
+
def global_store : PatFrag<(ops node:$val, node:$ptr),
(store node:$val, node:$ptr), [{
return isGlobalStore(dyn_cast<StoreSDNode>(N));
@@ -165,6 +197,14 @@ def sextloadi8_global : PatFrag<(ops node:$ptr), (sextloadi8 node:$ptr), [{
return isGlobalLoad(dyn_cast<LoadSDNode>(N));
}]>;
+def az_extloadi8_flat : PatFrag<(ops node:$ptr), (az_extloadi8 node:$ptr), [{
+ return isFlatLoad(dyn_cast<LoadSDNode>(N));
+}]>;
+
+def sextloadi8_flat : PatFrag<(ops node:$ptr), (sextloadi8 node:$ptr), [{
+ return isFlatLoad(dyn_cast<LoadSDNode>(N));
+}]>;
+
def az_extloadi8_constant : PatFrag<(ops node:$ptr), (az_extloadi8 node:$ptr), [{
return isConstantLoad(dyn_cast<LoadSDNode>(N), -1);
}]>;
@@ -193,6 +233,14 @@ def sextloadi16_global : PatFrag<(ops node:$ptr), (sextloadi16 node:$ptr), [{
return isGlobalLoad(dyn_cast<LoadSDNode>(N));
}]>;
+def az_extloadi16_flat : PatFrag<(ops node:$ptr), (az_extloadi16 node:$ptr), [{
+ return isFlatLoad(dyn_cast<LoadSDNode>(N));
+}]>;
+
+def sextloadi16_flat : PatFrag<(ops node:$ptr), (sextloadi16 node:$ptr), [{
+ return isFlatLoad(dyn_cast<LoadSDNode>(N));
+}]>;
+
def az_extloadi16_constant : PatFrag<(ops node:$ptr), (az_extloadi16 node:$ptr), [{
return isConstantLoad(dyn_cast<LoadSDNode>(N), -1);
}]>;
@@ -218,6 +266,11 @@ def az_extloadi32_global : PatFrag<(ops node:$ptr),
return isGlobalLoad(dyn_cast<LoadSDNode>(N));
}]>;
+def az_extloadi32_flat : PatFrag<(ops node:$ptr),
+ (az_extloadi32 node:$ptr), [{
+ return isFlatLoad(dyn_cast<LoadSDNode>(N));
+}]>;
+
def az_extloadi32_constant : PatFrag<(ops node:$ptr),
(az_extloadi32 node:$ptr), [{
return isConstantLoad(dyn_cast<LoadSDNode>(N), -1);
@@ -233,6 +286,16 @@ def truncstorei16_global : PatFrag<(ops node:$val, node:$ptr),
return isGlobalStore(dyn_cast<StoreSDNode>(N));
}]>;
+def truncstorei8_flat : PatFrag<(ops node:$val, node:$ptr),
+ (truncstorei8 node:$val, node:$ptr), [{
+ return isFlatStore(dyn_cast<StoreSDNode>(N));
+}]>;
+
+def truncstorei16_flat : PatFrag<(ops node:$val, node:$ptr),
+ (truncstorei16 node:$val, node:$ptr), [{
+ return isFlatStore(dyn_cast<StoreSDNode>(N));
+}]>;
+
def local_store : PatFrag<(ops node:$val, node:$ptr),
(store node:$val, node:$ptr), [{
return isLocalStore(dyn_cast<StoreSDNode>(N));
@@ -252,6 +315,17 @@ def local_load : PatFrag<(ops node:$ptr), (load node:$ptr), [{
return isLocalLoad(dyn_cast<LoadSDNode>(N));
}]>;
+class Aligned8Bytes <dag ops, dag frag> : PatFrag <ops, frag, [{
+ return cast<MemSDNode>(N)->getAlignment() % 8 == 0;
+}]>;
+
+def local_load_aligned8bytes : Aligned8Bytes <
+ (ops node:$ptr), (local_load node:$ptr)
+>;
+
+def local_store_aligned8bytes : Aligned8Bytes <
+ (ops node:$val, node:$ptr), (local_store node:$val, node:$ptr)
+>;
class local_binary_atomic_op<SDNode atomic_op> :
PatFrag<(ops node:$ptr, node:$value),
@@ -277,6 +351,7 @@ def mskor_global : PatFrag<(ops node:$val, node:$ptr),
return dyn_cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS;
}]>;
+
def atomic_cmp_swap_32_local :
PatFrag<(ops node:$ptr, node:$cmp, node:$swap),
(atomic_cmp_swap node:$ptr, node:$cmp, node:$swap), [{
@@ -293,6 +368,45 @@ def atomic_cmp_swap_64_local :
AN->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS;
}]>;
+def flat_load : PatFrag<(ops node:$ptr), (load node:$ptr), [{
+ return isFlatLoad(dyn_cast<LoadSDNode>(N));
+}]>;
+
+def flat_store : PatFrag<(ops node:$val, node:$ptr),
+ (store node:$val, node:$ptr), [{
+ return isFlatStore(dyn_cast<StoreSDNode>(N));
+}]>;
+
+def mskor_flat : PatFrag<(ops node:$val, node:$ptr),
+ (AMDGPUstore_mskor node:$val, node:$ptr), [{
+ return dyn_cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::FLAT_ADDRESS;
+}]>;
+
+class global_binary_atomic_op<SDNode atomic_op> : PatFrag<
+ (ops node:$ptr, node:$value),
+ (atomic_op node:$ptr, node:$value),
+ [{return cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS;}]
+>;
+
+def atomic_swap_global : global_binary_atomic_op<atomic_swap>;
+def atomic_add_global : global_binary_atomic_op<atomic_load_add>;
+def atomic_and_global : global_binary_atomic_op<atomic_load_and>;
+def atomic_max_global : global_binary_atomic_op<atomic_load_max>;
+def atomic_min_global : global_binary_atomic_op<atomic_load_min>;
+def atomic_or_global : global_binary_atomic_op<atomic_load_or>;
+def atomic_sub_global : global_binary_atomic_op<atomic_load_sub>;
+def atomic_umax_global : global_binary_atomic_op<atomic_load_umax>;
+def atomic_umin_global : global_binary_atomic_op<atomic_load_umin>;
+def atomic_xor_global : global_binary_atomic_op<atomic_load_xor>;
+
+//===----------------------------------------------------------------------===//
+// Misc Pattern Fragments
+//===----------------------------------------------------------------------===//
+
+def fmad : PatFrag <
+ (ops node:$src0, node:$src1, node:$src2),
+ (fadd (fmul node:$src0, node:$src1), node:$src2)
+>;
class Constants {
int TWO_PI = 0x40c90fdb;
@@ -412,8 +526,9 @@ class DwordAddrPat<ValueType vt, RegisterClass rc> : Pat <
// BFI_INT patterns
-multiclass BFIPatterns <Instruction BFI_INT, Instruction LoadImm32> {
-
+multiclass BFIPatterns <Instruction BFI_INT,
+ Instruction LoadImm32,
+ RegisterClass RC64> {
// Definition from ISA doc:
// (y & x) | (z & ~x)
def : Pat <
@@ -435,8 +550,8 @@ multiclass BFIPatterns <Instruction BFI_INT, Instruction LoadImm32> {
def : Pat <
(f64 (fcopysign f64:$src0, f64:$src1)),
- (INSERT_SUBREG (INSERT_SUBREG (f64 (IMPLICIT_DEF)),
- (i32 (EXTRACT_SUBREG $src0, sub0)), sub0),
+ (REG_SEQUENCE RC64,
+ (i32 (EXTRACT_SUBREG $src0, sub0)), sub0,
(BFI_INT (LoadImm32 0x7fffffff),
(i32 (EXTRACT_SUBREG $src0, sub1)),
(i32 (EXTRACT_SUBREG $src1, sub1))), sub1)
generated by cgit v1.1 at 2024-06-03 16:52:09 +0000