[NVPTX] Add prefetch tensormap variant

llvm/docs/NVPTXUsage.rst

@@ -971,6 +971,10 @@ Syntax:
   declare void  @llvm.nvvm.prefetch.L1(ptr %ptr)
   declare void  @llvm.nvvm.prefetch.L2(ptr %ptr)
 
+  declare void  @llvm.nvvm.prefetch.tensormap.p0(ptr %ptr)
+  declare void  @llvm.nvvm.prefetch.tensormap.p4(ptr addrspace(4) %const_ptr)
+  declare void  @llvm.nvvm.prefetch.tensormap.p101(ptr addrspace(101) %param_ptr)  
+
   declare void  @llvm.nvvm.prefetch.global.L2.evict.normal(ptr addrspace(1) %global_ptr)
   declare void  @llvm.nvvm.prefetch.global.L2.evict.last(ptr addrspace(1) %global_ptr)
 
@@ -983,7 +987,10 @@ The '``@llvm.nvvm.prefetch.*``' and '``@llvm.nvvm.prefetchu.*``' intrinsic
 correspond to the '``prefetch.*``;' and '``prefetchu.*``' family of PTX instructions. 
 The '``prefetch.*``' instructions bring the cache line containing the
 specified address in global or local memory address space into the 
-specified cache level (L1 or L2). The '`prefetchu.*``' instruction brings the cache line 
+specified cache level (L1 or L2). If the '``.tensormap``' qualifier is specified then the 
+prefetch instruction brings the cache line containing the specified address in the 
+'``.const``' or '``.param memory``' state space for subsequent use by the '``cp.async.bulk.tensor``' 
+instruction. The '`prefetchu.*``' instruction brings the cache line 
 containing the specified generic address into the specified uniform cache level.
 If no address space is specified, it is assumed to be generic address. The intrinsic 
 uses and eviction priority which can be accessed by the '``.level::eviction_priority``' modifier.

llvm/include/llvm/IR/IntrinsicsNVVM.td

@@ -137,6 +137,7 @@
 
 def llvm_global_ptr_ty  : LLVMQualPointerType<1>;         // (global)ptr
 def llvm_shared_ptr_ty  : LLVMQualPointerType<3>;         // (shared)ptr
+def llvm_constant_ptr_ty: LLVMQualPointerType<4>;         // (const)ptr
 def llvm_local_ptr_ty   : LLVMQualPointerType<5>;         // (local)ptr
 def llvm_tmem_ptr_ty    : LLVMQualPointerType<6>;         // (tensor memory)ptr
 def llvm_shared_cluster_ptr_ty : LLVMQualPointerType<7>;  // (shared_cluster)ptr
@@ -2087,15 +2088,17 @@ foreach dim = 1...5 in {
 // Intrinsics for Prefetch and Prefetchu
 let IntrProperties = [IntrArgMemOnly, ReadOnly<ArgIndex<0>>, NoCapture<ArgIndex<0>>] in {
   foreach level = ["L1", "L2"] in {
-    def int_nvvm_prefetch_ # level : Intrinsic<[], [llvm_ptr_ty]>;
-    def int_nvvm_prefetch_global_ # level : Intrinsic<[], [llvm_global_ptr_ty]>;
-    def int_nvvm_prefetch_local_ # level : Intrinsic<[], [llvm_local_ptr_ty]>;
+    def int_nvvm_prefetch_ # level : DefaultAttrsIntrinsic<[], [llvm_ptr_ty]>;
+    def int_nvvm_prefetch_global_ # level : DefaultAttrsIntrinsic<[], [llvm_global_ptr_ty]>;
+    def int_nvvm_prefetch_local_ # level : DefaultAttrsIntrinsic<[], [llvm_local_ptr_ty]>;
   }
 
+  def int_nvvm_prefetch_tensormap : DefaultAttrsIntrinsic<[], [llvm_anyptr_ty]>;
+
   foreach eviction_priority = ["evict_normal", "evict_last"] in
-    def int_nvvm_prefetch_global_L2_ # eviction_priority : Intrinsic<[], [llvm_global_ptr_ty]>;
+    def int_nvvm_prefetch_global_L2_ # eviction_priority : DefaultAttrsIntrinsic<[], [llvm_global_ptr_ty]>;
 
-  def int_nvvm_prefetchu_L1 : Intrinsic<[], [llvm_ptr_ty]>;
+  def int_nvvm_prefetchu_L1 : DefaultAttrsIntrinsic<[], [llvm_ptr_ty]>;
 }
 
 // applypriority

llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp

@@ -3973,6 +3973,18 @@ bool NVPTXTargetLowering::getTgtMemIntrinsic(
     return true;
   }
 
+  case Intrinsic::nvvm_prefetch_tensormap: {
+    auto &DL = I.getDataLayout();
+    Info.opc = ISD::INTRINSIC_VOID;
+    Info.memVT = getPointerTy(DL);
+    Info.ptrVal = I.getArgOperand(0);
+    Info.offset = 0;
+    Info.flags =
+        MachineMemOperand::MOLoad | MachineMemOperand::MODereferenceable;
+    Info.align.reset();
+    return true;
+  }
+
   case Intrinsic::nvvm_ldu_global_i:
   case Intrinsic::nvvm_ldu_global_f:
   case Intrinsic::nvvm_ldu_global_p: {

llvm/lib/Target/NVPTX/NVPTXIntrinsics.td

@@ -39,6 +39,12 @@ def AS_match {
   code global = [{
    return ChkMemSDNodeAddressSpace(N, llvm::ADDRESS_SPACE_GLOBAL);
   }];
+  code const = [{
+   return ChkMemSDNodeAddressSpace(N, llvm::ADDRESS_SPACE_CONST);
+  }];
+  code param = [{
+   return ChkMemSDNodeAddressSpace(N, llvm::ADDRESS_SPACE_PARAM);
+  }];
 }
 
 // A node that will be replaced with the current PTX version.
@@ -744,35 +750,50 @@ foreach dim = [1, 2, 3, 4, 5] in {
   }
 }
 
-//Prefetch and Prefetchu 
+//Prefetchu and Prefetch
+
+class PREFETCH_CONST_CHK<dag frag>
+          : PatFrag<!setdagop(frag, ops), frag, AS_match.const>;
+class PREFETCH_GENERIC_CHK<dag frag>
+          : PatFrag<!setdagop(frag, ops), frag, AS_match.generic>;
+class PREFETCH_PARAM_CHK<dag frag>
+          : PatFrag<!setdagop(frag, ops), frag, AS_match.param>;
+
+defvar frag_pat = (int_nvvm_prefetch_tensormap node:$addr);
+
+def prefetch_tensormap_const : PREFETCH_CONST_CHK<frag_pat>;
+def prefetch_tensormap_gen   : PREFETCH_GENERIC_CHK<frag_pat>;
+def prefetch_tensormap_param : PREFETCH_PARAM_CHK<frag_pat>;
 
-class PREFETCH_INTRS<string InstName> :
+def PREFETCH_CONST_TENSORMAP : NVPTXInst<(outs), (ins ADDR:$addr),
+          "prefetch.const.tensormap [$addr];",
+          [(prefetch_tensormap_const addr:$addr)]>;
+
+def PREFETCH_GENERIC_TENSORMAP : NVPTXInst<(outs), (ins ADDR:$addr),
+          "prefetch.tensormap [$addr];",
+          [(prefetch_tensormap_gen addr:$addr)]>; 
+
+def PREFETCH_PARAM_TENSORMAP : NVPTXInst<(outs), (ins ADDR:$addr),
+          "prefetch.param.tensormap [$addr];",
+          [(prefetch_tensormap_param addr:$addr)]>; 
+
+
+class PREFETCH_INTRS<string InstName, Intrinsic Intr> :
           BasicNVPTXInst<(outs), (ins ADDR:$addr),
           InstName,
-          [(!cast<Intrinsic>(!strconcat("int_nvvm_",
-          !subst(".", "_", InstName))) addr:$addr)]>,
-          Requires<[hasPTX<80>, hasSM<90>]>;
-
-
-def PREFETCH_L1 : PREFETCH_INTRS<"prefetch.L1">;
-def PREFETCH_L2 : PREFETCH_INTRS<"prefetch.L2">;
-def PREFETCH_GLOBAL_L1 : PREFETCH_INTRS<"prefetch.global.L1">;
-def PREFETCH_LOCAL_L1  : PREFETCH_INTRS<"prefetch.local.L1">;
-def PREFETCH_GLOBAL_L2 : PREFETCH_INTRS<"prefetch.global.L2">;
-def PREFETCH_LOCAL_L2 : PREFETCH_INTRS<"prefetch.local.L2">;
-
-def PREFETCH_GLOBAL_L2_EVICT_NORMAL : BasicNVPTXInst<(outs), (ins ADDR:$addr),
-                                      "prefetch.global.L2::evict_normal",
-                                      [(int_nvvm_prefetch_global_L2_evict_normal addr:$addr)]>,
-                                      Requires<[hasPTX<80>, hasSM<90>]>;
-
-def PREFETCH_GLOBAL_L2_EVICT_LAST   : BasicNVPTXInst<(outs), (ins ADDR:$addr),
-                                      "prefetch.global.L2::evict_last",
-                                      [(int_nvvm_prefetch_global_L2_evict_last addr:$addr)]>,
-                                      Requires<[hasPTX<80>, hasSM<90>]>;
-
-
-def PREFETCHU_L1 : PREFETCH_INTRS<"prefetchu.L1">;
+          [(Intr addr:$addr)]>;
+
+def PREFETCHU_L1 : PREFETCH_INTRS<"prefetchu.L1", int_nvvm_prefetchu_L1>;   
+def PREFETCH_L1 : PREFETCH_INTRS<"prefetch.L1", int_nvvm_prefetch_L1>;
+def PREFETCH_L2 : PREFETCH_INTRS<"prefetch.L2", int_nvvm_prefetch_L2>;
+def PREFETCH_GLOBAL_L1 : PREFETCH_INTRS<"prefetch.global.L1", int_nvvm_prefetch_global_L1>;
+def PREFETCH_LOCAL_L1 : PREFETCH_INTRS<"prefetch.local.L1", int_nvvm_prefetch_local_L1>;
+def PREFETCH_GLOBAL_L2 : PREFETCH_INTRS<"prefetch.global.L2", int_nvvm_prefetch_global_L2>;
+def PREFETCH_LOCAL_L2 : PREFETCH_INTRS<"prefetch.local.L2", int_nvvm_prefetch_local_L2>;
+def PREFETCH_GLOBAL_L2_EVICT_NORMAL : PREFETCH_INTRS<"prefetch.global.L2::evict_normal", 
+                                      int_nvvm_prefetch_global_L2_evict_normal>;
+def PREFETCH_GLOBAL_L2_EVICT_LAST : PREFETCH_INTRS<"prefetch.global.L2::evict_last", 
+                                    int_nvvm_prefetch_global_L2_evict_last>;
 
 //Applypriority intrinsics
 class APPLYPRIORITY_L2_INTRS<string addrspace> :

llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp

@@ -564,7 +564,8 @@ bool NVPTXTTIImpl::collectFlatAddressOperands(SmallVectorImpl<int> &OpIndexes,
   case Intrinsic::nvvm_isspacep_global:
   case Intrinsic::nvvm_isspacep_local:
   case Intrinsic::nvvm_isspacep_shared:
-  case Intrinsic::nvvm_isspacep_shared_cluster: {
+  case Intrinsic::nvvm_isspacep_shared_cluster:
+  case Intrinsic::nvvm_prefetch_tensormap: {
     OpIndexes.push_back(0);
     return true;
   }
@@ -587,6 +588,9 @@ Value *NVPTXTTIImpl::rewriteIntrinsicWithAddressSpace(IntrinsicInst *II,
       return ConstantInt::get(II->getType(), *R);
     return nullptr;
   }
+  case Intrinsic::nvvm_prefetch_tensormap: {
+    return nullptr;
+  }
   }
   return nullptr;
 }

llvm/test/CodeGen/NVPTX/prefetch.ll

@@ -12,6 +12,10 @@ declare void  @llvm.nvvm.prefetch.local.L2(ptr addrspace(5) %local_ptr)
 declare void  @llvm.nvvm.prefetch.L1(ptr %ptr)
 declare void  @llvm.nvvm.prefetch.L2(ptr %ptr)
 
+declare void  @llvm.nvvm.prefetch.tensormap.p0(ptr %ptr)
+declare void  @llvm.nvvm.prefetch.tensormap.p4(ptr addrspace(4) %const_ptr)
+declare void  @llvm.nvvm.prefetch.tensormap.p101(ptr addrspace(101) %param_ptr)
+
 declare void  @llvm.nvvm.prefetch.global.L2.evict.normal(ptr addrspace(1) %global_ptr)
 declare void  @llvm.nvvm.prefetch.global.L2.evict.last(ptr addrspace(1) %global_ptr)
 
@@ -78,4 +82,43 @@ define void @prefetchu_l1(ptr %ptr) {
 ; CHECK-PTX64-NEXT:    ret;
   tail call void @llvm.nvvm.prefetchu.L1(ptr %ptr)
   ret void
+}
+
+define void @prefetch_tensormap(ptr %ptr) {
+; CHECK-PTX64-LABEL: prefetch_tensormap(
+; CHECK-PTX64:       {
+; CHECK-PTX64-NEXT:    .reg .b64 %rd<2>;
+; CHECK-PTX64-EMPTY:
+; CHECK-PTX64-NEXT:  // %bb.0:
+; CHECK-PTX64-NEXT:    ld.param.b64 %rd1, [prefetch_tensormap_param_0];
+; CHECK-PTX64-NEXT:    prefetch.tensormap [%rd1];
+; CHECK-PTX64-NEXT:    ret;
+  tail call void @llvm.nvvm.prefetch.tensormap.p0(ptr %ptr)
+  ret void
+}
+
+define void @prefetch_const_tensormap(ptr addrspace(4) %const_ptr) {
+; CHECK-PTX64-LABEL: prefetch_const_tensormap(
+; CHECK-PTX64:       {
+; CHECK-PTX64-NEXT:    .reg .b64 %rd<2>;
+; CHECK-PTX64-EMPTY:
+; CHECK-PTX64-NEXT:  // %bb.0:
+; CHECK-PTX64-NEXT:    ld.param.b64 %rd1, [prefetch_const_tensormap_param_0];
+; CHECK-PTX64-NEXT:    prefetch.const.tensormap [%rd1];
+; CHECK-PTX64-NEXT:    ret;
+  tail call void @llvm.nvvm.prefetch.tensormap.p4(ptr addrspace(4) %const_ptr)
+  ret void
+}
+
+define void @prefetch_param_tensormap(ptr addrspace(101) %param_ptr) {
+; CHECK-PTX64-LABEL: prefetch_param_tensormap(
+; CHECK-PTX64:       {
+; CHECK-PTX64-NEXT:    .reg .b64 %rd<2>;
+; CHECK-PTX64-EMPTY:
+; CHECK-PTX64-NEXT:  // %bb.0:
+; CHECK-PTX64-NEXT:    ld.param.b64 %rd1, [prefetch_param_tensormap_param_0];
+; CHECK-PTX64-NEXT:    prefetch.param.tensormap [%rd1];
+; CHECK-PTX64-NEXT:    ret;
+  tail call void @llvm.nvvm.prefetch.tensormap.p101(ptr addrspace(101) %param_ptr)
+  ret void
 }