tvm
Loading...
Searching...
No Matches
gpu.h
Go to the documentation of this file.
1/*
2 * Licensed to the Apache Software Foundation (ASF) under one
3 * or more contributor license agreements. See the NOTICE file
4 * distributed with this work for additional information
5 * regarding copyright ownership. The ASF licenses this file
6 * to you under the Apache License, Version 2.0 (the
7 * "License"); you may not use this file except in compliance
8 * with the License. You may obtain a copy of the License at
9 *
10 * http://www.apache.org/licenses/LICENSE-2.0
11 *
12 * Unless required by applicable law or agreed to in writing,
13 * software distributed under the License is distributed on an
14 * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
15 * KIND, either express or implied. See the License for the
16 * specific language governing permissions and limitations
17 * under the License.
18 */
19
24#ifndef TVM_TIRX_OP_GPU_H_
25#define TVM_TIRX_OP_GPU_H_
26
27#include <tvm/ir/op.h>
28#include <tvm/ir/prim/expr.h>
29
30namespace tvm {
31namespace tirx {
32
37
45
55
65
73
81
93
105
117
129
134
154
164
165} // namespace tirx
166} // namespace tvm
167
168#endif // TVM_TIRX_OP_GPU_H_
Non-nullable source-location metadata.
Definition location.h:91
Reference-counted handle to a canonical named operator.
Definition op.h:147
Typed reference/view over any Expr whose ExprNode::ty is PrimType.
Definition base_expr.h:447
Reference to the shared immutable unknown location.
Definition location.h:105
Canonical primitive operators and their registration.
TIR expressions.
TVM_TIRX_DLL const Op & gpu_thread_return_op()
Return from a GPU thread without returning a function value.
TVM_TIRX_DLL const Op & gpu_thread_filter_op()
Mark an opaque predicate that selects one thread on an active-set axis.
TVM_TIRX_DLL const Op & gpu_warp_shuffle_down_op()
Read the value from a higher lane in the warp.
TVM_TIRX_DLL const Op & gpu_warp_shuffle_up_op()
Read the value from a lower lane in the warp.
TVM_TIRX_DLL const Op & gpu_active_thread_selector_op()
Analysis-only active-thread selector.
TVM_TIRX_DLL const Op & gpu_warp_shuffle_op()
Read the value from a selected lane in the warp.
TVM_TIRX_DLL PrimExpr gpu_thread_return(Location loc=UnknownLoc())
Return from a GPU thread without returning a function value.
TVM_TIRX_DLL const Op & gpu_thread_allreduce_op()
Cross-thread reduction with an explicit typed combiner and identities.
TVM_TIRX_DLL const Op & gpu_storage_sync_op()
Synchronize accesses in a storage scope.
TVM_TIRX_DLL const Op & gpu_warp_activemask_op()
Return the bit mask of active lanes in the warp.
TVM_TIRX_DLL const Op & gpu_thread_invariant_op()
Mark a condition to be thread invariant.
TVM_TIRX_DLL const Op & gpu_warp_shuffle_xor_op()
Read the value from the lane selected by an XOR mask.
TVM_TIRX_DLL const Op & gpu_dp4a_op()
Dot product of two int8x4 vectors and add an optional accumulator.
An object that builds and maintains block scope and StmtSref mapping for Dependence analysis.
Definition attr.h:23
#define TVM_TIRX_DLL
Definition base.h:83