llvm.org GIT mirror llvm / release_34 lib / Target / R600 / AMDGPUInstructions.td
release_34

Tree @release_34 (Download .tar.gz)

AMDGPUInstructions.td @release_34raw · history · blame

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
//===-- AMDGPUInstructions.td - Common instruction defs ---*- tablegen -*-===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file contains instruction defs that are common to all hw codegen
// targets.
//
//===----------------------------------------------------------------------===//

class AMDGPUInst <dag outs, dag ins, string asm, list<dag> pattern> : Instruction {
  field bit isRegisterLoad = 0;
  field bit isRegisterStore = 0;

  let Namespace = "AMDGPU";
  let OutOperandList = outs;
  let InOperandList = ins;
  let AsmString = asm;
  let Pattern = pattern;
  let Itinerary = NullALU;

  let TSFlags{63} = isRegisterLoad;
  let TSFlags{62} = isRegisterStore;
}

class AMDGPUShaderInst <dag outs, dag ins, string asm, list<dag> pattern>
    : AMDGPUInst<outs, ins, asm, pattern> {

  field bits<32> Inst = 0xffffffff;

}

def InstFlag : OperandWithDefaultOps <i32, (ops (i32 0))>;
def ADDRIndirect : ComplexPattern<iPTR, 2, "SelectADDRIndirect", [], []>;

//===----------------------------------------------------------------------===//
// PatLeafs for floating-point comparisons
//===----------------------------------------------------------------------===//

def COND_OEQ : PatLeaf <
  (cond),
  [{return N->get() == ISD::SETOEQ || N->get() == ISD::SETEQ;}]
>;

def COND_OGT : PatLeaf <
  (cond),
  [{return N->get() == ISD::SETOGT || N->get() == ISD::SETGT;}]
>;

def COND_OGE : PatLeaf <
  (cond),
  [{return N->get() == ISD::SETOGE || N->get() == ISD::SETGE;}]
>;

def COND_OLT : PatLeaf <
  (cond),
  [{return N->get() == ISD::SETOLT || N->get() == ISD::SETLT;}]
>;

def COND_OLE : PatLeaf <
  (cond),
  [{return N->get() == ISD::SETOLE || N->get() == ISD::SETLE;}]
>;

def COND_UNE : PatLeaf <
  (cond),
  [{return N->get() == ISD::SETUNE || N->get() == ISD::SETNE;}]
>;

def COND_O : PatLeaf <(cond), [{return N->get() == ISD::SETO;}]>;
def COND_UO : PatLeaf <(cond), [{return N->get() == ISD::SETUO;}]>;

//===----------------------------------------------------------------------===//
// PatLeafs for unsigned comparisons
//===----------------------------------------------------------------------===//

def COND_UGT : PatLeaf <(cond), [{return N->get() == ISD::SETUGT;}]>;
def COND_UGE : PatLeaf <(cond), [{return N->get() == ISD::SETUGE;}]>;
def COND_ULT : PatLeaf <(cond), [{return N->get() == ISD::SETULT;}]>;
def COND_ULE : PatLeaf <(cond), [{return N->get() == ISD::SETULE;}]>;

//===----------------------------------------------------------------------===//
// PatLeafs for signed comparisons
//===----------------------------------------------------------------------===//

def COND_SGT : PatLeaf <(cond), [{return N->get() == ISD::SETGT;}]>;
def COND_SGE : PatLeaf <(cond), [{return N->get() == ISD::SETGE;}]>;
def COND_SLT : PatLeaf <(cond), [{return N->get() == ISD::SETLT;}]>;
def COND_SLE : PatLeaf <(cond), [{return N->get() == ISD::SETLE;}]>;

//===----------------------------------------------------------------------===//
// PatLeafs for integer equality
//===----------------------------------------------------------------------===//

def COND_EQ : PatLeaf <
  (cond),
  [{return N->get() == ISD::SETEQ || N->get() == ISD::SETUEQ;}]
>;

def COND_NE : PatLeaf <
  (cond),
  [{return N->get() == ISD::SETNE || N->get() == ISD::SETUNE;}]
>;

def COND_NULL : PatLeaf <
  (cond),
  [{return false;}]
>;

//===----------------------------------------------------------------------===//
// Load/Store Pattern Fragments
//===----------------------------------------------------------------------===//

def az_extload : PatFrag<(ops node:$ptr), (unindexedload node:$ptr), [{
  LoadSDNode *L = cast<LoadSDNode>(N);
  return L->getExtensionType() == ISD::ZEXTLOAD ||
         L->getExtensionType() == ISD::EXTLOAD;
}]>;

def az_extloadi8 : PatFrag<(ops node:$ptr), (az_extload node:$ptr), [{
  return cast<LoadSDNode>(N)->getMemoryVT() == MVT::i8;
}]>;

def az_extloadi8_global : PatFrag<(ops node:$ptr), (az_extloadi8 node:$ptr), [{
    return isGlobalLoad(dyn_cast<LoadSDNode>(N));
}]>;

def sextloadi8_global : PatFrag<(ops node:$ptr), (sextloadi8 node:$ptr), [{
    return isGlobalLoad(dyn_cast<LoadSDNode>(N));
}]>;

def az_extloadi8_constant : PatFrag<(ops node:$ptr), (az_extloadi8 node:$ptr), [{
    return isConstantLoad(dyn_cast<LoadSDNode>(N), -1);
}]>;

def sextloadi8_constant : PatFrag<(ops node:$ptr), (sextloadi8 node:$ptr), [{
    return isConstantLoad(dyn_cast<LoadSDNode>(N), -1);
}]>;

def az_extloadi8_local : PatFrag<(ops node:$ptr), (az_extloadi8 node:$ptr), [{
    return isLocalLoad(dyn_cast<LoadSDNode>(N));
}]>;

def sextloadi8_local : PatFrag<(ops node:$ptr), (sextloadi8 node:$ptr), [{
    return isLocalLoad(dyn_cast<LoadSDNode>(N));
}]>;

def az_extloadi16 : PatFrag<(ops node:$ptr), (az_extload node:$ptr), [{
  return cast<LoadSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;

def az_extloadi16_global : PatFrag<(ops node:$ptr), (az_extloadi16 node:$ptr), [{
    return isGlobalLoad(dyn_cast<LoadSDNode>(N));
}]>;

def sextloadi16_global : PatFrag<(ops node:$ptr), (sextloadi16 node:$ptr), [{
    return isGlobalLoad(dyn_cast<LoadSDNode>(N));
}]>;

def az_extloadi16_constant : PatFrag<(ops node:$ptr), (az_extloadi16 node:$ptr), [{
    return isConstantLoad(dyn_cast<LoadSDNode>(N), -1);
}]>;

def sextloadi16_constant : PatFrag<(ops node:$ptr), (sextloadi16 node:$ptr), [{
    return isConstantLoad(dyn_cast<LoadSDNode>(N), -1);
}]>;

def az_extloadi16_local : PatFrag<(ops node:$ptr), (az_extloadi16 node:$ptr), [{
    return isLocalLoad(dyn_cast<LoadSDNode>(N));
}]>;

def sextloadi16_local : PatFrag<(ops node:$ptr), (sextloadi16 node:$ptr), [{
    return isLocalLoad(dyn_cast<LoadSDNode>(N));
}]>;

def az_extloadi32 : PatFrag<(ops node:$ptr), (az_extload node:$ptr), [{
  return cast<LoadSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;

def az_extloadi32_global : PatFrag<(ops node:$ptr),
                                   (az_extloadi32 node:$ptr), [{
  return isGlobalLoad(dyn_cast<LoadSDNode>(N));
}]>;

def az_extloadi32_constant : PatFrag<(ops node:$ptr),
                                     (az_extloadi32 node:$ptr), [{
  return isConstantLoad(dyn_cast<LoadSDNode>(N), -1);
}]>;

def truncstorei8_global : PatFrag<(ops node:$val, node:$ptr),
                                  (truncstorei8 node:$val, node:$ptr), [{
  return isGlobalStore(dyn_cast<StoreSDNode>(N));
}]>;

def truncstorei16_global : PatFrag<(ops node:$val, node:$ptr),
                                  (truncstorei16 node:$val, node:$ptr), [{
  return isGlobalStore(dyn_cast<StoreSDNode>(N));
}]>;

def local_store : PatFrag<(ops node:$val, node:$ptr),
                             (store node:$val, node:$ptr), [{
  return isLocalStore(dyn_cast<StoreSDNode>(N));
}]>;

def truncstorei8_local : PatFrag<(ops node:$val, node:$ptr),
                                  (truncstorei8 node:$val, node:$ptr), [{
  return isLocalStore(dyn_cast<StoreSDNode>(N));
}]>;

def truncstorei16_local : PatFrag<(ops node:$val, node:$ptr),
                                  (truncstorei16 node:$val, node:$ptr), [{
  return isLocalStore(dyn_cast<StoreSDNode>(N));
}]>;

def local_load : PatFrag<(ops node:$ptr), (load node:$ptr), [{
    return isLocalLoad(dyn_cast<LoadSDNode>(N));
}]>;

def atomic_load_add_local : PatFrag<(ops node:$ptr, node:$value),
                                    (atomic_load_add node:$ptr, node:$value), [{
  return dyn_cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS;
}]>;

def atomic_load_sub_local : PatFrag<(ops node:$ptr, node:$value),
                                    (atomic_load_sub node:$ptr, node:$value), [{
  return dyn_cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS;
}]>;

def mskor_global : PatFrag<(ops node:$val, node:$ptr),
                            (AMDGPUstore_mskor node:$val, node:$ptr), [{
  return dyn_cast<MemSDNode>(N)->getAddressSpace() == AMDGPUAS::GLOBAL_ADDRESS;
}]>;

class Constants {
int TWO_PI = 0x40c90fdb;
int PI = 0x40490fdb;
int TWO_PI_INV = 0x3e22f983;
int FP_UINT_MAX_PLUS_1 = 0x4f800000;    // 1 << 32 in floating point encoding
}
def CONST : Constants;

def FP_ZERO : PatLeaf <
  (fpimm),
  [{return N->getValueAPF().isZero();}]
>;

def FP_ONE : PatLeaf <
  (fpimm),
  [{return N->isExactlyValue(1.0);}]
>;

def U24 : ComplexPattern<i32, 1, "SelectU24", [], []>;
def I24 : ComplexPattern<i32, 1, "SelectI24", [], []>;

let isCodeGenOnly = 1, isPseudo = 1 in {

let usesCustomInserter = 1  in {

class CLAMP <RegisterClass rc> : AMDGPUShaderInst <
  (outs rc:$dst),
  (ins rc:$src0),
  "CLAMP $dst, $src0",
  [(set f32:$dst, (int_AMDIL_clamp f32:$src0, (f32 FP_ZERO), (f32 FP_ONE)))]
>;

class FABS <RegisterClass rc> : AMDGPUShaderInst <
  (outs rc:$dst),
  (ins rc:$src0),
  "FABS $dst, $src0",
  [(set f32:$dst, (fabs f32:$src0))]
>;

class FNEG <RegisterClass rc> : AMDGPUShaderInst <
  (outs rc:$dst),
  (ins rc:$src0),
  "FNEG $dst, $src0",
  [(set f32:$dst, (fneg f32:$src0))]
>;

} // usesCustomInserter = 1

multiclass RegisterLoadStore <RegisterClass dstClass, Operand addrClass,
                    ComplexPattern addrPat> {
let UseNamedOperandTable = 1 in {

  def RegisterLoad : AMDGPUShaderInst <
    (outs dstClass:$dst),
    (ins addrClass:$addr, i32imm:$chan),
    "RegisterLoad $dst, $addr",
    [(set i32:$dst, (AMDGPUregister_load addrPat:$addr, (i32 timm:$chan)))]
  > {
    let isRegisterLoad = 1;
  }

  def RegisterStore : AMDGPUShaderInst <
    (outs),
    (ins dstClass:$val, addrClass:$addr, i32imm:$chan),
    "RegisterStore $val, $addr",
    [(AMDGPUregister_store i32:$val, addrPat:$addr, (i32 timm:$chan))]
  > {
    let isRegisterStore = 1;
  }
}
}

} // End isCodeGenOnly = 1, isPseudo = 1

/* Generic helper patterns for intrinsics */
/* -------------------------------------- */

class POW_Common <AMDGPUInst log_ieee, AMDGPUInst exp_ieee, AMDGPUInst mul>
  : Pat <
  (fpow f32:$src0, f32:$src1),
  (exp_ieee (mul f32:$src1, (log_ieee f32:$src0)))
>;

/* Other helper patterns */
/* --------------------- */

/* Extract element pattern */
class Extract_Element <ValueType sub_type, ValueType vec_type, int sub_idx, 
                       SubRegIndex sub_reg>
  : Pat<
  (sub_type (vector_extract vec_type:$src, sub_idx)),
  (EXTRACT_SUBREG $src, sub_reg)
>;

/* Insert element pattern */
class Insert_Element <ValueType elem_type, ValueType vec_type,
                      int sub_idx, SubRegIndex sub_reg>
  : Pat <
  (vector_insert vec_type:$vec, elem_type:$elem, sub_idx),
  (INSERT_SUBREG $vec, $elem, sub_reg)
>;

class Vector4_Build <ValueType vecType, ValueType elemType> : Pat <
  (vecType (build_vector elemType:$x, elemType:$y, elemType:$z, elemType:$w)),
  (INSERT_SUBREG (INSERT_SUBREG (INSERT_SUBREG (INSERT_SUBREG
    (vecType (IMPLICIT_DEF)), $x, sub0), $y, sub1), $z, sub2), $w, sub3)
>;

// XXX: Convert to new syntax and use COPY_TO_REG, once the DFAPacketizer
// can handle COPY instructions.
// bitconvert pattern
class BitConvert <ValueType dt, ValueType st, RegisterClass rc> : Pat <
  (dt (bitconvert (st rc:$src0))),
  (dt rc:$src0)
>;

// XXX: Convert to new syntax and use COPY_TO_REG, once the DFAPacketizer
// can handle COPY instructions.
class DwordAddrPat<ValueType vt, RegisterClass rc> : Pat <
  (vt (AMDGPUdwordaddr (vt rc:$addr))),
  (vt rc:$addr)
>;

// BFI_INT patterns

multiclass BFIPatterns <Instruction BFI_INT> {

  // Definition from ISA doc:
  // (y & x) | (z & ~x)
  def : Pat <
    (or (and i32:$y, i32:$x), (and i32:$z, (not i32:$x))),
    (BFI_INT $x, $y, $z)
  >;

  // SHA-256 Ch function
  // z ^ (x & (y ^ z))
  def : Pat <
    (xor i32:$z, (and i32:$x, (xor i32:$y, i32:$z))),
    (BFI_INT $x, $y, $z)
  >;

}

// SHA-256 Ma patterns

// ((x & z) | (y & (x | z))) -> BFI_INT (XOR x, y), z, y
class SHA256MaPattern <Instruction BFI_INT, Instruction XOR> : Pat <
  (or (and i32:$x, i32:$z), (and i32:$y, (or i32:$x, i32:$z))),
  (BFI_INT (XOR i32:$x, i32:$y), i32:$z, i32:$y)
>;

// Bitfield extract patterns

/*

XXX: The BFE pattern is not working correctly because the XForm is not being
applied.

def legalshift32 : ImmLeaf <i32, [{return Imm >=0 && Imm < 32;}]>;
def bfemask : PatLeaf <(imm), [{return isMask_32(N->getZExtValue());}],
                            SDNodeXForm<imm, [{ return CurDAG->getTargetConstant(CountTrailingOnes_32(N->getZExtValue()), MVT::i32);}]>>;

class BFEPattern <Instruction BFE> : Pat <
  (and (srl i32:$x, legalshift32:$y), bfemask:$z),
  (BFE $x, $y, $z)
>;

*/

// rotr pattern
class ROTRPattern <Instruction BIT_ALIGN> : Pat <
  (rotr i32:$src0, i32:$src1),
  (BIT_ALIGN $src0, $src0, $src1)
>;

// 24-bit arithmetic patterns
def umul24 : PatFrag <(ops node:$x, node:$y), (mul node:$x, node:$y)>;

/*
class UMUL24Pattern <Instruction UMUL24> : Pat <
  (mul U24:$x, U24:$y),
  (UMUL24 $x, $y)
>;
*/

include "R600Instructions.td"

include "SIInstrInfo.td"