-
Notifications
You must be signed in to change notification settings - Fork 8
/
Copy pathcode_gen_cpu_int8.py
executable file
·524 lines (440 loc) · 19.1 KB
/
code_gen_cpu_int8.py
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
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
# this program basically does a constexpr and generates cuda code
import textwrap
import numpy as np
from code_fragments import *
from utils import *
import argparse
parser = argparse.ArgumentParser(description='CodeGen V1')
parser.add_argument('--A_dim', type=int, default=12)
parser.add_argument('--B_dim', type=int, default=12)
parser.add_argument('--C_dim', type=int, default=12)
parser.add_argument('--AT', type=int, default=12)
parser.add_argument('--C_blocks', type=int, default=12)
parser.add_argument('--CT',type=int, default=1)
parser.add_argument('--B_blocks',type=int,default=1)
parser.add_argument('--Gy', type=int, default=12)
parser.add_argument('--infile', default=None, type=str)
parser.add_argument('--infile_bias', default=None, type=str)
parser.add_argument('--outfile', default=None, type=str)
parser.add_argument('--outfile_asm', default= None, type = str)
parser.add_argument('--in_format', default="NCHW",type=str)
parser.add_argument('--out_format', default="NCHW",type=str)
parser.add_argument('--Tsb',type=float,default=1)
parser.add_argument('--fuse',default=False,action='store_true')
parser.add_argument('--x86',default=False,action='store_true')
parser.add_argument('--arm',default=False,action='store_true')
parser.add_argument('--threads',type = int, default=4)
parser.add_argument('--avx512',default=False,action='store_true')
parser.add_argument('--no_relu',default=False,action='store_true')
parser.add_argument('--no_row_skip',default=False,action='store_true')
args = parser.parse_args()
GY = args.Gy
FUSE_END = args.fuse
NO_RELU = args.no_relu
print(FUSE_END)
TSB_MULT = args.Tsb
A_dim = args.A_dim
B_dim = args.B_dim
C_dim = args.C_dim
THREADS = args.threads
AT = args.AT
C_blocks = args.C_blocks
AVX512 = False
if args.avx512:
AVX512 = True
input_file = args.infile
outfile = args.outfile
outfile_asm = args.outfile_asm
#assert C_dim % C_blocks == 0
GSY = C_dim // C_blocks
TSB =int( GSY * TSB_MULT)
TSZ = C_dim // C_blocks if C_dim % C_blocks == 0 else C_dim // C_blocks + 1
X86 = args.x86
ARM = args.arm
NRS = args.no_row_skip
CT = args.CT
B_blocks = args.B_blocks
assert CT %4 == 0
assert not (X86 and ARM)
if X86:
print("Generating X86 vector intrinsics")
elif ARM:
print("Generating Arm intrinsics")
else:
assert False
if AVX512:
VEC = 16
else:
VEC = 8
IN_FORMAT = args.in_format
OUT_FORMAT = args.out_format
if IN_FORMAT == "NHWC" or OUT_FORMAT == "NHWC":
assert False
input_file_bias = args.infile_bias
if input_file_bias:
bias = np.load(input_file_bias)
#global AB_vals
AB_vals = []
A_idx = []
B_idx = []
AB_block_offs = [0]
#global off
off = 0
if X86:
LOAD_CACHE_ASM = """
vmovdqu8 IDX1(%r8,%r11,1), %xmmNUM;
vmovdqu8 IDX3(%r8,%r11,1), %xmm29;
vbroadcasti32x4 IDX2(%r8,%r11,1),%ymmNUM {%k1};
vbroadcasti32x4 IDX4(%r8,%r11,1), %ymm29 {%k1};
vpermt2d %zmm29, %zmm27, %zmmNUM;
vpshufb %zmm28, %zmmNUM, %zmmNUM;
"""
elif ARM:
LOAD_CACHE = """
RC = vld1q_f32(&BC[IDX + C_offset + lane]);
"""
if X86:
if AVX512:
LOAD_WEIGHT_ASM = """vpbroadcastd OFF(%rcx), %zmm31;
"""
MAIN_PROGRAM_ASM="""vpdpbusd %zmmNUM,%zmm31,%zmmTAR;
"""
elif ARM:
MAIN_PROGRAM ="""
val = vdupq_n_f32(VAL);
ACC[IDX_1] = vmlaq_f32(ACC[IDX_1], RC, val);
"""
def emit_load_block(index, currloadreg):
return LOAD_CACHE_ASM.replace("IDX1",str(index[0])).replace("IDX2",str(index[1])). \
replace("IDX3",str(index[2])).replace("IDX4",str(index[3])).replace("NUM",str(currloadreg))
def emit_compute_block(Ny_idx,vals,currloadreg, virg=False):
global off
new_block_asm = LOAD_WEIGHT_ASM.replace("OFF",str(off * 4 ))
for i in range(CT):
new_block_asm += MAIN_PROGRAM_ASM.replace("TAR",str(Ny_idx +i * AT)).replace("NUM",str(currloadreg - i))
global AB_vals
AB_vals.extend(vals)
global A_idx
A_idx.extend([Ny_idx] * 4)
off += 1
return new_block_asm
def ny_to_a(ny_idx,groupId,blockId, A_dim = None, A_offset = None):
if A_offset is None:
A_offset = blockId * (AT)
return A_offset + ny_idx
def generate_from_B(Ny_indices, B_indices,BA,block,NY,BB_offset, GY = None,A_offset=None):
program = ""
asm = """
..B1.NUM1:
xorl %r10d, %r10d;
..B1.NUM2:
imul $16, %r10d, %r11d;
add %r9d, %r11d;
movslq %r11d, %r11;
add $CT, %r10d;
""".replace("NUM1",str(BB_offset + block*2+2)).replace("NUM2",str(BB_offset + block * 2 + 3)).replace("STRIDE",str(8)).replace("CT",str(CT))
#print(A_offset)
if input_file_bias is not None:
for i in range(NY):
for j in range(CT):
if BB_offset > 0:
asm += "\t\tvmovups " + str(mapping[A_offset + i] * C_dim * 4 + j * VEC * 4) + "(%rdx,%r11,4)" + (",%zmm" if AVX512 else ",%ymm") + str(i + j * AT) + ";\n"
else:
asm += "\tvpbroadcastd " + str(mapping[A_offset+i] * 4) + "(%rsi), %zmm" + str(i + AT * j) + ";\n"
#asm += "\tvpbroadcastd " + "%xmm" + str(i + AT * j) + (", %zmm" if AVX512 else "(%rsi), %ymm") + str(i + AT * j) + ";\n"
else:
for i in range(NY):
for j in range(CT):
if BB_offset > 0:
asm += "\t\tvmovups " + str(mapping[A_offset + i] * C_dim * 4 + j * VEC * 4) + "(%rdx,%r11,4)" + (",%zmm" if AVX512 else ",%ymm") + str(i + j * AT) + ";\n"
else:
asm += "\tvxorps " + ("%zmm" if AVX512 else "%ymm") + str(i + AT * j) + "," + \
("%zmm" if AVX512 else "%ymm") + str(i + AT * j) + "," + ("%zmm" if AVX512 else "%ymm") + str(i + AT * j) + ";\n"
done = set()
loads = ""
computes = ""
if AVX512:
TOK = 24
else:
TOK = 13
currloadreg = TOK
# pad end of the zipped list
padded_Ny_indices = []
padded_B_indices = []
counter = 0
old_B_idx = -1
for ny_idx, b_idx in zip(Ny_indices[0],B_indices[0]):
assert ny_idx == 0 # for now. We are going to handle AT for quantized at a later date if at all.
if(b_idx < ny_idx):
pad_len = ((counter - 1) // 4 + 1 ) * 4 - counter
padded_Ny_indices.extend([-1] * pad_len)
padded_B_indices.extend([-1] * pad_len)
padded_Ny_indices.append(ny_idx)
padded_B_indices.append(b_idx)
counter = 1
else:
padded_Ny_indices.append(ny_idx)
padded_B_indices.append(b_idx)
counter += 1
pad_len = ((counter - 1) // 4 + 1 ) * 4 - counter
padded_Ny_indices.extend([-1] * pad_len)
padded_B_indices.extend([-1] * pad_len)
#print(padded_B_indices,len(padded_B_indices))
#for ny_idx, b_idx in zip(padded_Ny_indices[0],padded_B_indices[0]):
for pos in range(0,len(padded_Ny_indices),4):
b_indices = padded_B_indices[pos:pos+4]
currloadreg = TOK #(currloadreg - TOK + 1) % 6 + TOK
asm += loads
asm += computes
loads = ""
computes = ""
ny_idx = 0
a_idx = ny_to_a(ny_idx,0,block,A_dim = A_dim, A_offset=A_offset)
global B_idx
if -1 in b_indices:
assert(b_indices[0] != -1)
for i in range(CT):
load_block_asm = """
vxorps %zmm29, %zmm29, % zmm29;
vxorps %zmmNUM, %zmmNUM, %zmmNUM;
vmovdqu8 IDX1(%r8,%r11,1), %xmmNUM;
""".replace("IDX1",str(b_indices[0] * C_dim + i * VEC)).replace("NUM",str(currloadreg-i))
if b_indices[2] != -1:
load_block_asm += """
vmovdqu8 IDX3(%r8,%r11,1), %xmm29;
""".replace("IDX3",str(b_indices[2] * C_dim + i * VEC))
if b_indices[1] != -1:
load_block_asm += """
vbroadcasti32x4 IDX2(%r8,%r11,1),%ymmNUM {%k1};
""".replace("IDX2",str(b_indices[1] * C_dim + i * VEC)).replace("NUM",str(currloadreg-i))
if b_indices[3] != -1:
load_block_asm += """
vbroadcasti32x4 IDX3(%r8,%r11,1),%ymm29 {%k1};
""".replace("IDX3",str(b_indices[3] * C_dim + i * VEC))
load_block_asm += """
vpermt2d %zmm29, %zmm27, %zmmNUM;
vpshufb %zmm28, %zmmNUM, %zmmNUM;
""".replace("NUM",str(currloadreg-i))
loads += load_block_asm
#print(b_indices)
num_vals = np.where(np.array(b_indices) == -1)[0][0]
#print(num_vals)
values = np.array([BA[b_indices[i],a_idx] for i in range(num_vals)] + [0 for j in range(4-num_vals)]).astype(np.int8)
#print(values)
B_idx.extend(b_indices)
compute_block_asm = emit_compute_block(ny_idx , values, currloadreg , virg = ny_idx not in done)
computes += compute_block_asm
else:
for i in range(CT):
load_block_asm = emit_load_block([k * C_dim + i * VEC for k in b_indices], currloadreg - i)
loads += load_block_asm
values = BA[b_indices,a_idx]
B_idx.extend(b_indices)
compute_block_asm = emit_compute_block(ny_idx , values, currloadreg , virg = ny_idx not in done)
computes += compute_block_asm
done.add(ny_idx)
asm += loads
asm += computes
#print(block,group)
#program += GROUP_CONTROL_END + "\n"
global AB_block_offs
AB_block_offs.append(len(AB_vals))
return program, asm, done
def get_idx_balanced(block,BA,A_offset,block_NY,B_bounds = [0,B_dim], GY=None):
BA = BA[B_bounds[0]:B_bounds[1]]
Ny_indices = [[] for i in range(GY)]
B_indices = [[] for i in range(GY)]
nnz = np.sum(np.abs(BA[:,A_offset:A_offset + block_NY]) > EPS )
nnz_per_group = nnz // GY
curr_group = 0
curr_nnz = 0
for B_idx in range(B_dim // B_blocks):
for ny in range(block_NY):
assert curr_group < GY
A_idx = ny_to_a(ny,curr_group,block,A_dim = A_dim, A_offset=A_offset)
if np.abs(BA[B_idx,A_idx]) > EPS:
B_indices[curr_group].append(B_idx + B_bounds[0])
Ny_indices[curr_group].append(ny)
curr_nnz += 1
if curr_nnz > nnz_per_group:
curr_group += 1
curr_nnz = 0
return Ny_indices, B_indices
def no_load_balance(BA):
#assert A_dim % A_blocks == 0
interval = AT
bounds = [interval * i for i in range(A_blocks)] + [A_dim]
return bounds , interval
def load_balancer2(BA):
total_nnz = (np.abs(BA) > EPS).sum()
nnz_per_block = total_nnz / A_blocks
sums = np.sum(np.abs(BA) > EPS, axis = 0)
cs = np.cumsum(sums)
bounds = [np.argmax(cs > nnz_per_block * i) for i in range(A_blocks)]
bounds = bounds + [A_dim]
nnzs = np.diff(bounds)
NY = np.max(nnzs)
return bounds, NY
# name is the name of the numpy file
def gencode(BA,outfile,C_dim,A_blocks,C_blocks,GY,name=None):
program = ""
asm_program = """
# -- Begin _spmm
.text
# mark_begin;
.align 16,0x90
.globl _spmm
# --- mm(void *)
_spmm:
# parameter 1: %rdi
..B1.1: # Preds ..B1.0
# Execution count [9.00e-01]
.cfi_startproc
..___tag_value__spmm.1:
..L2:
#45.1
pushq %rbp #45.1
.cfi_def_cfa_offset 16
movq %rsp, %rbp #45.1
.cfi_def_cfa 6, 16
.cfi_offset 6, -16
andq $-32, %rsp #45.1
subq $96, %rsp #45.1
movq (%rdi), %rcx #47.38
movq 8(%rdi), %rsi #48.46
movq 16(%rdi), %r8 #49.41
movq 24(%rdi), %rdx #50.22
movl 36(%rdi), %eax
movl 32(%rdi), %edi #51.21
vxorps ZERO, ZERO, ZERO #59.19
decl %eax
decl %edi
imul $TSZ, %eax, %r9d
mov $0xf0 , %ebx;
kmovb %ebx, %k1
vpmovzxbd vpermt2d_control(%rip), % zmm27;
vbroadcasti32x4 vpshufb_control(%rip), % zmm28;
""".replace("BOUND",str(C_blocks//THREADS)).replace("TSZ",str(TSZ)).replace("ZERO","%zmm30" if AVX512 else "%ymm14")
#assert A_dim % A_blocks == 0
#assert C_dim % C_blocks == 0
B_dim = BA.shape[0]
# if IN_FORMAT == "NCHW" and OUT_FORMAT == "NCHW":
# bounds, NY = load_balancer2(BA)
# else:
bounds, NY = no_load_balance(BA)
program += START_NONFUSED.replace("OUTPUT_FORMAT",OUT_FORMAT).replace("INPUT_FORMAT",IN_FORMAT).replace("Ny",str(NY)).replace("GY",str(GY)).replace("A_dim",str(A_dim)).replace(
"C_dim",str(C_dim)).replace("B_dim",str(B_dim)).replace("A_BLOCKS",str(A_blocks)).replace("C_BLOCKS",str(C_blocks)).replace("BOUND",str(C_blocks//4)).replace("X86_DEF",str(int(X86))).replace("ARM_DEF",str(int(ARM))) + "\n"
assert B_dim % B_blocks == 0
block_size = B_dim // B_blocks
for b_block in range(B_blocks):
bb_offset = b_block * A_blocks * 2
for block in range(A_blocks):
A_offset = bounds[block]
block_NY = bounds[block+1] - A_offset
program += BLOCK_CONTROL_START.replace("BLOCK", str(block)).replace("Ny",str(block_NY)) + "\n"
Ny_indices, B_indices = get_idx_balanced(block,BA,A_offset,block_NY,B_bounds = [b_block * block_size, (b_block + 1) * block_size],GY=GY)
#import pdb;pdb.set_trace()
ccode, asm, done = generate_from_B(Ny_indices,B_indices,BA,block,block_NY,bb_offset, GY=GY,A_offset=A_offset)
#ccode = generate_c_stem(block_NY)
program += textwrap.indent(ccode,"\t") + "\n"
asm_program += textwrap.indent(asm,"\t") + "\n"
if OUT_FORMAT == "NCHW":
if FUSE_END:
if GY > 1:
print("End fusion strategy not valid.")
for i in range(block_NY):
program += BLOCK_END_REDUCTION.replace("OFFSET",str(mapping[A_offset + i] * C_dim)).replace("IDX",str(i)).replace("BIAS",str(A_offset+i))
if not NO_RELU:
for j in range(CT):
asm_program += "\t\tvmaxps %ymm" + str(i + j * AT) + (", %zmm30," if AVX512 else ", %ymm14,") + "%ymm" + str(i + j * AT) + ";\n"
for j in range(CT):
asm_program += "\t\tvmovdqu32 %ymm" + str(i + j * AT) + ", " + str(mapping[A_offset + i] * C_dim * 4 + j * VEC * 4) + "(%rdx,%r11,4);\n"
asm_program += """
cmpl $END, %r10d;
jb ..B1.NUM;
""".replace("NUM",str(bb_offset + block * 2 + 3)).replace("END",str(TSZ // VEC))
program += "\t}"
else:
if b_block == B_blocks - 1:
for i in range(block_NY):
for j in range(CT):
asm_program += "\t\tvpmovdb %zmm" + str(i + j * AT) + "," + "%xmm" + str(i + j * AT) + ";\n"
asm_program += """
vinserti32x4 $1,%xmm1,%zmm0,%zmm0;
vinserti32x4 $2,%xmm2,%zmm0,%zmm0;
vinserti32x4 $3,%xmm3,%zmm0,%zmm0;
"""
asm_program += "vmovdqu32 %zmm" + str(i) + ", " + str(mapping[A_offset + i] * C_dim ) + "(%rdx,%r11,1);\n"
else:
for i in range(block_NY):
for j in range(CT):
asm_program += "\t\tvmovdqu32 %zmm" + str(i + j * AT) + ", " + str(mapping[A_offset + i] * C_dim * 4 + j * VEC * 4) + "(%rdx,%r11,4);\n"
asm_program += """
cmpl $END, %r10d;
jb ..B1.NUM;
""".replace("NUM",str(bb_offset + block * 2 + 3)).replace("END",str(TSZ // VEC))
else:
program += BLOCK_END_NHWC.replace("A_offset",str(A_offset)).replace("Ny",str(block_NY)).replace("A_BLOCKS",str(A_blocks)).replace(
"C_BLOCKS", str(C_blocks)).replace("A_dim",str(A_dim)).replace("C_dim",str(C_dim)).replace("B_dim",str(B_dim)) + "\n"
# program += BLOCK_CONTROL_END
program += END_NONFUSED.replace("AB_sparse_tidy.npy",name)
open(outfile,"w").write(program.replace("B_dim",str(B_dim)))
asm_program += """
..B1.NUM1: # Preds ..B1.17
# Execution count [2.80e+01]
decl %eax #44.37
subl $TSZ, %r9d #44.37
cmpl %eax, %edi #44.33
jl ..B1.2 # Prob 96% #44.33
# LOE rcx rbx rbp rsi rdi r12 r13 r14 r15 eax dl ymm15
..B1.NUM2: # Preds ..B1.18
# Execution count [1.00e+00]
vzeroupper #2398.1
movq %rbp, %rsp
popq %rbp
#call pthread_exit@PLT #2416.1
ret
..___tag_value__spmm.13:
.align 16,0x90
# LOE
.cfi_endproc
# mark_end;
.type _spmm,@function
#.size _spmm,-_spmm
..LN_spmm.0:
.section .rodata
.balign 32
vpermt2d_control: .byte 0,4,16,20, 1,5,17,21, 2, 6, 18, 22,3,7,19,23
vpshufb_control: .byte 0,4,8,12, 1,5,9,13, 2,6,10,14, 3,7,11,15
# -- End _spmm
""".replace("TSZ",str(TSZ)).replace("CBLOCKS",str(C_blocks)).replace("NUM1",str(B_blocks * A_blocks *2 + 2)).replace("NUM2",str(B_blocks * A_blocks * 2 + 3))
if AVX512:
asm_program = asm_program.replace("ymm","zmm")
open(outfile_asm,"w").write(asm_program)
BA = np.load(input_file)
print(BA.shape)
BA = BA.squeeze()
"""
We are going to redo BA here to remove some empty rows
"""
if NRS:
A_dim = BA.shape[1]
mapping = {i:i for i in range(A_dim)}
else:
nnz_cols = np.unique(np.where(BA)[1])
mapping = {i : nnz_cols[i] for i in range(len(nnz_cols))}
#print(mapping)
BA = BA[:,nnz_cols]
A_dim = len(nnz_cols)
if A_dim % AT == 0:
A_blocks = A_dim // AT
else:
A_blocks = A_dim // AT + 1
print("Reduced A dimension " + str(A_dim))
gencode(BA,outfile,C_dim,A_blocks,C_blocks,GY,name=input_file)
np.save("AB_vals.npy",np.array(AB_vals))
np.save("AB_block_off.npy",np.array(AB_block_offs).astype(np.int32))
np.save("A_idx.npy",np.array(A_idx).astype(np.int32))
np.save("B_idx.npy",np.array(B_idx).astype(np.int32))
if input_file_bias:
np.save("bias.npy",bias.squeeze())