Skip to content

Commit b495e54

Browse files
authored
Fix declaration of input arguments in the x86_64 SCAL microkernels (#1966)
* Tag arguments 0 and 1 as both input and output (see #1964)
1 parent d5e6940 commit b495e54

9 files changed

+111
-111
lines changed

kernel/x86_64/cscal_microk_bulldozer-2.c

Lines changed: 16 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -116,11 +116,11 @@ static void cscal_kernel_16( BLASLONG n, FLOAT *alpha, FLOAT *x)
116116
"vzeroupper \n\t"
117117

118118
:
119-
:
120-
"r" (n), // 0
121-
"r" (x), // 1
119+
"+r" (n), // 0
120+
"+r" (x) // 1
121+
:
122122
"r" (alpha) // 2
123-
: "cc", //"%0", "%1",
123+
: "cc",
124124
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
125125
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
126126
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -208,11 +208,11 @@ static void cscal_kernel_16_zero_r( BLASLONG n, FLOAT *alpha, FLOAT *x)
208208
"vzeroupper \n\t"
209209

210210
:
211-
:
212-
"r" (n), // 0
213-
"r" (x), // 1
211+
"+r" (n), // 0
212+
"+r" (x) // 1
213+
:
214214
"r" (alpha) // 2
215-
: "cc", //"%0", "%1",
215+
: "cc",
216216
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
217217
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
218218
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -285,11 +285,11 @@ static void cscal_kernel_16_zero_i( BLASLONG n, FLOAT *alpha, FLOAT *x)
285285
"vzeroupper \n\t"
286286

287287
:
288-
:
289-
"r" (n), // 0
290-
"r" (x), // 1
288+
"+r" (n), // 0
289+
"+r" (x) // 1
290+
:
291291
"r" (alpha) // 2
292-
: "cc", //"%0", "%1",
292+
: "cc",
293293
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
294294
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
295295
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -330,11 +330,11 @@ static void cscal_kernel_16_zero( BLASLONG n, FLOAT *alpha, FLOAT *x)
330330
"vzeroupper \n\t"
331331

332332
:
333-
:
334-
"r" (n), // 0
335-
"r" (x), // 1
333+
"+r" (n), // 0
334+
"+r" (x) // 1
335+
:
336336
"r" (alpha) // 2
337-
: "cc", //"%0", "%1",
337+
: "cc",
338338
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
339339
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
340340
"%xmm8", "%xmm9", "%xmm10", "%xmm11",

kernel/x86_64/cscal_microk_haswell-2.c

Lines changed: 15 additions & 15 deletions
Original file line numberDiff line numberDiff line change
@@ -116,11 +116,11 @@ static void cscal_kernel_16( BLASLONG n, FLOAT *alpha, FLOAT *x)
116116
"vzeroupper \n\t"
117117

118118
:
119-
:
120-
"r" (n), // 0
121-
"r" (x), // 1
119+
"+r" (n), // 0
120+
"+r" (x) // 1
121+
:
122122
"r" (alpha) // 2
123-
: "cc", //"0", "1",
123+
: "cc",
124124
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
125125
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
126126
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -208,9 +208,9 @@ static void cscal_kernel_16_zero_r( BLASLONG n, FLOAT *alpha, FLOAT *x)
208208
"vzeroupper \n\t"
209209

210210
:
211-
:
212-
"r" (n), // 0
213-
"r" (x), // 1
211+
"+r" (n), // 0
212+
"+r" (x) // 1
213+
:
214214
"r" (alpha) // 2
215215
: "cc", // "0", "1",
216216
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
@@ -285,9 +285,9 @@ static void cscal_kernel_16_zero_i( BLASLONG n, FLOAT *alpha, FLOAT *x)
285285
"vzeroupper \n\t"
286286

287287
:
288-
:
289-
"r" (n), // 0
290-
"r" (x), // 1
288+
"+r" (n), // 0
289+
"+r" (x) // 1
290+
:
291291
"r" (alpha) // 2
292292
: "cc", //"%0", "%1",
293293
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
@@ -329,12 +329,12 @@ static void cscal_kernel_16_zero( BLASLONG n, FLOAT *alpha, FLOAT *x)
329329

330330
"vzeroupper \n\t"
331331

332-
:
333-
:
334-
"r" (n), // 0
335-
"r" (x), // 1
332+
:
333+
"+r" (n), // 0
334+
"+r" (x) // 1
335+
:
336336
"r" (alpha) // 2
337-
: "cc", //"0", "1",
337+
: "cc",
338338
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
339339
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
340340
"%xmm8", "%xmm9", "%xmm10", "%xmm11",

kernel/x86_64/cscal_microk_steamroller-2.c

Lines changed: 16 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -117,11 +117,11 @@ static void cscal_kernel_16( BLASLONG n, FLOAT *alpha, FLOAT *x)
117117
"vzeroupper \n\t"
118118

119119
:
120-
:
121-
"r" (n), // 0
122-
"r" (x), // 1
120+
"+r" (n), // 0
121+
"+r" (x) // 1
122+
:
123123
"r" (alpha) // 2
124-
: "cc", //"0", "1",
124+
: "cc",
125125
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
126126
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
127127
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -208,12 +208,12 @@ static void cscal_kernel_16_zero_r( BLASLONG n, FLOAT *alpha, FLOAT *x)
208208

209209
"vzeroupper \n\t"
210210

211+
:
212+
"+r" (n), // 0
213+
"+r" (x) // 1
211214
:
212-
:
213-
"r" (n), // 0
214-
"r" (x), // 1
215215
"r" (alpha) // 2
216-
: "cc", //"0", "1",
216+
: "cc",
217217
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
218218
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
219219
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -286,11 +286,11 @@ static void cscal_kernel_16_zero_i( BLASLONG n, FLOAT *alpha, FLOAT *x)
286286
"vzeroupper \n\t"
287287

288288
:
289-
:
290-
"r" (n), // 0
291-
"r" (x), // 1
289+
"+r" (n), // 0
290+
"+r" (x) // 1
291+
:
292292
"r" (alpha) // 2
293-
: "cc", //"%0", "%1",
293+
: "cc",
294294
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
295295
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
296296
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -331,11 +331,11 @@ static void cscal_kernel_16_zero( BLASLONG n, FLOAT *alpha, FLOAT *x)
331331
"vzeroupper \n\t"
332332

333333
:
334-
:
335-
"r" (n), // 0
336-
"r" (x), // 1
334+
"+r" (n), // 0
335+
"+r" (x) // 1
336+
:
337337
"r" (alpha) // 2
338-
: "cc", //"0", "1",
338+
: "cc",
339339
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
340340
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
341341
"%xmm8", "%xmm9", "%xmm10", "%xmm11",

kernel/x86_64/dscal_microk_bulldozer-2.c

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -122,9 +122,9 @@ static void dscal_kernel_8( BLASLONG n, FLOAT *alpha, FLOAT *x)
122122
"vzeroupper \n\t"
123123

124124
:
125-
:
126-
"r" (n1), // 0
127-
"r" (x), // 1
125+
"+r" (n1), // 0
126+
"+r" (x) // 1
127+
:
128128
"r" (alpha), // 2
129129
"r" (n2) // 3
130130
: "cc",
@@ -188,9 +188,9 @@ static void dscal_kernel_8_zero( BLASLONG n, FLOAT *alpha, FLOAT *x)
188188
"vzeroupper \n\t"
189189

190190
:
191-
:
192-
"r" (n1), // 0
193-
"r" (x), // 1
191+
"+r" (n1), // 0
192+
"+r" (x) // 1
193+
:
194194
"r" (alpha), // 2
195195
"r" (n2) // 3
196196
: "cc",

kernel/x86_64/dscal_microk_haswell-2.c

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -122,9 +122,9 @@ static void dscal_kernel_8( BLASLONG n, FLOAT *alpha, FLOAT *x)
122122
"vzeroupper \n\t"
123123

124124
:
125-
:
126-
"r" (n1), // 0
127-
"r" (x), // 1
125+
"+r" (n1), // 0
126+
"+r" (x) // 1
127+
:
128128
"r" (alpha), // 2
129129
"r" (n2) // 3
130130
: "cc",
@@ -187,10 +187,10 @@ static void dscal_kernel_8_zero( BLASLONG n, FLOAT *alpha, FLOAT *x)
187187

188188
"vzeroupper \n\t"
189189

190+
:
191+
"+r" (n1), // 0
192+
"+r" (x) // 1
190193
:
191-
:
192-
"r" (n1), // 0
193-
"r" (x), // 1
194194
"r" (alpha), // 2
195195
"r" (n2) // 3
196196
: "cc",

kernel/x86_64/dscal_microk_sandy-2.c

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -122,9 +122,9 @@ static void dscal_kernel_8( BLASLONG n, FLOAT *alpha, FLOAT *x)
122122
"vzeroupper \n\t"
123123

124124
:
125-
:
126-
"r" (n1), // 0
127-
"r" (x), // 1
125+
"+r" (n1), // 0
126+
"+r" (x) // 1
127+
:
128128
"r" (alpha), // 2
129129
"r" (n2) // 3
130130
: "cc",
@@ -187,10 +187,10 @@ static void dscal_kernel_8_zero( BLASLONG n, FLOAT *alpha, FLOAT *x)
187187

188188
"vzeroupper \n\t"
189189

190+
:
191+
"+r" (n1), // 0
192+
"+r" (x) // 1
190193
:
191-
:
192-
"r" (n1), // 0
193-
"r" (x), // 1
194194
"r" (alpha), // 2
195195
"r" (n2) // 3
196196
: "cc",

kernel/x86_64/zscal_microk_bulldozer-2.c

Lines changed: 14 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -116,11 +116,11 @@ static void zscal_kernel_8( BLASLONG n, FLOAT *alpha, FLOAT *x)
116116
"vzeroupper \n\t"
117117

118118
:
119-
:
120-
"r" (n), // 0
121-
"r" (x), // 1
119+
"+r" (n), // 0
120+
"+r" (x) // 1
121+
:
122122
"r" (alpha) // 2
123-
: "cc", //"%0", "%1",
123+
: "cc",
124124
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
125125
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
126126
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -208,11 +208,11 @@ static void zscal_kernel_8_zero_r( BLASLONG n, FLOAT *alpha, FLOAT *x)
208208
"vzeroupper \n\t"
209209

210210
:
211-
:
212-
"r" (n), // 0
213-
"r" (x), // 1
211+
"+r" (n), // 0
212+
"+r" (x) // 1
213+
:
214214
"r" (alpha) // 2
215-
: "cc", //"%0", "%1",
215+
: "cc",
216216
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
217217
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
218218
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -285,9 +285,9 @@ static void zscal_kernel_8_zero_i( BLASLONG n, FLOAT *alpha, FLOAT *x)
285285
"vzeroupper \n\t"
286286

287287
:
288-
:
289-
"r" (n), // 0
290-
"r" (x), // 1
288+
"+r" (n), // 0
289+
"+r" (x) // 1
290+
:
291291
"r" (alpha) // 2
292292
: "cc", //"%0", "%1",
293293
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
@@ -329,10 +329,10 @@ static void zscal_kernel_8_zero( BLASLONG n, FLOAT *alpha, FLOAT *x)
329329

330330
"vzeroupper \n\t"
331331

332+
:
333+
"+r" (n), // 0
334+
"+r" (x) // 1
332335
:
333-
:
334-
"r" (n), // 0
335-
"r" (x), // 1
336336
"r" (alpha) // 2
337337
: "cc", //"%0", "%1",
338338
"%xmm0", "%xmm1", "%xmm2", "%xmm3",

kernel/x86_64/zscal_microk_haswell-2.c

Lines changed: 16 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -116,11 +116,11 @@ static void zscal_kernel_8( BLASLONG n, FLOAT *alpha, FLOAT *x)
116116
"vzeroupper \n\t"
117117

118118
:
119-
:
120-
"r" (n), // 0
121-
"r" (x), // 1
119+
"+r" (n), // 0
120+
"+r" (x) // 1
121+
:
122122
"r" (alpha) // 2
123-
: "cc", //"%0", "%1",
123+
: "cc",
124124
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
125125
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
126126
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -208,11 +208,11 @@ static void zscal_kernel_8_zero_r( BLASLONG n, FLOAT *alpha, FLOAT *x)
208208
"vzeroupper \n\t"
209209

210210
:
211-
:
212-
"r" (n), // 0
213-
"r" (x), // 1
211+
"+r" (n), // 0
212+
"+r" (x) // 1
213+
:
214214
"r" (alpha) // 2
215-
: "cc", //"%0", "%1",
215+
: "cc",
216216
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
217217
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
218218
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -285,11 +285,11 @@ static void zscal_kernel_8_zero_i( BLASLONG n, FLOAT *alpha, FLOAT *x)
285285
"vzeroupper \n\t"
286286

287287
:
288-
:
289-
"r" (n), // 0
290-
"r" (x), // 1
288+
"+r" (n), // 0
289+
"+r" (x) // 1
290+
:
291291
"r" (alpha) // 2
292-
: "cc", //"%0", "%1",
292+
: "cc",
293293
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
294294
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
295295
"%xmm8", "%xmm9", "%xmm10", "%xmm11",
@@ -330,11 +330,11 @@ static void zscal_kernel_8_zero( BLASLONG n, FLOAT *alpha, FLOAT *x)
330330
"vzeroupper \n\t"
331331

332332
:
333-
:
334-
"r" (n), // 0
335-
"r" (x), // 1
333+
"+r" (n), // 0
334+
"+r" (x) // 1
335+
:
336336
"r" (alpha) // 2
337-
: "cc", //"%0", "%1",
337+
: "cc",
338338
"%xmm0", "%xmm1", "%xmm2", "%xmm3",
339339
"%xmm4", "%xmm5", "%xmm6", "%xmm7",
340340
"%xmm8", "%xmm9", "%xmm10", "%xmm11",

0 commit comments

Comments
 (0)