Skip to content

Commit 32ed6e3

Browse files
author
Chip-Kerchner
committed
Merge branch 'develop' of https://github.com/openmathlib/openblas into develop
2 parents cb9aa2a + ba3bfe8 commit 32ed6e3

5 files changed

Lines changed: 102 additions & 16 deletions

File tree

common_interface.h

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -773,8 +773,8 @@ xdouble BLASFUNC(qlamc3)(xdouble *, xdouble *);
773773

774774
void BLASFUNC(saxpby) (blasint *, float *, float *, blasint *, float *, float *, blasint *);
775775
void BLASFUNC(daxpby) (blasint *, double *, double *, blasint *, double *, double *, blasint *);
776-
void BLASFUNC(caxpby) (blasint *, float *, float *, blasint *, float *, float *, blasint *);
777-
void BLASFUNC(zaxpby) (blasint *, double *, double *, blasint *, double *, double *, blasint *);
776+
void BLASFUNC(caxpby) (blasint *, void *, float *, blasint *, void *, float *, blasint *);
777+
void BLASFUNC(zaxpby) (blasint *, void *, double *, blasint *, void *, double *, blasint *);
778778

779779
void BLASFUNC(somatcopy) (char *, char *, blasint *, blasint *, float *, float *, blasint *, float *, blasint *);
780780
void BLASFUNC(domatcopy) (char *, char *, blasint *, blasint *, double *, double *, blasint *, double *, blasint *);

driver/others/dynamic_power.c

Lines changed: 4 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -69,7 +69,7 @@ static int cpuid(void)
6969
else if (arch == POWER_9) return CPU_POWER9;
7070
#endif
7171
#ifdef POWER_10
72-
else if (arch == POWER_10) return CPU_POWER10;
72+
else if (arch >= POWER_10) return CPU_POWER10;
7373
#endif
7474
return CPU_UNKNOWN;
7575
}
@@ -339,6 +339,9 @@ void gotoblas_dynamic_init(void) {
339339
if (gotoblas && gotoblas -> init) {
340340
strncpy(coren,gotoblas_corename(),20);
341341
sprintf(coremsg, "Core: %s\n",coren);
342+
if (getenv("GET_OPENBLAS_CORETYPE")) {
343+
fprintf(stderr, "%s", coremsg);
344+
}
342345
openblas_warning(2, coremsg);
343346
gotoblas -> init();
344347
} else {

exports/gensymbol

Lines changed: 87 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -60,6 +60,7 @@ cblasobjsc="
6060
cblas_ctbsv cblas_ctpmv cblas_ctpsv cblas_ctrmm cblas_ctrmv cblas_ctrsm cblas_ctrsv
6161
cblas_scnrm2 cblas_scasum cblas_cgemmt
6262
cblas_icamax cblas_icamin cblas_icmin cblas_icmax cblas_scsum cblas_cimatcopy cblas_comatcopy
63+
cblas_caxpyc cblas_crotg cblas_csrot cblas_scamax cblas_scamin
6364
"
6465
cblasobjsd="
6566
cblas_dasum cblas_daxpy cblas_dcopy cblas_ddot
@@ -69,6 +70,7 @@ cblasobjsd="
6970
cblas_dsyr2k cblas_dsyr cblas_dsyrk cblas_dtbmv cblas_dtbsv cblas_dtpmv cblas_dtpsv
7071
cblas_dtrmm cblas_dtrmv cblas_dtrsm cblas_dtrsv cblas_daxpby cblas_dgeadd cblas_dgemmt
7172
cblas_idamax cblas_idamin cblas_idmin cblas_idmax cblas_dsum cblas_dimatcopy cblas_domatcopy
73+
cblas_damax cblas_damin
7274
"
7375

7476
cblasobjss="
@@ -80,6 +82,7 @@ cblasobjss="
8082
cblas_stbmv cblas_stbsv cblas_stpmv cblas_stpsv cblas_strmm cblas_strmv cblas_strsm
8183
cblas_strsv cblas_sgeadd cblas_sgemmt
8284
cblas_isamax cblas_isamin cblas_ismin cblas_ismax cblas_ssum cblas_simatcopy cblas_somatcopy
85+
cblas_samax cblas_samin
8386
"
8487

8588
cblasobjsz="
@@ -91,6 +94,7 @@ cblasobjsz="
9194
cblas_ztrsv cblas_cdotc_sub cblas_cdotu_sub cblas_zdotc_sub cblas_zdotu_sub
9295
cblas_zaxpby cblas_zgeadd cblas_zgemmt
9396
cblas_izamax cblas_izamin cblas_izmin cblas_izmax cblas_dzsum cblas_zimatcopy cblas_zomatcopy
97+
cblas_zaxpyc cblas_zdrot cblas_zrotg cblas_dzamax cblas_dzamin
9498
"
9599

96100
cblasobjs="cblas_xerbla"
@@ -861,6 +865,53 @@ lapackobjs2z="$lapackobjs2z
861865
zgedmd
862866
zgedmdq
863867
"
868+
869+
#functions added post 3.11
870+
871+
lapackobjs2c="$lapackobjs2c
872+
claqp2rk
873+
claqp3rk
874+
ctrsyl3
875+
"
876+
# claqz0
877+
# claqz1
878+
# claqz2
879+
# claqz3
880+
# clatrs3
881+
882+
lapackobjs2d="$lapackobjs2d
883+
dgelqs
884+
dgelst
885+
dgeqp3rk
886+
dgeqrs
887+
dlaqp2rk
888+
dlaqp3rk
889+
dlarmm
890+
dlatrs3
891+
dtrsyl3
892+
"
893+
# dlaqz0
894+
# dlaqz1
895+
# dlaqz2
896+
# dlaqz3
897+
# dlaqz4
898+
899+
lapackobjs2z="$lapackobjs2z
900+
zgelqs
901+
zgelst
902+
zgeqp3rk
903+
zgeqrs
904+
zlaqp2rk
905+
zlaqp3rk
906+
zlatrs3
907+
zrscl
908+
ztrsyl3
909+
"
910+
# zlaqz0
911+
# zlaqz1
912+
# zlaqz2
913+
# zlaqz3
914+
864915
lapack_extendedprecision_objs="
865916
zposvxx clagge clatms chesvxx cposvxx cgesvxx ssyrfssx csyrfsx
866917
dlagsy dsysvxx sporfsx slatms zlatms zherfsx csysvxx
@@ -1622,6 +1673,14 @@ lapackeobjsc="
16221673
LAPACKE_cgetsqrhrt_work
16231674
LAPACKE_cungtsqr_row
16241675
LAPACKE_cungtsqr_row_work
1676+
LAPACKE_clangb
1677+
LAPACKE_clangb_work
1678+
LAPACKE_ctrsyl3
1679+
LAPACKE_ctrsyl3_work
1680+
LAPACKE_ctz_nancheck
1681+
LAPACKE_ctz_trans
1682+
LAPACKE_cunhr_col
1683+
LAPACKE_cunhr_col_work
16251684
"
16261685

16271686
lapackeobjsd="
@@ -2239,6 +2298,14 @@ lapackeobjsd="
22392298
LAPACKE_dgetsqrhrt_work
22402299
LAPACKE_dorgtsqr_row
22412300
LAPACKE_dorgtsqr_row_work
2301+
LAPACKE_dlangb
2302+
LAPACKE_dlangb_work
2303+
LAPACKE_dorhr_col
2304+
LAPACKE_dorhr_col_work
2305+
LAPACKE_dtrsyl3
2306+
LAPACKE_dtrsyl3_work
2307+
LAPACKE_dtz_nancheck
2308+
LAPACKE_dtz_trans
22422309
"
22432310

22442311
lapackeobjss="
@@ -2848,6 +2915,14 @@ lapackeobjss="
28482915
LAPACKE_sgetsqrhrt_work
28492916
LAPACKE_sorgtsqr_row
28502917
LAPACKE_sorgtsqr_row_work
2918+
LAPACKE_slangb
2919+
LAPACKE_slangb_work
2920+
LAPACKE_sorhr_col
2921+
LAPACKE_sorhr_col_work
2922+
LAPACKE_strsyl3
2923+
LAPACKE_strsyl3_work
2924+
LAPACKE_stz_nancheck
2925+
LAPACKE_stz_trans
28512926
"
28522927

28532928
lapackeobjsz="
@@ -3515,6 +3590,14 @@ lapackeobjsz="
35153590
LAPACKE_zgetsqrhrt_work
35163591
LAPACKE_zungtsqr_row
35173592
LAPACKE_zungtsqr_row_work
3593+
LAPACKE_zlangb
3594+
LAPACKE_zlangb_work
3595+
LAPACKE_ztrsyl3
3596+
LAPACKE_ztrsyl3_work
3597+
LAPACKE_ztz_nancheck
3598+
LAPACKE_ztz_trans
3599+
LAPACKE_zunhr_col
3600+
LAPACKE_zunhr_col_work
35183601
"
35193602
## @(SRCX_OBJ) from `lapack-3.4.1/lapacke/src/Makefile`
35203603
## Not exported: requires LAPACKE_EXTENDED to be set and depends on the
@@ -3616,6 +3699,7 @@ lapack_embeded_underscore_objs_s="
36163699
ssysv_aa_2stage ssytrf_aa_2stage
36173700
ssytrs_aa_2stage
36183701
slaorhr_col_getrfnp slaorhr_col_getrfnp2 sorhr_col
3702+
slarfb_gett
36193703
"
36203704
lapack_embeded_underscore_objs_c="
36213705
chetf2_rook chetrf_rook chetri_rook
@@ -3641,6 +3725,7 @@ lapack_embeded_underscore_objs_c="
36413725
csysv_aa_2stage csytrf_aa_2stage
36423726
csytrs_aa_2stage
36433727
claunhr_col_getrfnp claunhr_col_getrfnp2 cunhr_col
3728+
clarfb_gett
36443729
"
36453730
lapack_embeded_underscore_objs_d="
36463731
dlasyf_rook
@@ -3658,6 +3743,7 @@ lapack_embeded_underscore_objs_d="
36583743
dsysv_aa_2stage
36593744
dsytrf_aa_2stage dsytrs_aa_2stage
36603745
dlaorhr_col_getrfnp dlaorhr_col_getrfnp2 dorhr_col
3746+
dlarfb_gett
36613747
"
36623748
lapack_embeded_underscore_objs_z="
36633749
zhetf2_rook zhetrf_rook zhetri_rook
@@ -3682,6 +3768,7 @@ lapack_embeded_underscore_objs_z="
36823768
zhetrs_aa_2stage zsysv_aa_2stage
36833769
zsytrf_aa_2stage zsytrs_aa_2stage
36843770
zlaunhr_col_getrfnp zlaunhr_col_getrfnp2 zunhr_col
3771+
zlarfb_gett
36853772
"
36863773

36873774
dirname=`pwd -P`/../lapack-netlib

interface/zaxpby.c

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -39,12 +39,14 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
3939

4040
#ifndef CBLAS
4141

42-
void NAME(blasint *N, FLOAT *ALPHA, FLOAT *x, blasint *INCX, FLOAT *BETA, FLOAT *y, blasint *INCY)
42+
void NAME(blasint *N, void *VALPHA, FLOAT *x, blasint *INCX, void *VBETA, FLOAT *y, blasint *INCY)
4343
{
4444

4545
blasint n = *N;
4646
blasint incx = *INCX;
4747
blasint incy = *INCY;
48+
FLOAT* ALPHA = (FLOAT*) VALPHA;
49+
FLOAT* BETA = (FLOAT*) VBETA;
4850

4951
#else
5052

test/compare_sgemm_sbgemm.c

Lines changed: 6 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -81,16 +81,6 @@ float16to32 (bfloat16_bits f16)
8181
return f32.v;
8282
}
8383

84-
float
85-
float32to16 (float32_bits f32)
86-
{
87-
bfloat16_bits f16;
88-
f16.bits.s = f32.bits.s;
89-
f16.bits.e = f32.bits.e;
90-
f16.bits.m = (uint32_t) f32.bits.m >> 16;
91-
return f32.v;
92-
}
93-
9484
int
9585
main (int argc, char *argv[])
9686
{
@@ -110,6 +100,8 @@ main (int argc, char *argv[])
110100
float C[m * n];
111101
bfloat16_bits AA[m * k], BB[k * n];
112102
float DD[m * n], CC[m * n];
103+
bfloat16 atmp,btmp;
104+
blasint one=1;
113105

114106
for (j = 0; j < m; j++)
115107
{
@@ -118,8 +110,10 @@ main (int argc, char *argv[])
118110
A[j * k + i] = ((FLOAT) rand () / (FLOAT) RAND_MAX) + 0.5;
119111
B[j * k + i] = ((FLOAT) rand () / (FLOAT) RAND_MAX) + 0.5;
120112
C[j * k + i] = 0;
121-
AA[j * k + i].v = float32to16( A[j * k + i] );
122-
BB[j * k + i].v = float32to16( B[j * k + i] );
113+
sbstobf16_(&one, &A[j*k+i], &one, &atmp, &one);
114+
sbstobf16_(&one, &B[j*k+i], &one, &btmp, &one);
115+
AA[j * k + i].v = atmp;
116+
BB[j * k + i].v = btmp;
123117
CC[j * k + i] = 0;
124118
DD[j * k + i] = 0;
125119
}

0 commit comments

Comments
 (0)