Power10 vector integer multiply, divide, modulo insns

opcodes/
	* ppc-opc.c (powerpc_opcodes): Add vdivuw, vdivud, vdivsw, vmulld,
	vdivsd, vmulhuw, vdiveuw, vmulhud, vdiveud, vmulhsw, vdivesw,
	vmulhsd, vdivesd, vmoduw, vmodud, vmodsw, vmodsd.
gas/
	* testsuite/gas/ppc/vec_mul.s,
	* testsuite/gas/ppc/vec_mul.d: New test.
	* testsuite/gas/ppc/ppc.exp: Run it.
This commit is contained in:
Alan Modra
2020-05-11 09:32:56 +09:30
parent 3ff0a5ba64
commit f4791f1afa
6 changed files with 76 additions and 0 deletions

View File

@ -1,3 +1,9 @@
2020-05-11 Alan Modra <amodra@gmail.com>
* testsuite/gas/ppc/vec_mul.s,
* testsuite/gas/ppc/vec_mul.d: New test.
* testsuite/gas/ppc/ppc.exp: Run it.
2020-05-11 Peter Bergner <bergner@linux.ibm.com>
* testsuite/gas/ppc/byte_rev.d,

View File

@ -132,3 +132,4 @@ if { [supports_ppc64] } then {
run_dump_test "prefix-reloc"
}
run_dump_test "byte_rev"
run_dump_test "vec_mul"

View File

@ -0,0 +1,27 @@
#as: -mpower10
#objdump: -dr -Mpower10
#name: vector integer mul/div/mod
.*
Disassembly of section \.text:
0+0 <_start>:
.*: (10 22 1b 89|89 1b 22 10) vmulhsw v1,v2,v3
.*: (10 85 32 89|89 32 85 10) vmulhuw v4,v5,v6
.*: (10 e8 4b c9|c9 4b e8 10) vmulhsd v7,v8,v9
.*: (11 4b 62 c9|c9 62 4b 11) vmulhud v10,v11,v12
.*: (11 ae 79 c9|c9 79 ae 11) vmulld v13,v14,v15
.*: (12 11 91 8b|8b 91 11 12) vdivsw v16,v17,v18
.*: (12 74 a8 8b|8b a8 74 12) vdivuw v19,v20,v21
.*: (12 d7 c3 8b|8b c3 d7 12) vdivesw v22,v23,v24
.*: (13 3a da 8b|8b da 3a 13) vdiveuw v25,v26,v27
.*: (13 9d f1 cb|cb f1 9d 13) vdivsd v28,v29,v30
.*: (13 e0 08 cb|cb 08 e0 13) vdivud v31,v0,v1
.*: (10 43 23 cb|cb 23 43 10) vdivesd v2,v3,v4
.*: (10 a6 3a cb|cb 3a a6 10) vdiveud v5,v6,v7
.*: (11 09 57 8b|8b 57 09 11) vmodsw v8,v9,v10
.*: (11 6c 6e 8b|8b 6e 6c 11) vmoduw v11,v12,v13
.*: (11 cf 87 cb|cb 87 cf 11) vmodsd v14,v15,v16
.*: (12 32 9e cb|cb 9e 32 12) vmodud v17,v18,v19

View File

@ -0,0 +1,19 @@
.text
_start:
vmulhsw 1,2,3
vmulhuw 4,5,6
vmulhsd 7,8,9
vmulhud 10,11,12
vmulld 13,14,15
vdivsw 16,17,18
vdivuw 19,20,21
vdivesw 22,23,24
vdiveuw 25,26,27
vdivsd 28,29,30
vdivud 31,0,1
vdivesd 2,3,4
vdiveud 5,6,7
vmodsw 8,9,10
vmoduw 11,12,13
vmodsd 14,15,16
vmodud 17,18,19

View File

@ -1,3 +1,9 @@
2020-05-11 Alan Modra <amodra@gmail.com>
* ppc-opc.c (powerpc_opcodes): Add vdivuw, vdivud, vdivsw, vmulld,
vdivsd, vmulhuw, vdiveuw, vmulhud, vdiveud, vmulhsw, vdivesw,
vmulhsd, vdivesd, vmoduw, vmodud, vmodsw, vmodsd.
2020-05-11 Peter Bergner <bergner@linux.ibm.com>
* ppc-opc.c (powerpc_opcodes) <brd, brh, brw>: New mnemonics.

View File

@ -3983,6 +3983,7 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"vcmpnew", VXR(4, 135,0), VXR_MASK, PPCVEC3, 0, {VD, VA, VB}},
{"vmulouw", VX (4, 136), VX_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"vmuluwm", VX (4, 137), VX_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"vdivuw", VX (4, 139), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vmrghw", VX (4, 140), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vpkuhus", VX (4, 142), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"ps_mr", XRC(4, 72,0), XRA_MASK, PPCPS, 0, {FRT, FRB}},
@ -3996,6 +3997,7 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"vrldmi", VX (4, 197), VX_MASK, PPCVEC3, 0, {VD, VA, VB}},
{"vcmpeqfp", VXR(4, 198,0), VXR_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vcmpequd", VXR(4, 199,0), VXR_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"vdivud", VX (4, 203), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vpkuwus", VX (4, 206), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"machhws", XO (4, 108,0,0), XO_MASK, MULHW, 0, {RT, RA, RB}},
{"machhws.", XO (4, 108,0,1), XO_MASK, MULHW, 0, {RT, RA, RB}},
@ -4036,6 +4038,7 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"vcmpnezw", VXR(4, 391,0), VXR_MASK, PPCVEC3, 0, {VD, VA, VB}},
{"vmulosw", VX (4, 392), VX_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"vexptefp", VX (4, 394), VXVA_MASK, PPCVEC, 0, {VD, VB}},
{"vdivsw", VX (4, 395), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vmrglw", VX (4, 396), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vpkshss", VX (4, 398), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"macchwsu", XO (4, 204,0,0), XO_MASK, MULHW, 0, {RT, RA, RB}},
@ -4044,7 +4047,9 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"vsl", VX (4, 452), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vrldnm", VX (4, 453), VX_MASK, PPCVEC3, 0, {VD, VA, VB}},
{"vcmpgefp", VXR(4, 454,0), VXR_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vmulld", VX (4, 457), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vlogefp", VX (4, 458), VXVA_MASK, PPCVEC, 0, {VD, VB}},
{"vdivsd", VX (4, 459), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vpkswss", VX (4, 462), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"macchws", XO (4, 236,0,0), XO_MASK, MULHW, 0, {RT, RA, RB}},
{"macchws.", XO (4, 236,0,1), XO_MASK, MULHW, 0, {RT, RA, RB}},
@ -4134,9 +4139,11 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"evfssqrt", VX_RB_CONST(4, 647, 0), VX_RB_CONST_MASK, PPCEFS2, 0, {RD, RA}},
{"vmuleuw", VX (4, 648), VX_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"evfsmul", VX (4, 648), VX_MASK, PPCSPE, 0, {RS, RA, RB}},
{"vmulhuw", VX (4, 649), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"evfsdiv", VX (4, 649), VX_MASK, PPCSPE, 0, {RS, RA, RB}},
{"evfsnmadd", VX (4, 650), VX_MASK, PPCSPE, 0, {RS, RA, RB}},
{"vrfip", VX (4, 650), VXVA_MASK, PPCVEC, 0, {VD, VB}},
{"vdiveuw", VX (4, 651), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"evfsnmsub", VX (4, 651), VX_MASK, PPCSPE, 0, {RS, RA, RB}},
{"evfscmpgt", VX (4, 652), VX_MASK, PPCSPE, 0, {CRFD, RA, RB}},
{"vspltw", VX (4, 652), VXUIMM2_MASK, PPCVEC, 0, {VD, VB, UIMM2}},
@ -4193,9 +4200,11 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"efssqrt", VX_RB_CONST(4, 711, 0), VX_RB_CONST_MASK,PPCEFS2, 0, {RD, RA}},
{"vcmpgtud", VXR(4, 711,0), VXR_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"efsmul", VX (4, 712), VX_MASK, PPCEFS, 0, {RS, RA, RB}},
{"vmulhud", VX (4, 713), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"efsdiv", VX (4, 713), VX_MASK, PPCEFS, 0, {RS, RA, RB}},
{"efsnmadd", VX (4, 714), VX_MASK, PPCEFS2, 0, {RS, RA, RB}},
{"vrfim", VX (4, 714), VXVA_MASK, PPCVEC, 0, {VD, VB}},
{"vdiveud", VX (4, 715), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"efsnmsub", VX (4, 715), VX_MASK, PPCEFS2, 0, {RS, RA, RB}},
{"efscmpgt", VX (4, 716), VX_MASK, PPCEFS, 0, {CRFD, RA, RB}},
{"vextractd", VX (4, 717), VXUIMM4_MASK, PPCVEC3, 0, {VD, VB, UIMM4}},
@ -4331,8 +4340,10 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"vsraw", VX (4, 900), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vcmpgtsw", VXR(4, 902,0), VXR_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vmulesw", VX (4, 904), VX_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"vmulhsw", VX (4, 905), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vctuxs", VX (4, 906), VX_MASK, PPCVEC, 0, {VD, VB, UIMM}},
{"vcfpuxws", VX (4, 906), VX_MASK, PPCVEC, 0, {VD, VB, UIMM}},
{"vdivesw", VX (4, 907), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vspltisw", VX (4, 908), VXVB_MASK, PPCVEC, 0, {VD, SIMM}},
{"vinsertw", VX (4, 909), VXUIMM4_MASK, PPCVEC3, 0, {VD, VB, UIMM4}},
{"maclhwsu", XO (4, 460,0,0), XO_MASK, MULHW, 0, {RT, RA, RB}},
@ -4341,8 +4352,10 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"vsrad", VX (4, 964), VX_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"vcmpbfp", VXR(4, 966,0), VXR_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vcmpgtsd", VXR(4, 967,0), VXR_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"vmulhsd", VX (4, 969), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vctsxs", VX (4, 970), VX_MASK, PPCVEC, 0, {VD, VB, UIMM}},
{"vcfpsxws", VX (4, 970), VX_MASK, PPCVEC, 0, {VD, VB, UIMM}},
{"vdivesd", VX (4, 971), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vinsertd", VX (4, 973), VXUIMM4_MASK, PPCVEC3, 0, {VD, VB, UIMM4}},
{"vupklpx", VX (4, 974), VXVA_MASK, PPCVEC, 0, {VD, VB}},
{"maclhws", XO (4, 492,0,0), XO_MASK, MULHW, 0, {RT, RA, RB}},
@ -4632,6 +4645,7 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"udi10fcm.", APU(4, 835,0), APU_MASK, PPC440, 0, {URT, URA, URB}},
{"udi10fcm", APU(4, 835,1), APU_MASK, PPC440, 0, {URT, URA, URB}},
{"vsum2sws", VX (4,1672), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vmoduw", VX (4,1675), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vmrgow", VX (4,1676), VX_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"vextuwlx", VX (4,1677), VX_MASK, PPCVEC3, 0, {RT, RA, VB}},
{"vshasigmad", VX (4,1730), VX_MASK, PPCVEC2, 0, {VD, VA, ST, SIX}},
@ -4640,6 +4654,7 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"udi11fcm.", APU(4, 867,0), APU_MASK, PPC440, 0, {URT, URA, URB}},
{"vcmpgtud.", VXR(4, 711,1), VXR_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"udi11fcm", APU(4, 867,1), APU_MASK, PPC440, 0, {URT, URA, URB}},
{"vmodud", VX (4,1739), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vupklsw", VX (4,1742), VXVA_MASK, PPCVEC2, 0, {VD, VB}},
{"vsubsbs", VX (4,1792), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vclzb", VX (4,1794), VXVA_MASK, PPCVEC2, 0, {VD, VB}},
@ -4671,6 +4686,7 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"udi14fcm.", APU(4, 963,0), APU_MASK, PPC440, 0, {URT, URA, URB}},
{"udi14fcm", APU(4, 963,1), APU_MASK, PPC440, 0, {URT, URA, URB}},
{"vsumsws", VX (4,1928), VX_MASK, PPCVEC, 0, {VD, VA, VB}},
{"vmodsw", VX (4,1931), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"vmrgew", VX (4,1932), VX_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"vextuwrx", VX (4,1933), VX_MASK, PPCVEC3, 0, {RT, RA, VB}},
{"maclhwsuo", XO (4, 460,1,0), XO_MASK, MULHW, 0, {RT, RA, RB}},
@ -4681,6 +4697,7 @@ const struct powerpc_opcode powerpc_opcodes[] = {
{"udi15fcm.", APU(4, 995,0), APU_MASK, PPC440, 0, {URT, URA, URB}},
{"vcmpgtsd.", VXR(4, 967,1), VXR_MASK, PPCVEC2, 0, {VD, VA, VB}},
{"udi15fcm", APU(4, 995,1), APU_MASK, PPC440, 0, {URT, URA, URB}},
{"vmodsd", VX (4,1995), VX_MASK, POWER10, 0, {VD, VA, VB}},
{"maclhwso", XO (4, 492,1,0), XO_MASK, MULHW, 0, {RT, RA, RB}},
{"maclhwso.", XO (4, 492,1,1), XO_MASK, MULHW, 0, {RT, RA, RB}},
{"nmaclhwso", XO (4, 494,1,0), XO_MASK, MULHW, 0, {RT, RA, RB}},