Apple M1 Microarchitecture Research by Dougall Johnson

Firestorm: Overview | Base Instructions | SIMD and FP Instructions
Icestorm:  Overview | Base Instructions | SIMD and FP Instructions

SQRSHL (vector, 2S)

Test 1: uops

Code:

  sqrshl v0.2s, v0.2s, v1.2s
  movi v0.16b, 1
  movi v1.16b, 2

(no loop instructions)

1000 unrolls and 1 iteration

Retires: 1.000

Issues: 1.000

Integer unit issues: 0.000

Load/store unit issues: 0.000

SIMD/FP unit issues: 1.000

retire (01)cycle (02)031e3f4e51inst issue (52)~issue fp/simd (54)~dispatch fp/simd (57)huge thing fp/simd (5b)60696d6edispatch stall (70)scheduler rewind (75)scheduler stall (76)~dispatch op (78)~map op fp/simd (7e)~map lookup fp/simd (81)8283pipeline redirect (84)85inst all (8c)inst neon or fp (9a)a0a8accfd5d6ddinst fetch restart (de)e0? fp/simd (ee)f5f6f7f8fd
10043037230103254825100010001000398313030183037303724153289510001000200030373037111001100000073116112630100030383038303830383038
1004303723061254825100010001000398313130183037303724153289510001000200030373037111001100000073116112630100030383038303830383038
1004303723061254825100010001000398313130183037303724153289510001000200030373037111001100020073116112646100030863074312330853085
10043083230612548251000100010003983131301830373037241532895100010002000303730371110011000002473116112630100030383038303830383038
1004303722361254825100010001000398313130183037303724153289910001000200030373037111001100000073116112630100030383038303830383038
1004303722061254825100010001000398313130183037303724153289510001000200030373037111001100000073116112630100030383038303830383038
1004303723061254825100010001000398313130183037303724153289510001000200030373037111001100000073116112630100030383038303830383038
1004303722061254825100010001000398313030183037303724153289510001000200030373037111001100000073116112630100030383038303830383038
1004303723061254825100010001000398313030183037303724153289510001000233630373084211001100000073116112630100030383038303830383038
1004303722061254825100010001000398313030183037303724153289510001000200030373037111001100000073116112630100030383086303830383038

Test 2: Latency 1->2

Code:

  sqrshl v0.2s, v0.2s, v1.2s
  movi v0.16b, 1
  movi v1.16b, 2

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 3.0037

retire (01)cycle (02)03080b18191e1f3f4e51inst issue (52)~issue int (53)~issue fp/simd (54)~dispatch int (56)~dispatch fp/simd (57)huge thing int (59)huge thing fp/simd (5b)60696d6edispatch stall (70)72scheduler rewind (75)scheduler stall (76)~dispatch op (78)~map op int (7c)~map op fp/simd (7e)~map lookup int (7f)~map lookup fp/simd (81)8283pipeline redirect (84)85inst all (8c)inst branch (8d)inst branch taken (90)inst b.cc (94)inst integer (97)inst neon or fp (9a)9fa0a1a6a8acc2cfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10204300372250000006129548251010010010000100100005004277313130018300373003728265032874510100200100002002000030037300371110201100991001001000010000046071021622296340100001003003830038300383003830038
10204300372250000006129548251010010010000100100005004277313130018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071021622296340100001003003830038300383003830038
1020430037225000000612954825101001001000010010000500427731313001830037300372826503287451010020010000200200003003730037111020110099100100100001000420192582869376232988427100001003037030180303243036830371
1020430373226110780461652082949416410185138100481391104371942854550303423046930498283010482890911489220114972322298430463304645110201100991001001000010072239266080912464323001036100001003046730511302303051330517
102043046522800410132379258882946522510203138100801511149074242868121303783046730515282980522892211642220100002002000030037300371110201100991001001000010000000071021622296340100001003003830038300383003830038
10204300372250000006129548251010010010000100100005004278303130018300373003728265032874510100200100002002000030037300371110201100991001001000010000029071021622296340100001003003830038300383003830038
10204300372250000006129548251010010010000100100005004277313130018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071021622296340100001003003830038300383003830038
10204300372250000006129548251010010010000100100005004277313130018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071021622296340100001003003830038300383003830038
10204300372250000006129548251010010010000100100005004277313130018300373003728265032874510100200100002002000030037300371110201100991001001000010000020071021622296340100001003003830038300383003830038
10204300372250000006129548251010010010000100100005004277313130018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071021622296340100001003003830038300383003830038

1000 unrolls and 10 iterations

Result (median cycles for code): 3.0037

retire (01)cycle (02)03080b18191e1f3a3f4e51inst issue (52)~issue int (53)~issue fp/simd (54)~dispatch int (56)~dispatch fp/simd (57)huge thing int (59)huge thing fp/simd (5b)60696d6edispatch stall (70)scheduler rewind (75)scheduler stall (76)~dispatch op (78)~map op int (7c)~map op fp/simd (7e)~map lookup int (7f)~map lookup fp/simd (81)8283pipeline redirect (84)85inst all (8c)inst branch (8d)inst branch taken (90)inst b.cc (94)inst integer (97)inst neon or fp (9a)9fa0a1a6a8a9acc2cfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
100243003722400001800441295482510010101000010100005042773131300183003730037282873287671001020100002020000300373003711100211091010100001004227222626283331056429968610000103046130464304663049930461
10024302292280110911888800649529467217100931910080141149055428952613041430310304992832024289371136124116192423274300373003711100211091010100001000300607705805529980410000103046430323303223041530451
1002430508228001091200792158762945821910051141008012113416542895261303423046130507283244528948115023011649262324630414304601111002110910101000010200102791506402162229630210000103046530464304073041730370
1002430365229008911977040653729458241100921910075141149082429200813037830511305102832447289741150830114702623286305563027611110021109101010000100021903086308492162229630010000103008530038300383003830038
10024300372251089105670406090294671971009213100801311490764284098130378305583046328324512895311502241163424220023046030555111100211091010100001000000906402162229630010000103003830038300823003830085
10024300372330000000104295482510010101000010100005042777971300183003730037282873287671001020100002020000300373003711100211091010100001000010606402162229630010000103003830038300383003830038
100243003722500000006129548251001010100001010000504277313130018300373003728287328767100102010000202000030037300371110021109101010000100005501206402162229630010000103003830038300383003830038
10024300372250000000612954825100101010000101000050427731313001830037300372828732876710010201000020200003003730037111002110910101000010000320306401162229630010000103003830038300383003830038
100243003722500000006129548251001010100001010000504277313130018300373003728287328767100102010000202000030037300371110021109101010000100002701206402162229630010000103003830038300383003830038
1002430037225000000061295482510010101000010100005042773131300183003730037282873287671001020100002020000300373003711100211091010100001000020306402162229630010000103003830038300383003830038

Test 3: Latency 1->3

Code:

  sqrshl v0.2s, v1.2s, v0.2s
  movi v0.16b, 1
  movi v1.16b, 2

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 3.0037

retire (01)cycle (02)03090b1e3f4e51inst issue (52)~issue int (53)~issue fp/simd (54)~dispatch int (56)~dispatch fp/simd (57)huge thing int (59)huge thing fp/simd (5b)60696d6edispatch stall (70)scheduler rewind (75)scheduler stall (76)~dispatch op (78)~map op int (7c)~map op fp/simd (7e)~map lookup int (7f)~map lookup fp/simd (81)8283pipeline redirect (84)85inst all (8c)inst branch (8d)inst branch taken (90)inst b.cc (94)inst integer (97)inst neon or fp (9a)9fa6a8acbranch mispredict (cb)cfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
102043003722500078129548251010010010000100100005004277313130018300373003728265328745101002001000020020000300373003711102011009910010010000100020071011611296340100001003003830038300383003830038
10204300372250006129548251010010010000100100005004277313030018300373003728265328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
102043003722500082295482510100100100001001000050042773131300183003730037282653287451010020010000200200003003730037111020110099100100100001000620071011611296340100001003003830038300383003830038
10204300372250006129548251010010010000100100005004277313030018300373003728265328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
10204300372250006129548251010010010000100100005004277313130018300373003728265328745101002001000020020000300373003711102011009910010010000100000171011611296340100001003003830038300383003830038
102043003722400061295482510100100100001001000050042773131300183003730037282653287451010020010000200200003003730037111020110099100100100001000110071011611296340100001003003830038300383003830038
10204300372250096129548251010010010000100100005004277313030018300373003728265328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
1020430037225000384529548251010010010000100100005004277313130018300373003728265328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
10204300372240006129548251010010010000100100005004277313130018300373003728265328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
10204300372250006129548251010010010000100100005004277313030018300373003728265328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038

1000 unrolls and 10 iterations

Result (median cycles for code): 3.0037

retire (01)cycle (02)030918191e1f3f4e51inst issue (52)~issue int (53)~issue fp/simd (54)~dispatch int (56)~dispatch fp/simd (57)huge thing int (59)huge thing fp/simd (5b)696d6edispatch stall (70)scheduler rewind (75)scheduler stall (76)~dispatch op (78)~map op int (7c)~map op fp/simd (7e)~map lookup int (7f)~map lookup fp/simd (81)8283pipeline redirect (84)85inst all (8c)inst branch (8d)inst branch taken (90)inst b.cc (94)inst integer (97)inst neon or fp (9a)9fa0a6a8a9acc2cfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
100243003722500000612954825100101010000101000050427731330018300373003728287328767100102010000202000030037300371110021109101010000100000006402162229630010000103003830038300383003830038
100243003722500000612954825100101010000101000050427731330018300373003728287328767100102010000202000030037300371110021109101010000100010006402162229630010000103003830038300383003830038
100243003722500000612954825100101010000101000050427731330018300373003728287328767100102010000202000030037300371110021109101010000100000006402162229630010000103003830038300383003830038
100243003722400000612954825100101010000101000050427731330018300373003728287328767100102010000202000030037300371110021109101010000100000006402162229630010000103003830038300383003830038
100243003722400000612954825100101010000101000050427731330018300373003728287328767100102010000202000030037300371110021109101010000100000006402162229630010000103003830038300383003830038
100243003722500000612954825100101010000101000050427731330018300373003728287328767100102010000202000030037300371110021109101010000100000006402162229630010000103003830038300383003830038
100243003722500000612954825100101010000101000050427731330018300373003728287328767100102010000202000030037300371110021109101010000100010006402162229630010000103003830038300383003830038
100243003722500000612954825100101010000101000050427731330018300373003728287328767100102010000202000030037300371110021109101010000100000006402162229630010000103003830038300383003830038
100243003722500000612954825100101010000121000050427731330018300373003728287328767100102010000202000030037300371110021109101010000100000006402162229630010000103003830038300383003830038
100243003722500000612954825100101010000101000050427731330018300373003728287328767100102010000202000030037300371110021109101010000100000006402162229630010000103003830038300383003830038

Test 4: throughput

Count: 8

Code:

  sqrshl v0.2s, v8.2s, v9.2s
  sqrshl v1.2s, v8.2s, v9.2s
  sqrshl v2.2s, v8.2s, v9.2s
  sqrshl v3.2s, v8.2s, v9.2s
  sqrshl v4.2s, v8.2s, v9.2s
  sqrshl v5.2s, v8.2s, v9.2s
  sqrshl v6.2s, v8.2s, v9.2s
  sqrshl v7.2s, v8.2s, v9.2s
  movi v8.16b, 9
  movi v9.16b, 10

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code divided by count): 0.2505

retire (01)cycle (02)033f51inst issue (52)~issue int (53)~issue fp/simd (54)~dispatch int (56)~dispatch fp/simd (57)huge thing int (59)huge thing fp/simd (5b)60696d6edispatch stall (70)scheduler rewind (75)scheduler stall (76)~dispatch op (78)~map op int (7c)~map op fp/simd (7e)~map lookup int (7f)~map lookup fp/simd (81)8283pipeline redirect (84)85inst all (8c)inst branch (8d)inst branch taken (90)inst b.cc (94)inst integer (97)inst neon or fp (9a)9fa6a8accfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
802042005815041258010010080000100800005006400001200202003920039997339997801002008000020016000020039200391180201100991001008000010000051102161120036800001002004020040200402004020040
8020420039150412580100100800001008000050064000002002020039200399973399978010020080000200160000200392003911802011009910010080000100002751101161120036800001002004020040200402004020040
802042003915041258010010080000100800005006400000200202003920039997339997801002008000020016000020039200391180201100991001008000010000051101161120036800001002004020040200402004020040
802042003915041258010010080000100800005006400000200202003920039997339997801002008000020016000020039200391180201100991001008000010000051101161120036800001002004020040200402004020040
8020420039150412580100100800001008000050064000002002020039200399973399978010020080000200160000200392003911802011009910010080000100007251101161120036800001002004020040200402004020040
802042003915041258010010080000100800005006400000200202003920039997339997801002008000020016000020039200391180201100991001008000010000051101161120036800001002004020040200402004020040
802042003915041258010010080000100800005006400000200202003920039997339997801002008000020016000020039200391180201100991001008000010000051101161120036800001002004020040200402004020040
8020420039150136258010010080000100800005006400000200202003920039997339997801002008000020016000020039200391180201100991001008000010000051101161120036800001002004020040200402004020040
802042003915041258010010080000100800005006400000200202003920039997339997801002008000020016000020039200391180201100991001008000010001051101161120036800001002004020040200402004020040
8020420039150412580100100800001008010550064000002002020039200399973399978010020080000200160000200392003911802011009910010080000100006651101161120036800001002004020040200402004020040

1000 unrolls and 10 iterations

Result (median cycles for code divided by count): 0.2505

retire (01)cycle (02)030a181e3f51inst issue (52)~issue int (53)~issue fp/simd (54)~dispatch int (56)~dispatch fp/simd (57)huge thing int (59)huge thing fp/simd (5b)5f60696b6d6edispatch stall (70)scheduler rewind (75)scheduler stall (76)~dispatch op (78)~map op int (7c)~map op fp/simd (7e)~map lookup int (7f)~map lookup fp/simd (81)8283pipeline redirect (84)85inst all (8c)inst branch (8d)inst branch taken (90)inst b.cc (94)inst integer (97)inst neon or fp (9a)9fa8a9acbranch mispredict (cb)cfd2icache miss (d3)d5d6daddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
80024200481500004025800101080000108000050640000002002002003920039999631001980010208000020160000200392003911800211091010800001010005020004160352003680000102004020040200402004020040
80024200391500006125800101080000108000050640000002002002003920039999631001980010208000020160000200392003911800211091010800001000005020005160532003680000102004020040200402004020040
80024200391500004025800101080000108000050640000002002002003920039999631001980010208000020160000200392003911800211091010800001010005020003160352003680000102004020040200402004020040
80024200391500004025800101080000108000050640000002002002003920039999631001980010208000020160000200392003911800211091010800001000163805020003160352003680000102004020040200402004020040
80024200391500004025800101080000108000050640000002002002003920039999631001980010208000020160000200392003911800211091010800001000315020004160452003680000102004020040200402004020040
80024200391500006125800101080000108000050640000002002002003920039999631001980010208000020160000200392003911800211091010800001010005020005160532003680000102004020040200402004020040
80024200391500004025800101080000108000050640000002002002003920039999631001980010208000020160000200392003911800211091010800001020005020005160352003680000102004020040200402004020040
80024200391500004025800101080000108000050640000002002002003920039999631001980010208000020160000200392003911800211091010800001010005020003160532003680000102004020040200402004020040
80024200391500004025800101080000108000050640000002002002003920039999631001980010208000020160000200392003911800211091010800001000005020003160532003680000102004020040200402004020040
80024200391500004025800101080000108000050640000002002002003920039999631001980010208000020160000200392003911800211091010800001030005020005160532003680000102004020040200402004020040