Apple M1 Microarchitecture Research by Dougall Johnson

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

AESE

Test 1: uops

Code:

  aese v0.16b, v1.16b
  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)03181e3f4e51inst 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)a8accfd5d6ddinst fetch restart (de)e0? fp/simd (ee)f5f6f7f8fd
1004303723006125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
10043037220126125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303723006125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303722006125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
100430372301216125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303723006125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303723006125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303723006125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303723096125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303723006125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038

Test 2: Latency 1->1

Code:

  aese v0.16b, v1.16b
  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)03093f4e51inst 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)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)9fa0a1a8a9accdcfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10204300372250612954844101001001000010010000500427731330018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071011611296340100001003003830038300383003830038
10204300372250612954825101001001000010010000500427731330018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071011611296340100001003003830038300383003830038
10204300372250612954825101001001000010010000500427731330018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071011611296340100001003003830038300383003830038
10204300372250612954825101001001000010010000500427731330018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071011611296340100001003003830038300383003830038
102043003722501242954825101001001000010010000500427731330018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071011611296340100001003003830038300383003830038
10204300372250612954825101001001000010010000500427731330018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071011611296340100001003003830038300383003830038
10204300372250612954825101001001000010010000500427731330018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071011611296340100001003003830038300853003830038
10204300372250612954825101001001000010010000500427731330018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071011611296340100001003003830038300383003830038
10204300372250612954825101001001000010010000500427731330018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071011611296340100001003003830038300383003830038
10204300372250612954825101001001000010010000500427731330018300373003728265032874510100200100002002000030037300371110201100991001001000010000000071011611296340100001003003830038300383003830038

1000 unrolls and 10 iterations

Result (median cycles for code): 3.0037

retire (01)cycle (02)03040818191e3f4e51inst 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)9fa0a1a9accfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10024300372250000061295482510010101000010100005042773131300183003730037282873287671001020100002020000300373003711100211091010100001000006402162229630010000103003830038300383003830038
10024300372250000061295482510010101000010100005042773131300543008530037282873287671001020100002020354300373003711100211091010100001000027656613243329702210000103013230084301333013330085
100243003722500001321021295482510010101000010100005042773130300183013330227282873287671001020100002020000300373003711100211091010100001030256006402162229630010000103003830038300383003830038
10024300372250000061295482510010101000010100005042773130300183003730037282873287671001020100002020000300373003711100211091010100001000006402162229630010000103003830038300383003830038
10024300372250010061295482510010101000010100005042773130300183003730037282873287671001020100002020000300373003711100211091010100001000006402242229630010000103003830038300383003830038
10024300372250000061295482510010101000010100005042773131300183003730037282873287671001020100002020000300373003711100211091010100001000006402162229630010000103003830038300383003830038
10024300372250000061295482510010101000010100005042773131300183003730037282873287671001020100002020000300373003711100211091010100001000006402162229630010000103003830038300383003830038
10024300372250000061295482510010101000010100005042773130300183003730037282873287671001020100002020000300373003711100211091010100001000006402162229630010000103008530038300383003830038
100243003724147000061295482510010101000010100005042773131300183003730037282873287671001020100002020000300373003711100211091010100001000006402162229630010000103003830038300383003830038
10024300372250000082295482510010101000810101496542773130300183003730037282873287671001020100002220982300843003731100211091010100001000006402162229630010000103003830038300383003830038

Test 3: Latency 1->2

Code:

  aese v0.16b, v0.16b
  movi v0.16b, 1

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 3.0037

retire (01)cycle (02)0318191e1f3a3f4e51inst 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)60696b6d6edispatch 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)9fa0a1a6a7a8a9acc5branch mispredict (cb)cdcfd5d6d8ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
102043003722500000612954725101001001000010010000500427716013001803003730037282717287401010020010008200203443003730037111020110099100100100001000020002813111718016000296450100001003003830038300383003830038
10204300372250000046642954725101001001000010210000500427716013001803003730037282717287401010020010008200200163003730037111020110099100100100001000000000111742016000296462100001003008530038300383008730085
10204300372261012006129547251010010010000100100005004277160030018030037300372829272874010100200100082002001630037300371110201100991001001000010004010231117640247800296463100001003003830086300853003830229
102043008422500900612952025101121001000010410300522427851213009003013230037282716287401010020010008200200163003730085111020110099100100100001002000400111764016001297200100001003008630038300383003830038
10204300372260000061295472510100100100001001000050042771601300180300373003728271162874110100200100082002001630037300373110201100991001001000010000002001117620165500297270100001003003830038300383003830038
10204300372250221006129538251010010810000100100005444279864130054030037300372827113287411010020010008204200163003730037111020110099100100100001000000100111717016022297010100001003013330038301333003830133
10204301332250039011392953843101001001000010010000500427716013001803003730037282711728740104082001000820820016300373003711102011009910010010000100003004178111722224000296450100001003003830038300383003830038
1020430037225001590032832952962101001041000810610000516427986413009003008530037282716287411040720010008200200163003730037111020110099100100100001000000000111717032010296460100001003003830038300383003830038
1020430037225020880612954725101001001000010010000500427851203001803003730037282717287411010020010008200200163003730037111020110099100100100001000030200111718016000296450100001003003830038300383003830038
1020430037225000880822954725101001001000010010000500427716003001803003730037282926287411010020010008200200163003730037111020110099100100100001000000100111718396000296453100001003003830086300843013330085

1000 unrolls and 10 iterations

Result (median cycles for code): 3.0037

retire (01)cycle (02)0309181e3f4e51inst 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)9fa6a8acc2cfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10024300372250006129547251001010100001010000504277160130018300373003728286328767100102010000202000030037300371110021109101010000100000640216222962910000103003830038300383003830038
10024300372250006129547251001010100001010000504277160030018300373003728286328767100102010000202000030037300371110021109101010000100000640216222962910000103003830038300383003830038
10024300372250006129547251001010100001010000504277160030018300373003728286328767100102010000202000030037300371110021109101010000100000640216222962910000103003830038300383003830038
10024300372250006129547251001010100001010000504277160130018300373003728286328767100102010000202000030037300371110021109101010000100000640216222962910000103003830038300383003830038
10024300372250006129547251001010100001010000504277160130018300373003728286328767100102010000202000030037300371110021109101010000100000640216222962910000103003830038300383003830038
10024300372250096129547251001010100001010000504277160130018300373003728286328767100102010000202000030037300371110021109101010000100000640216222962910000103003830038300383003830038
100243003722500034629547251001010100001010000504277160030018300373003728286328767100102010000202000030037300371110021109101010000100000640216222962910000103003830038300383003830038
100243003722500072629547251001010100001010000504277160130018300373003728286328767100102010000202000030037300371110021109101010000100000640216222962910000103003830038300383003830038
10024300372240106129547251001010100001010000504277160030018300373003728286328767100102010000202000030037300371110021109101010000100000640216222962910000103003830038300383003830038
10024300372250006129547251001010100001010000504277160130018300373003728286328767100102010000202000030037300371110021109101010000100000640216222962910000103003830038300383003830038

Test 4: throughput

Count: 8

Code:

  movi v0.16b, 0
  aese v0.16b, v8.16b
  movi v1.16b, 0
  aese v1.16b, v8.16b
  movi v2.16b, 0
  aese v2.16b, v8.16b
  movi v3.16b, 0
  aese v3.16b, v8.16b
  movi v4.16b, 0
  aese v4.16b, v8.16b
  movi v5.16b, 0
  aese v5.16b, v8.16b
  movi v6.16b, 0
  aese v6.16b, v8.16b
  movi v7.16b, 0
  aese v7.16b, v8.16b
  movi v8.16b, 9

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code divided by count): 0.2508

retire (01)cycle (02)03070a1e3f51inst 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)60696d6escheduler 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)9fa8a9acc5branch mispredict (cb)cdcfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
160204200881511101362580116100800161008002850064019612004520065200656128012820080028200160056200652006511160201100991001001600001000001111012131633200621600001002006620066200662006620066
16020420065150110292580116100800161008002850064019612004520065200656128012820080028200160056200652006511160201100991001001600001000031111012131633200621600001002006620066200662006620066
16020420065151110292580116100800161008002850064019612004520065200656128012820080028200160056200652006511160201100991001001600001000001111012131632200621600001002006620066200662006620066
16020420065151110942580116100800161008002850064019612004520065200656128012820080028200160056200652006511160201100991001001600001002801381111012121633200621600001002006620066200662006620066
16020420065151110502580116100800161008002850064019612004520065200656128012820080028200160056200652006511160201100991001001600001003601471111012131633200621600001002006620066200662006620066
160204200651511101572580116100800161008002850064019612004520065200656128012820080028200160056200652006511160201100991001001600001000001111012131633200621600001002006620066200662006620066
160204200651501107152580116100800161008002850064019612004520065200656128012820080028200160056200652006511160201100991001001600001000001111012131633200621600001002006620066200662006620066
16020420156150110292580116100800161008002850064019612004520065200656128012820080028200160056200652006511160201100991001001600001000001111012131633200621600001002006620066200662006620066
160204200651501136477258011610080016100800285006401961200452006520065612801282008002820016005620065200651116020110099100100160000100390811111012131633200621600001002006620066200662006620066
16020420065151110292580116100800161008002850064019612004520065200656128012820080028200160056200652006511160201100991001001600001000001111012131633200621600001002006620066200662006620066

1000 unrolls and 10 iterations

Result (median cycles for code divided by count): 0.2506

retire (01)cycle (02)03050b1e3a3f51inst 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)5f6061696d6escheduler 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)9fa7a8accfd0d2icache miss (d3)d5d6d9dadbddinst fetch restart (de)e0ea? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
1600242008215000005492580010108000010800005064000011520027200462005032280010208000020160000200502005011160021109101016000010000100261132624412552004315160000102005120057200472005120051
16002420056150100051258001010800001080000506400000152002720050200463228001020800002016000020050200461116002110910101600001000010034832524421932004715160000102005120051200472005120047
16002420046150000037425800101080000108000050640000015200272005020050322800102080000201600002005020050111600211091010160000100001003111325244221132004730160000102005120051200512004720051
160024200501500000952580010108000010800005064000011520031200502005032280010208000020160000200502005011160021109101016000010000100271132724422532004330160000102005120051200512005120051
16002420050151000011132580010108000010800005064000011520027200502004632280010208000020160000200462005011160021109101016000010000100311141324412992004730160000102004720047200472004720047
1600242004615000004932580010108000010800005064000001520031200502005032280010208000020160000200502004611160021109101016000010040126100351142524222352004715160000102005120051200472005120051
160024200501500000452580010108000010800005064000001520027200462005032280010208000020160000200462005011160021109101016000010046171100261132324422932004715160000102004720051200512004720051
16002420050150000051258001010800001080000616400000152003120050200463228001020800002016000020050200461116002110910101600001000010029832520412932004330160000102005120047200512004720051
160024200501500000562580010108000010800005064000001520027200462004632280010208000020160000200502004611160021109101016000010003100291132324421532004715160000102004720047200472004720047
160024200461500000110258001010800001080000506400000152002720050200503228001020800002016000020050200501116002110910101600001004199100321131520211532004730160000102005120051200472005120051

Test 5: throughput

Count: 16

Code:

  aese v0.16b, v16.16b
  aese v1.16b, v16.16b
  aese v2.16b, v16.16b
  aese v3.16b, v16.16b
  aese v4.16b, v16.16b
  aese v5.16b, v16.16b
  aese v6.16b, v16.16b
  aese v7.16b, v16.16b
  aese v8.16b, v16.16b
  aese v9.16b, v16.16b
  aese v10.16b, v16.16b
  aese v11.16b, v16.16b
  aese v12.16b, v16.16b
  aese v13.16b, v16.16b
  aese v14.16b, v16.16b
  aese v15.16b, v16.16b
  movi v16.16b, 17

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code divided by count): 0.2502

retire (01)cycle (02)031e3f51inst 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)60696b6d6edispatch 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)9facc2c5branch mispredict (cb)cdcfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
16020440058300930251601081001600081001600205001280132140020040039400391997761999016012020016003220032006440039400391116020110099100100160000100001111011801600400361600001004004040040400404004040040
160204400392990404251601081001600081001600205001280132140020040150400391997761999016062820016003220032026440039400393116020110099100100160000100301111011801600400361600001004004040040400404004040040
160204400393000505251601081001600081001600205001280132040020040039400391997761999016012020016003220032006440039400391116020110099100100160000100001111011801600400361600001004004040040400404004040040
160204400392991272251601081001600081001600205001280132040020040039400391997761999016012020016003220032006440039400391116020110099100100160000100021111011801600400361600001004004040040400404004040040
16020440039300030251601081241600081001600205001280132140020040039400391997761999016012020016003220032006440039400391116020110099100100160000100001111011801600400361600001004004040040400404004040040
16020440039300030251601081001600081001600205001280132040020040039400391997761999016012020016003220032006440039400391116020110099100100160000100001111011801600400361600001004004040040400404004040040
16020440039300074251601081001600081001600205001280132140020040039400391997761999016012020016003220032006440039400391116020110099100100160000100001111011801600400361600001004004040040400404004040040
16020440039300030251601081001600081001600205001280132040020040039400391997761999016012020016003220032006440039400391116020110099100100160000100001111011801600400361600001004004040040400404004040040
16020440039300030251601081001600081001600205001280132140020040039400391997761999016012020016003220032006440039400391116020110099100100160000100001111011801600400361600001004004040040400404004040040
16020440039300030251601081001600081001600205001280132140020040039400391997761999016012020016003220032006440039400391116020110099100100160000100001111011801600400361600001004004040040400404004040040

1000 unrolls and 10 iterations

Result (median cycles for code divided by count): 0.2502

retire (01)cycle (02)030b18191e3a3f51inst 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)5f60696d6edispatch 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)9fa0a1a6a8a9accfd0d2icache miss (d3)d5d6d9dadbddinst fetch restart (de)e0eaeb? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
16002440050299000024725160010101600001016000050128000011400204003940039199963200191600102016000020320000400394003911160021109101016000010000000100263112016211191940036206160000104004040040400404004040040
16002440039300000024725160010101600001016000050128000011400204003940039199963200191600102016000020320000400394003911160021109101016000010000000100263121716412181640036206160000104004040040400404004040040
160024400393000000247251600101016000010160000501280000114002040039400391999632001916001020160000203200004003940039111600211091010160000100000001002632115162111617400362012160000104004040040400404004040040
1600244003930000063924725160010101600001016000050128000011400204003940039199963200191600102016000020320000400394003911160021109101016000010000000100263111816211201840036206160000104004040040400404004040040
1600244003930000068424725160010101600001016000050128000011400204003940039199963200191600102016000020320000400394003911160021109101016000010000000100263121716211171640036206160000104004040040400404004040040
1600244003930000082224725160010101600001016000050128000011400204003940039199963200191600102016000020320000400394003911160021109101016000010000000100266111916211181740036206160000104004040040400404004040040
16002440039300000024725160010101600001016000050128000011400204003940039199963200191600102016000020320000400394003911160021109101016000010000000100263112116211171940036206160000104004040040400404004040040
16002440039300000024725160010101600001016000050128000011400204003940039199963200191600102016000020320000400394003911160021109101016000010000000100263111916211171940036206160000104004040040400404004040040
1600244003930000027924725160090101600001016000050128000011400784003940039199963200191600102016000020320000400394003911160021109101016000010000000100263111916211171940036406160000104004040040400404004040040
1600244003929900087624725160010101600001016000050128000011400204003940039199963200191600102016000020320000400394003911160021109101016000010010006100263111816211182040036206160000104004040040400404004040040