Apple M1 Microarchitecture Research by Dougall Johnson

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

UQXTN2 (2D)

Test 1: uops

Code:

  uqxtn2 v0.4s, v1.2d
  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)03081e3f4e51inst 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
1004303723006125482510001000100039831313018303730372415328951000100020003037303711100110000973116112630100030383038303830383038
1004303723006125482510001000100039831313018303730372415328951000100020003037303711100110001073116112630100030383038303830383038
10043037230010325482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303722006125482510001000100039831303018303730372415328951000100020003037303711100110000373116112630100030383038303830383038
1004303723006125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303723006125482510001000100039831303018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303722006125482510001000100039831303018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303723006125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303723006125482510001000100039831303018303730372415328951000100020003037303711100110000073116112630100030383038303830383038
1004303722006125482510001000100039831313018303730372415328951000100020003037303711100110000073116112630100030383038303830383038

Test 2: Latency 1->1

Code:

  uqxtn2 v0.4s, v1.2d
  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)031e3a3f4e51inst 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)9fa0a8accdcfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
1020430037225150124295482510100100100001001000050042773131300180300373003728265328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
10204300372250061295482510100100100001001000050042773130300180300373003728265328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
102043003722500612954825101001001000012510000626427731303001803003730037282653287451010020010000200200003003730037111020110099100100100001000000710116112963425100001003003830038300383003830038
10204300372250061295482510100100100001001000050042773130300180300373003728290328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
102043003722590612954825101001001000010010000500427731303001803003730037282653287451010020010000200200003003730037111020110099100100100001000000710116112963425100001003003830038300383003830038
1020430037225906129548251010010010000100100005004277313130018030037300372826532874510100200100002002000030037300371110201100991001001000010000141071011611296340100001003003830038300383003830038
10204300372250061295482510100100100001001000050042773130300180300373003728265328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
10204300372250061295482510100100100001001000050042773131300180300373003728265328744101252001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
10204300372250061295482510100100100001001000050042773130300180300373003728265328745101002001000020020000300373003711102011009910010010000100000071011611296340100001003003830038300383003830038
10204300372250061295482510100100100001001000050042773130300180300373003728265328745101002001000020020000300373003711102011009910010010000100420071011611296340100001003003830038300383003830038

1000 unrolls and 10 iterations

Result (median cycles for code): 3.0037

retire (01)cycle (02)031e3f4e51inst 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)9fa0a1a8c5cdcfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
100243003722506129548251001010100001010000504278675030018300373003728287032876710010201000020200003003730037111002110910101000010000006403163329630010000103003830038300383003830038
100243003722506129548251001010100001010000504277313030018300373003728287032876710010201000020200003003730037111002110910101000010000006403163329630010000103003830038300383003830038
100243003722506129548251001010100001010000504277313030018300373003728287732876710010201000020200003003730037111002110910101000010000006403163329630010000103003830038300383003830038
100243003722506129548251001010100001010000504277313030018300373003728287032876710010201000020200003003730037111002110910101000010000006403163329630210000103003830038300383003830038
100243003722506129548251001010100001010000504277313030018300373003728287032876710010201000020200003003730037111002110910101000010000006403163329630010000103003830038300383003830038
100243003722506129548251001010100001010000504277313130018300373003728287032876710010201000020200003003730037111002110910101000010000006403163329630010000103003830038300383003830038
100243003722506129548251001010100001010000504277313030018300373003728287032876710010201000020200003003730037111002110910101000010000006403163329630010000103003830038300383003830038
100243003722506129548251001010100001010000504277313030018300373003728287032876710010201000020200003003730037111002110910101000010000006403163329630010000103003830038300383003830038
100243003722506129548251001010100001010000504277313030018300373003728287032876710010201000020200003003730037111002110910101000010000006403163329630010000103003830038300383003830038
100243003722406129548251001010100001010000504277313030018300373003728287032876710010201000020200003003730037111002110910101000010000006403163329630010000103003830038300383003830038

Test 3: Latency 1->2

Code:

  uqxtn2 v0.4s, v0.2d
  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)03040818191e1f3a3f4e51inst 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)9fa0a1a8acc5branch mispredict (cb)cdcfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10204300372251000000106295472510100100100001001000050042771601300183003730037282710628741101002001000820020016300373003711102011009910010010000100000011171801600296450100001003003830038300383003830038
1020430037225000000034629547251010010010000100100005334277160130234303213032328310035288501102422211001218223463036730373811020110099100100100001002011939011187217201299114100001003018030374303733037430371
1020430326227005592461604582948415410132118100561151090060942852341300543037030324282970292886711072222111672202232030322303628110201100991001001000010020222350111831085002990213100001003036930361303723037230360
102043036922701772676160134329493174101901231005611211050654428662413027030359303702830303928865110162261117722222014300373037281102011009910010010000100000011171701600296460100001003003830038300383003830038
1020430037225000000061295472510100100100001001000050042771601300183003730037282710728740101002001000820020016300373003711102011009910010010000100000011171801600296450100001003003830038300383003830038
10204300372250000000151295472510100100100001001000050042771601300183003730037282710628741101002001000820020016300373003711102011009910010010000100000011171801600296460100001003003830038300383003830038
1020430037225000000061295472510100100100001001000050042771601300183003730037282710628740101002001000820020016300373003711102011009910010010000100000011171701600296450100001003003830038300383003830038
1020430037225000000061295472510100100100001001000050042771601300183003730037282710628740101002001000820020016300373003711102011009910010010000100000011171701600296450100001003003830038300383003830038
1020430037225000000061295472510100100100001001000050042771601300183003730037282710628741101002001000820020016300373003711102011009910010010000100000011171801600296450100001003003830038300383003830038
1020430037225000000061295472510100100100001001000050042771601300183003730037282710728741101002001000820020016300373003711102011009910010010000100000011171701600296460100001003003830038300383003830038

1000 unrolls and 10 iterations

Result (median cycles for code): 3.0037

retire (01)cycle (02)03181e1f3f4e51inst 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)9fa0a8accfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
1002430037225000612954725100101010000101000050427716030018300373003728286032876710010201000020200003003730037111002110910101000010010640216222962910000103003830038300383003830038
1002430037224000612954725100101010000101000050427716030018300813003728286032876710010201000020200003003730037111002110910101000010003640216222962910000103003830038300383003830038
1002430037225000642954725100101010000101000050427716030018300373003728286032876710010201000020200003003730037111002110910101000010000640216222962910000103003830038300383003830038
100243003722542426414062954799100101010000101000050427716030018300373003728286032876710010201000020200003003730037111002110910101000010000640216222962910000103003830038300383003830038
1002430037225000612954725100101010000101000050427716030018300373003728286032876710010201000020200003003730037111002110910101000010000640216222962910000103003830038300383003830038
1002430037224060612954725100101010000101000050427716030018300373003728286032876710010201000020200003003730037111002110910101000010000640216222962910000103003830038300383003830038
1002430037225000612954725100101010000101000050427716030018300373003728286032876710010201000020200003003730037111002110910101000010000640216222962910000103003830038300383003830038
1002430037225000612954725100101010000101000050427716030018300373003728286032876710010201000020200003003730037111002110910101000010000640216222962910000103003830038300383003830038
1002430037224000612954725100101010000101000050427716030018300373003728286032876710010201000020200003003730037111002110910101000010000640216222962910000103003830038300383003830038
100243003722500064232954725100101010000101000050427716030018300373003728286032876710010201000020200003003730037111002110910101000010000640216222962910000103003830038300383003830038

Test 4: throughput

Count: 8

Code:

  movi v0.16b, 0
  uqxtn2 v0.4s, v8.2d
  movi v1.16b, 0
  uqxtn2 v1.4s, v8.2d
  movi v2.16b, 0
  uqxtn2 v2.4s, v8.2d
  movi v3.16b, 0
  uqxtn2 v3.4s, v8.2d
  movi v4.16b, 0
  uqxtn2 v4.4s, v8.2d
  movi v5.16b, 0
  uqxtn2 v5.4s, v8.2d
  movi v6.16b, 0
  uqxtn2 v6.4s, v8.2d
  movi v7.16b, 0
  uqxtn2 v7.4s, v8.2d
  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)0318191e1f3f51inst 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)9fa1a8a9acc2c5branch mispredict (cb)cdcfd5d6e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
160204200911500000712580116100800161008002850064019602004520065200656128012820080028200160056200652006511160201100991001001600001000000011110119216200621600001002006620066200662006620066
160204200651510000942580116100800161008002850064019602004520065200656128012820080028200160056200652006511160201100991001001600001000000011110119016200621600001002006620066200662006620066
16020420065150000037525801161008001610080028500640196020045200652006561280128200800282001600562006520065111602011009910010016000010004403011110119016200621600001002006620066200662006620066
160204200651500000502580116100800161008002850064019612004520065200656128012820080028200160056200652006511160201100991001001600001000000011110119016200621600001002006620066200662006620066
160204200651500021605992580116100800161008002850064019602004520065200656128012820080028200160056200652006511160201100991001001600001000000011110119016200621600001002006620066200662006620066
160204200651510000292580116100800161008002850064019602004520065200656128012820080028200160056200652006511160201100991001001600001000000011110119016200621600001002006620066200662006620066
160204200651510000522580116100800161008002850064019602004520065200656128012820080028200160056200652006511160201100991001001600001000000011110119016200621600001002006620066200662006620066
1602042006515000001152580116100800161008002850064019602004520065200656128012820080028200160056200652006511160201100991001001600001000200011110119016200621600001002006620066200662006620066
1602042006515000004152580116100800161008002850064019602004520065200656128012820080028200160056200652006511160201100991001001600001000203011110119016200621600001002006620066200662006620066
1602042006515000001132580116100800161008002850064019602004520065200656128012820080028200160056200652006511160201100991001001600001000000011110119016200621600001002006620066200662006620066

1000 unrolls and 10 iterations

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

retire (01)cycle (02)0304070a18191e1f3a3f51inst 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)9fa0a1a6a8a9accfd0d2icache miss (d3)d5d6d9dadbddinst fetch restart (de)e0eaec? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
16002420067150122000003462580010108000010800005064000001520033200522005232280010208000020160000200522005311160021109101016000010000000100598413720211343420043150160000102004720047200472004720047
16002420046150022000006262580010108000010800005064000011520027200462004632280010208000020160000200462004611160021109101016000010000000100598613520211362220043150160000102004720047200472004720047
16002420046150022000002552580010108000010800005064000011520027200462004632280010208000020160000200462004611160021109101016000010000000100588613420211273620043150160000102004720047200472004720047
16002420046150022000002462580010108000010800005064000011520027200462004632280010208000020160000200462004611160021109101016000010000000100428613620211353020043150160000102004720047200472004720047
16002420046150022000002202580010108000010800005064000011520027200462004632280010208000020160000200462004611160021109101016000010000000100558612220211332120043150160000102004720047200472004720047
16002420046150022000004322580010108000010800005064000001520033200522004632280010208000020160000200522005211160021109101016000010000000100588723520321213520043150160000102005320053200472005320053
16002420046150022000002192580010108000010800005064000001020027200462004632280010208000020160000200462004611160021109101016000010000000100578613520211193520043150160000102005320053200532005320053
16002420052150021000002282580010108000010800005064000011520027200462005232280010208000020160000200462004611160021109101016000010200100100588613620311363220049150160000102005320053200532005320053
1600242005215001300001999258001010800001080000506400000152003320052200523228001020800002016000020052200521116002110910101600001000000961005711713626211353520043150160000102004720053200472004720047
160024200461500120000020225800101080000108000050640000115200272004620046322800102080000201600002004620046111600211091010160000100000001006211722320322353620049150160000102005320053200532005320053

Test 5: throughput

Count: 16

Code:

  uqxtn2 v0.4s, v16.2d
  uqxtn2 v1.4s, v16.2d
  uqxtn2 v2.4s, v16.2d
  uqxtn2 v3.4s, v16.2d
  uqxtn2 v4.4s, v16.2d
  uqxtn2 v5.4s, v16.2d
  uqxtn2 v6.4s, v16.2d
  uqxtn2 v7.4s, v16.2d
  uqxtn2 v8.4s, v16.2d
  uqxtn2 v9.4s, v16.2d
  uqxtn2 v10.4s, v16.2d
  uqxtn2 v11.4s, v16.2d
  uqxtn2 v12.4s, v16.2d
  uqxtn2 v13.4s, v16.2d
  uqxtn2 v14.4s, v16.2d
  uqxtn2 v15.4s, v16.2d
  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)030a18191e1f3f51inst 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)9fa1a6a8a9acc2c5branch mispredict (cb)cdcfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
160204400593000009030251601081001600081001600205001280132400204003940039199776199901601202001600322003200644003940039111602011009910010016000010000000011110118216114003601600001004004040040400404004040040
160204400393000000030251601081001600081001600205001280132400204003940039199776199901601202001600322003200644003940039111602011009910010016000010000000011110118016014003601600001004004040040400404004040040
160204400392990000030251601081001600081001600205001280132400204003940039199776199901601202001600322003200644003940039111602011009910010016000010002300011110118016104003601600001004004040040400404004040040
160204400392990000030251601081001600081001600205001280132400204003940039199776199901601202001600322003200644003940039111602011009910010016000010040000011110118116004003601600001004004040040400404004040040
160204400393000000030251601081001600081001600205001280132400204003940039199776199901601202001600322003200644003940039111602011009910010016000010000000011110118216024003601600001004004040040400404004040040
160204400393000000030251601081001600081001600205001280132400204003940039199776199901601202001600322003200644003940039211602011009910010016000010000000011110118016114003601600001004004040040400404004040040
160204400393000000030251601081001600081001600205001280132400204003940039199776199901601202001600322003200644003940039111602011009910010016000010000000011110118016104003601600001004004040040400404004040040
160204400393000000030251605011001600081001600205001280132400204003940039199776199901601202001600322003200644003940039111602011009910010016000010000000011110118116014003601600001004004040040400404004040040
160204400392990000030251601081001600081001600205001280132400204003940039199776199901601202001600322003200644003940039111602011009910010016000010000000011110118016004003601600001004004040040400404004040040
160204400392990000051251601081001600081001600205001280132400204003940039199776199901601202001600322003200644003940039111602011009910010016000010000003011110118216024003601600001004004040040400404004040040

1000 unrolls and 10 iterations

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

retire (01)cycle (02)030809191e1f3a3f51inst 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)5f6061696d6edispatch 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)9fa1a6a8a9acc2cdcfd0d2icache miss (d3)d5d6d9dadbddinst fetch restart (de)e0eaeb? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
160024400513000000028865251600101016000010160000501280000115400204003940039199963200191600102016000020320000400394003911160021109101016000010000030010023831102821121940036206160000104004040040400404004040040
1600244003929910000156802516001010160000101600005012800001154002040039400391999632001916001020160000203200004003940039111600211091010160000100000300100238212016211102140036206160000104004040040400404004040040
16002440039300000001274725160010101600001016000050128000011540020400394003919996320019160010201600002032000040039400391116002110910101600001000000001002282171621172040036206160000104004040040400404004040040
16002440039300100001094125160010101600001016000050128000011540020400394003919996320019160010201600002032000040039400391116002110910101600001000000001002383181621120940036206160000104004040040400404004040040
16002440039299000001265325160010101600001016000050128000011540020400394003919996320019160010201600002032000040039400391116002110910101600001000620001002382192721172040036206160000104004040040400404004040040
1600244003930000000206052516001010160000101600005012800001154002040039400391999632001916001020160000203200004003940039111600211091010160000100000000100238212016211162040036206160000104004040040400404004040040
16002440039300000692648241251600101016000010160000501280000115400204003940039199963200191600102016000020320000400394003911160021109101016000010000000010023821201621120740036206160000104004040040400404004040040
16002440039300000009203251600101016000010160000501280000115400204003940039199963200191600102016000020320000400394003911160021109101016000010000000010023821171621120740036206160000104004040040400404004040040
160024400392990000011155251600101016000010160000501280000115400204003940039199963200191600102016000020320000400394003911160021109101016000010022000010023821201621120840036206160000104004040040400404004040040
160024400393001000012516251600101016000010160000501280000115400204003940039199963200191600102016000020320000400394003911160021109101016000010000000010023821201621192040036206160000104004040040400404004040040