Apple M1 Microarchitecture Research by Dougall Johnson

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

SLI (vector, 2D)

Test 1: uops

Code:

  sli v0.2d, v1.2d, #3
  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)a8accfd5d6ddinst fetch restart (de)e0? fp/simd (ee)f5f6f7f8fd
100420371506116872510001000100026468012018203720371572318951000100020002037203711100110000073216111787100020382038203820382038
100420371608416872510001000100026468012018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
100420371506116872510001000100026468012018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
100420371506116872510001000100026468012018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
100420371506116872510001000100026468012018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
1004203715246116872510001000100026468012018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
100420371506116872510001000100026468012018203720371572318951000100020002037203711100110000073116211787100020382038203820382038
100420371506116872510001000100026468012018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
100420371506116872510001000100026468012018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
100420371596116872510001000100026468012018203720371572318951000100020002037203711100110007073116111787100020382038203820382038

Test 2: Latency 1->1

Code:

  sli v0.2d, v1.2d, #3
  movi v0.16b, 1
  movi v1.16b, 2

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 2.0037

retire (01)cycle (02)03081e3f4e51inst 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)9fa8accfd5d6ddinst fetch restart (de)e0eb? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
102042003715007261196872510100100100001001000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000071011611197911100001002003820038200382003820038
102042003715001261196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000071011611197910100001002003820038200382003820038
1020420037150050782196872510100100100001001000050028489630200182003720037184223187451010020010000200200002003720037111020110099100100100001000071011611197910100001002003820038200382003820038
102042003715002161196872510100100100001001000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000071011611197910100001002003820038200382003820038
1020420037150046561196872510100100100001001000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000071011611197910100001002003820038200382003820038
10204200371500192441196872510100100100001001000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000071011611197910100001002003820038200382003820038
10204200371500061196872510100100100001001000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000071011611197910100001002003820038200382003820038
10204200371500061196872510100100100001001000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000071011611197910100001002003820038200382003820038
10204200371500061196872510100100100001201000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000071011611197910100001002003820038200382003820038
102042003715002461196872510100100100001001000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000071011611197910100001002003820038200382003820038

1000 unrolls and 10 iterations

Result (median cycles for code): 2.0037

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)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? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
100242003714902706119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100000640216221978510000102003820038200382003820038
100242003715103072619687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100030640216221978510000102003820038200382003820038
10024200371500006119687251001010100001010000502852812020018200372003718444318767100102010000202000020037200371110021109101010000100000640216221978510000102003820038200382003820038
100242003715000061196872510010101000010100005028476801200182003720037184443187671001020100002020000200372003711100211091010100001000120640216221978510000102003820038200382003820038
10024200371500006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100000640216221978510000102003820038200382003820038
10024200371500006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100000640216221978510000102003820038200382003820038
10024200371500006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100000640216221978510000102003820038200382003820038
10024200371500006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100000640216221978510000102003820038200382003820038
100242003715000061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000120640216221978510000102003820038200382003820038
10024200371500006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100000640216221978510000102003820038200382003820038

Test 3: Latency 1->2

Code:

  sli v0.2d, v0.2d, #3
  movi v0.16b, 1

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 2.0037

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)9fa8acc5branch mispredict (cb)cdcfd6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
102042003715006119686251010010010000100100005002847521120018200372003718428061874110100200100082002001620037200371110201100991001001000010001201117171600198000100001002003820038200382003820038
10204200371500611968625101001001000010010000500284752112001820037200371842806187411010020010008200200162003720037111020110099100100100001000121117181600198010100001002003820038200382003820038
1020420037150061196862510100100100001001000050028475211200182003720037184280718740101002001000820020016200372003711102011009910010010000100001117181600198010100001002003820038200382003820038
1020420037150082196862510100100100001001000050028475211200182003720037184280618741101002001000820020016200372003711102011009910010010000100001117171600198012100001002003820038200382003820038
10204200371502461196862510100100100001001000050028475211200182003720037184280618741101002001000820020016200372003711102011009910010010000100001117181600198010100001002003820038200382003820038
102042003715006119686251010010010000100100005002847521120018200372003718428061874110100200100082002001620037200371110201100991001001000010001531117181600198010100001002003820038200382003820038
1020420037150061196862510100100100001001000050028475211200182003720037184280618740101002001000820020016200372003711102011009910010010000100001117181600198000100001002003820038200382003820038
10204200371500611968625101001001000010010000500284752112001820037200371842806187411010020010008200200162003720037111020110099100100100001000151117181600198000100001002003820038200382003820038
1020420037150061196862510100100100001001000050028496931200182003720037184280618740101002001000820020016200372003711102011009910010010000100001117181600198010100001002003820038200382003820038
1020420037150061196862510100100100001001000050028475211200182003720037184280618740101002001000820020016200372003711102011009910010010000100001117171600198010100001002003820038200382003820038

1000 unrolls and 10 iterations

Result (median cycles for code): 2.0037

retire (01)cycle (02)033f4e5051inst 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)9fa8accfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
100242003715061196860251001010100001010000502847521120018200372003718443318767100102010000202000020037200371110021109101010000100114640316331978610000102003820038200382003820038
1002420037150611968620021251001010100001010000502847521120018200372003718443318767100102010000202000020037200371110021109101010000100138640316331978610000102003820038200382003820038
1002420037150611968620021251001010100001010000502847521120018200372003718443318767100102010000202000020037200371110021109101010000100117640316331978610000102003820038200382003820038
10024200371506119686200212510010101000010100005028475211200182003720037184433187671001020100002020000200372003711100211091010100001003640316331978610000102003820038200382003820038
1002420037150611968602510010101000010100005028475211200182003720037184433187671001020100002020000200372003711100211091010100001000640316331978610000102003820038200382003820038
1002420037150611968602510010101000010100005028475211200182003720037184433187671001020100002020000200372003711100211091010100001016640316331978610000102003820038200382003820038
1002420037150611968602510010101000010100005028475211200182003720037184433187671001020100002020000200372003711100211091010100001000640316331978610000102003820038200382003820038
1002420037150611968602510010101000010100005028475211200182003720037184433187671001020100002020000200372003711100211091010100001000640316331978610000102003820038200382003820038
1002420037150611968602510010101000010100005028475211200182003720037184433187671001020100002020000200372003711100211091010100001000640316331978610000102003820038200382003820038
1002420037150611968602510010101000010100005028475211200182003720037184433187671001020100002020000200372003711100211091010100001000640316331978610000102003820038200382003820038

Test 4: throughput

Count: 8

Code:

  movi v0.16b, 0
  sli v0.2d, v8.2d, #3
  movi v1.16b, 0
  sli v1.2d, v8.2d, #3
  movi v2.16b, 0
  sli v2.2d, v8.2d, #3
  movi v3.16b, 0
  sli v3.2d, v8.2d, #3
  movi v4.16b, 0
  sli v4.2d, v8.2d, #3
  movi v5.16b, 0
  sli v5.2d, v8.2d, #3
  movi v6.16b, 0
  sli v6.2d, v8.2d, #3
  movi v7.16b, 0
  sli v7.2d, v8.2d, #3
  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)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)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)9fa7a8acc5branch mispredict (cb)cdcfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
16020420091150292580116100800161028002850064019612004420065200656128012820080028200160056200652006511160201100991001001600001002801111011901600200621600001002006620066200662006620066
16020420065150292580116100800161008002850064019602004420065200656128012820080028200160056200652006511160201100991001001600001000301111011901600200621600001002006620066200662006620066
16020420065150292580116100800161008002850064019612004420065200656128012820080028200160056200652006511160201100991001001600001000401111011901600200621600001002006620066200662006620066
16020420065151292580116100800161008002850064019612004420065200656128012820080028200160056200652006511160201100991001001600001000001111011901600200621600001002006620066200662006620066
160204200651512192580116100800161008002850064019612004420065200656128012820080028200160056200652006511160201100991001001600001000101111011901600200621600001002006620066200662006620066
16020420065150292580116100800161008002850064019612004420065200656128012820080028200160056200652006511160201100991001001600001000201111011901600200621600001002006620066200662006620066
16020420065150292580116100800161008002850064019612004420065200656128012820080028200160056200652006511160201100991001001600001000501111011901600200621600001002006620066200662006620066
16020420065150292580116100800161008002850064019612004420065200656128012820080028200160056200652006511160201100991001001600001000001111011901600200621600001002006620066200662006620066
16020420065150292580116100800161008002850064019612004420065200656128012820080028200160056200652006511160201100991001001600001000101111011901600200621600001002006620066200662006620066
16020420065150292580116100800161008002850064019612004420065200656128012820080028200160056200652006511160201100991001001600001000101111011901600200621600001002006620066200662006620066

1000 unrolls and 10 iterations

Result (median cycles for code divided by count): 0.2507

retire (01)cycle (02)0307193a3f51inst 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)5f60696d6escheduler 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)9fa8accfd0d2icache miss (d3)d5d6d9dadbddinst fetch restart (de)e0eaec? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
160024200751500004429800101080000108000050640000012003120052200503218001020800002016000020050200501116002110910101600001000100463112025211231820047201160000102005120053200532026520053
1600242005215000044278001010800001080000506400001120031200522005032180010208000020160000200522005011160021109101016000010446100413111825211221820047201160000102005320051200512026020051
160024200501500104427800101080000108000050640000112003120050200503218001020800002016000020050200501116002110910101600001003100453112325211182220049211160000102005320051200512024920051
160024200521500004427800101080000108000050640000112004020050200503218001020800002016000020050200501116002110910101600001006100463412425211182220047201160000102005320053200512026720062
160024200501500005029800101080000108000050640000012003120050200503218001020800002016000020050200591116002110910101600001000100403111725211211720047201160000102005120051200512024720060
160024200611500004429800101080000108000050640000012004020059200593218001020800002016000020059200591116002110910101600001000100466522036422221820056402160000102006020060200602026120051
160024200501500004429800101080297108000050640000012004020059200593218001020800002016000020059200591116002110910101600001006100466222036422202020056402160000102006020060200602028020060
1600242005915000071527800101080000108000050640000112003120050200503218001020800002016000020052200521116002110910101600001060100436221734422162020049402160000102006020060200602025120060
160024200611510005029800101080000108000050640000012004020059200503218001020800002016000020059200591116002110910101600001006100436522136422171820058402160000102006020060200602026620060
160024200611500005029800101080000108000050640000012004020061200593218001020800002016000020061200591116002110910101600001000100446521934422212320058412160000102006020060200602026820062

Test 5: throughput

Count: 16

Code:

  sli v0.2d, v16.2d, #3
  sli v1.2d, v16.2d, #3
  sli v2.2d, v16.2d, #3
  sli v3.2d, v16.2d, #3
  sli v4.2d, v16.2d, #3
  sli v5.2d, v16.2d, #3
  sli v6.2d, v16.2d, #3
  sli v7.2d, v16.2d, #3
  sli v8.2d, v16.2d, #3
  sli v9.2d, v16.2d, #3
  sli v10.2d, v16.2d, #3
  sli v11.2d, v16.2d, #3
  sli v12.2d, v16.2d, #3
  sli v13.2d, v16.2d, #3
  sli v14.2d, v16.2d, #3
  sli v15.2d, v16.2d, #3
  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)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)9fa0a1a6a8acc5branch mispredict (cb)cdcfd6e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
16020440038300921925160108100160008100160020500128013214001940038400381997761998916012020016003220032006440038400381116020110099100100160000100000101111011816400351600001004003940039400394003940039
16020440038300071251601081001600081001600205001280132140019400384003819977619989160120200160032200320064400384003811160201100991001001600001000002561111011816400351600001004003940039400394003940039
160204400383001271251601081001601081001600205001280132140019400384003819977619989160120200160032200320064400384008611160201100991001001600001002205001111011816400351600001004003940039400394003940039
16020440038300029251601081001600081001600205001280132140019400384003819977619989160120200160032200320064400384003811160201100991001001600001000001031111011816400351600001004003940039400394003940039
1602044003830002925160108100160008100160020500128013214001940038400381997761998916012020016003220032006440038400381116020110099100100160000100000801111011816400351600001004003940039400394003940039
16020440038300217125160108100160008100160020500128013214001940038400381997761998916012020016003220032006440038400381116020110099100100160000100000001111011816400351600001004003940039400394003940039
16020440038300029251601081001600081001600205001280132140019400384003819977619989160120200160032200320064400384003811160201100991001001600001000005901111011816400351600001004003940039400394003940039
1602044003830002925160108100160008100160020500128013214001940038400381997761998916012020016003220032006440038400381116020110099100100160000100000001111011816400351600001004003940039400394003940039
1602044003829902925160108100160008100160020500128013214001940038400381997761998916012020016003220032006440038400381116020110099100100160000100000101111011816400351600001004003940039400394003940039
1602044003829902925160108100160008100160020500128013214001940038400381997761998916012020016003220032006440038400381116020110099100100160000100000101111011816400351600001004003940039400394003940039

1000 unrolls and 10 iterations

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

retire (01)cycle (02)03081e3f51inst 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)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)9fa0a6a8a9accfd0d2icache miss (d3)d5d6d9dadbddinst fetch restart (de)e0eaeb? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
16002440046300004525160010101600001016000050128000011400194003840038199960320018160010201601322032000040038400381116002110910101600001000000100256211216422910400354016160000104003940039400394003940039
1600244003829900452516001010160000101600005012800001140019400384003819996032001816001020160000203200004003840038111600211091010160000100000010025622916422107400354016160000104003940039400394003940039
160024400383000045251600101016000010160000501280000104001940038400381999603200181600102016000020320000400384003811160021109101016000010000001002231151621151040035208160000104003940039400394003940039
160024400382990051251600101016000010160000501280000114001940038400381999603200181600102016000020320000400384003811160021109101016000010200001002532151622110540035408160000104003940039400394003940039
16002440038300004525160010101600001016000050128000011400194003840038199960320018160010201600002032000040038400381116002110910101600001000000100253121016412105400352016160000104003940039400394003940039
16002440038300004525160010101600001016000050128000000400194003840038199960320018160010201600002032000040038400381116002110910101600001000000100256221016422510400354016160000104003940039400394003940039
1600244009130000512516001010160000101600005012800001040019400384003819996032001816001020160000203200004003840038111600211091010160000100000010025312516412510400352016160000104003940039400394003940039
1600244003830000512516001010160000101600005012800001140019400384003819996032001816001020160000203200004003840038111600211091010160000100000010022311101621161140035208160000104003940039400394003940039
1600244009031300452516001010160000101600005012800001140019400384003819996032001816001020160000203200004003840038111600211091010160000100000910022311101621151040035208160000104003940039400394003940039
1600244003829900710251600101016000010160000501280000014001940038400381999603200181600102016000020320000400384003811160021109101016000010000001002231151621151040035208160000104003940039400394003940039