Apple M1 Microarchitecture Research by Dougall Johnson

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

FMAX (vector, 4H)

Test 1: uops

Code:

  fmax v0.4h, v0.4h, v1.4h
  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)03183f4e51inst 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)accfd5d6ddinst fetch restart (de)e0? fp/simd (ee)f5f6f7f8fd
10042037150611687251000100010002646800201820372037157131895100010002000203720371110011000073116111785100020382038203820382038
10042037150611687251000100010002646801201820372037157231895100010002000203720371110011000075116111787100020382038203820382038
100420371502511687251000100010002646800201820372037157231895100010002000203720371110011000073116111787100020382038203820382038
10042037150611687251000100010002646800201820372037157231895100010002000203720371110011000373116111787100020382038203820382038
10042037150611687251000100010002646800201820372037157231895100010002000203720371110011000073116111787100020382038203820382038
10042037150611687251000100010002646801201820372037157231895100010002000203720371110011000073116111787100020382038203820382038
10042037150611687251000100010002646801201820372037157231895100010002000203720371110011000073116111787100020382038203820382038
10042037160611687251000100010002646800201820372037157231895100010002000203720371110011000073116111787100020382038203820382038
10042037150611687251000100010002646801201820372037157231895100010002000203720371110011000073116211787100020382038203820382038
10042037150611687251000100010002646801201820372037157131895100010002000203720371110011000073116111787100020382038203820382038

Test 2: Latency 1->2

Code:

  fmax v0.4h, v0.4h, v1.4h
  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)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)9fa8acc5cfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
102042003715000061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
102042003715000061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
102042003715000061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
1020420037150100457196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
102042003715000061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
102042003715000061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
102042003715000061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
102042003715000061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001001907101161119791100001002003820038200382003820038
10204200371500153104179196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
102042007415000061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038

1000 unrolls and 10 iterations

Result (median cycles for code): 2.0037

retire (01)cycle (02)0318191e1f3f4e51inst 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)9fa8a9cfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10024200371500000124196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
1002420037150000061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
10024200371500000130196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
10024200371500000149196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
10024200371500020788170196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
10024200371500000536196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
1002420037150000061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
1002420037150000061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
10024200371500069061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
10024200371500000145196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038

Test 3: Latency 1->3

Code:

  fmax v0.4h, v1.4h, v0.4h
  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)0318191e3f4e51inst 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)9fa8acc5branch mispredict (cb)cdcfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10204200371500006119687441010010010000100100005002847680120018200372003718429718740101002001000820020016200372003711102011009910010010000100001117170160019801100001002003820038200382003820038
10204200371500006619687251010010010000100100005002847680120018200372003718429718740101002001000820020016200372003711102011009910010010000100001117170160019801100001002003820038200382003820038
10204200371500006119687251010010010000100100005002847680120018200372003718429618741101002001000820020016200372003711102011009910010010000100001117170160019801100001002003820038200382003820038
10204200371500006119687251010010010000100100005002847680120018200372003718429618741101002001000820020016200372003711102011009910010010000100001117170160019836100001002003820038200382003820038
10204200371500006119687251010010010000100100005002847680120018200372003718429718741101002001000820020016200372003711102011009910010010000100001117170160019801100001002003820038200382003820038
1020420037150131716119687251010010010000100100005002847680120018200372003718429618741101002001000820020016200372003711102011009910010010000100001117170160019801100001002003820038200382003820038
10204200371500006119687251010010010000100100005002847680120018200372003718429718741101002001000820020016200372003711102011009910010010000100001117180160019801100001002003820038200382003820038
10204200371500008419687251010010010000100100005002847680020018200372003718429618740101002001000820020016200372003711102011009910010010000100001117170160019802100001002003820038200382003820038
10204200371500006119687251010010010000100100005002847680020018200372003718429718740101002001000820020016200372003711102011009910010010000100001117170160019802100001002003820038200382003820038
10204200371500006119687251010010010000100100005002847680020018200372003718429618741101002001000820020016200372003711102011009910010010000100001117170160019802100001002003820038200382003820038

1000 unrolls and 10 iterations

Result (median cycles for code): 2.0037

retire (01)cycle (02)030b181e1f3a3f4e51inst 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)9fcfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
100242003715000900143196872510010101000010100005028476802001820037200371844431876710010201000020200002003720037111002110910101000010640316331978510000102003820038200382003820038
10024200371500100061196872510010101000010100005028476802001820037200371844431876710010201000020200002003720037111002110910101000010640316331978510000102003820038200382003820038
10024200371500000061196872510010101000010100005028476802001820037200371844431876710010201000020200002003720037111002110910101000010640316331978510000102003820038200382003820038
10024200371500000061196872510010101000010100005028476802001820037200371844431876710010201000020200002003720037211002110910101000010640316331978510000102003820038200382003820038
10024200371500000066196872510010101000010100005028476802001820037200371844431876710010201000020200002003720037111002110910101000010640316331978510000102003820038200382003820038
100242003715000000145196872510010101000010100005028476802001820037200371844431876710010201000020200002003720037111002110910101000010640316331978510000102003820038200382003820038
10024200371500000182196872510010101000010100005028476802001820037200371844431876710010201000020200002003720037111002110910101000010640316331978510000102003820038200382003820038
10024200371500000061196872510010101000010100005028476802001820037200371844431876710010201000020200002003720037111002110910101000010640316331978510000102003820038200382003820038
100242003715000180061196872510010101000010100005028476802001820037200371844431876710010201000020200002003720037111002110910101000010640316331978510000102003820038200382003820038
10024200371500000061196872510010101000010100005028476802001820037200371844431876710010201000020200002003720037111002110910101000010640316331978510000102003820038200382003820038

Test 4: throughput

Count: 8

Code:

  fmax v0.4h, v8.4h, v9.4h
  fmax v1.4h, v8.4h, v9.4h
  fmax v2.4h, v8.4h, v9.4h
  fmax v3.4h, v8.4h, v9.4h
  fmax v4.4h, v8.4h, v9.4h
  fmax v5.4h, v8.4h, v9.4h
  fmax v6.4h, v8.4h, v9.4h
  fmax v7.4h, v8.4h, v9.4h
  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)030b18191e1f3f51inst 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
80204200491500000014725801001008000010080000500640000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511021611200350800001002003920039200392003920039
8020420038150000004025801001008000010080000500640000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511011611200790800001002003920039200392003920039
8020420189155000004025801001008000010080000500640000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511011611200350800001002003920039200392003920039
8020420038150000004025801001008000010080000500640000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511011611200350800001002003920039200392003920039
8020420038150000004025801001008000010080000500640000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511011611200350800001002003920039200392003920039
8020420038150000004025801001008000010080000500640000020019200382003899733999680100200800002001600002003820038118020110099100100800001002430000511011611200350800001002003920039200392003920039
8020420038150000004025801001008000010080000500640000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511011611200350800001002003920039200392003920039
8020420038150000004025801001008000010080000500640000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511011611200350800001002003920039200392003920039
8020420038150000004025801001008000010080000500640000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511011611200350800001002003920039200392003920039
8020420038150000004025801001008000010080000500640000020019200382003899723999680100200800002001600002003820038118020110099100100800001000000000511011611200350800001002003920039200392003920039

1000 unrolls and 10 iterations

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

retire (01)cycle (02)03080b18191e1f3a3f51inst 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)9fa0a1a6a7a8a9acc2cfd5d6dbddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
80024200391500000000392580010108000010800005064000002001920038200389996031001880010208000020160000200382003811800211091010800001000000000502031602320035080000102003920039200392003920039
80024200381500000000392580010108000010800005064000002001920038200389996031001880010208000020160000200382003811800211091010800001000000000502021602320035080000102003920039200392003920039
80024200381500000000392580010108000010800005064000002001920038200389996031001880010208000020160000200382003811800211091010800001000000200503731603220035080000102003920039200392003920039
80024200381500000000392580010108000010800005064000002001920038200389996031001880010208000020160000200382003811800211091010800001000000000502031603220035080000102003920039200392003920039
80024200381500000000392580010108000010800005064000002001920038200389996031001880010208000020160000200382003811800211091010800001000000000502031703320035080000102003920039200392003920039
80024200381550000000392580010108000010800005064000002001920038200389996031001880010208000020160000200382003811800211091010800001000000000502021604320035080000102003920039200392003920039
80024200381500000000392580010108000010800005064000002001920038200389996031001880010208000020160000200382003811800211091010800001000000000502031602320035080000102003920039200392003920039
8002420038150000090010225800101080000108000050640000020019200382003899960310018800102080000201600002003820038118002110910108000010000000120502021603320035080000102003920039200392003920039
80024200381500000000622580010108000010800005064000002001920038200389996031001880010208000020160000200382003811800211091010800001000000000502031603320035080000102003920039200392003920039
80024200381500000000392580010108000010800005064000002001920038200389996031001880010208000020160000200382003811800211091010800001000000000502031603320035080000102003920039200392003920039