Apple M1 Microarchitecture Research by Dougall Johnson

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

SADDW (vector, 4S)

Test 1: uops

Code:

  saddw v0.4s, v0.4s, 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)030b191e1f3f4e51inst 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
1004203715000061168725100010001000264680120182037203715723189510001000200020372037111001100001573116111787100020382038203820382038
1004203715000061168725100010001000264680120182037203715723189510001000200020372037111001100001873116111787100020382038203820382038
1004203715000061168725100010001000264680120182037203715723189510001000200020372037111001100001873116111787100020382038203820382038
100420371600006116872510001000100026468012018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
100420371500006116872510001000100026468012018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
1004203715000061168725100010001000264680120182037203715723189510001000200020372037111001100002173116111787100020382038203820382038
100420371500006116872510001000100026468012018203720371572318951000100020002037203711100110000373116111787100020382038203820382038
100420371500006116872510001000100026468012018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
100420371500006116872510001000100026322712018203720371572318951000100020002037203711100110000973116111787100020382038203820382038
1004203715000061168725100010001000264680120182037203715723189510001000200020372037111001100002173116111787100020382038203820382038

Test 2: Latency 1->2

Code:

  saddw v0.4s, v0.4s, 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)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)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
102042003715018611968725101001001000010010000500284768002001820037200371842231874510100200100002002000020037200371110201100991001001000010046007101161119791100001002003820038200382003820038
1020420037150061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
1020420037150061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
1020420037150061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
1020420037150061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
1020420037150061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
1020420037150061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
1020420037150061196672510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
10204200371500212196872510100102100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038
102042003715016561196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000007101161119791100001002003820038200382003820038

1000 unrolls and 10 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)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)9faccfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
100242003715006119687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100640216221978510000102003820038200382003820038
100242003715006119687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100640216221978510000102003820038200382003820038
100242003715008419687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100640216221978510000102003820038200382003820038
100242003715006119687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100640216221978510000102003820038200382003820038
100242003715006119687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100640216221978510000102003820038200382003820038
100242003715006119676251001013100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100640216221978510000102003820038200382003820038
100242003715006119687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100640216221978510000102003820038200382003820038
100242003715006119687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100640216221978510000102003820038200382003820038
100242003715006119687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100640216221978510000102003820038200382003820038
100242003715006119687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100640216221978510000102003820038200382003820038

Test 3: Latency 1->3

Code:

  saddw v0.4s, v1.4s, 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)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)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)9fa8a9acc2c5branch mispredict (cb)cdcfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10204200371490611968725101001001000010010000500284768012001820037200371842971874110100200100082002001620037200371110201100991001001000010000021117170160019802100001002003820038200382003820038
10204200371505521261965425101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000007101161119791100001002003820038200382003820038
1020420037150911061968725101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000007101161119791100001002003820038200382003820038
102042003715007261968725101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000007101161119791100001002003820038200382003820038
10204200371509821968725101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000300007101161119791100001002003820038200382003820038
10204200371506611968725101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000007101161119791100001002003820038200382003820038
10204200371500611968725101001131000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000007101161119791100001002003820038200382003820038
10204200371500611968725101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000007101161119791100001002003820038200382003820038
102042003715024611968725101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000007101161119791100001002003820038200382003820038
10204200371500611968725101001001000010310000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000007101161119791100001002003820038200382003820038

1000 unrolls and 10 iterations

Result (median cycles for code): 2.0037

retire (01)cycle (02)030818191e1f3f4e51inst 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)9fa0a6a7a8a9acc2branch mispredict (cb)cdcfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
1002420037150000006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100000000006403162219785010000102003820038200382003820038
1002420037150000006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100001000006402162219785010000102003820038200382003820038
1002420037150000006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100000030006402162219785010000102003820038200382003820038
1002420037150000006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100001060006402162219785010000102003820038200382003820038
1002420037150000006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100000000006402162219785010000102003820038200382003820038
1002420037150000906119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100001000006402162219785010000102003820038200382003820038
1002420037150000006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100006030006402162219785010000102003820038200382003820038
1002420037150000006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100002030006402162219785010000102003820038200732003820038
1002420037150000006119687251001010100001010000502847680120018200372003718444318767100102010000202000020037200371110021109101010000100000030006402162219785010000102003820038200382003820038
1002420037150000006119687251001010100001010000502847680020018200372003718444318767100102010000202000020037200371110021109101010000100003000006402162219785010000102003820038200382003820038

Test 4: throughput

Count: 8

Code:

  saddw v0.4s, v8.4s, v9.4h
  saddw v1.4s, v8.4s, v9.4h
  saddw v2.4s, v8.4s, v9.4h
  saddw v3.4s, v8.4s, v9.4h
  saddw v4.4s, v8.4s, v9.4h
  saddw v5.4s, v8.4s, v9.4h
  saddw v6.4s, v8.4s, v9.4h
  saddw v7.4s, v8.4s, 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)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)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)9fa1a6a8a9acbranch mispredict (cb)cfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
80204200581500230258010010080000100800005006400002001920038200389973399968010020080000200160000200382003811802011009910010080000100001030511021611200350800001002003920039200392003920039
8020420038150040258010010080000100800005006400002001920038200389973399968010020080000200160000200382003811802011009910010080000100001000511011611200350800001002003920039200392003920039
80204200381500402580100100800001008000050064000020019200382009199736999680100200800002001601942003820038118020110099100100800001000000005110116112003519800001002003920095200392003920039
8020420098150040258010010080000100800005006400002001920038200389973399968010020080000200160000200382003811802011009910010080000100000200511011611200350800001002003920039200392003920039
80204200381500402580100100800001008000050064000020019200382003899733999680100200800002001600002003820038118020110099100100800001000055090511011611200350800001002003920039200392003920039
8020420038150040258010010080000100800005006400002001920038200389973399968010020080000200160000200382003811802011009910010080000100001000511011611200350800001002003920039200392003920039
8020420038150040438021310080000100803885006400002001920038200389973399968010020080000200160000200382003811802011009910010080000100001030511011611200350800001002003920039200392003920039
8020420038150040258010010080000100800005006400002001920038200389973399968010020080000200160000200382003811802011009910010080000100000000511011611200350800001002003920039200392003920039
8020420038150040258010010080000100800005006400002001920038200389973399968010020080000200160000200382003811802011009910010080000100001030511011611200350800001002003920039200392003920039
8020420038150040258010010080000100800005006400002001920038200389973399968010020080000200160000200382003811802011009910010080000100000000511011611200350800001002003920039200392003920039

1000 unrolls and 10 iterations

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

retire (01)cycle (02)0304083f51inst 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)9fa0a1a8acc2cficache miss (d3)itlb miss (d4)d5d6ddinst fetch restart (de)e0ea? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
8002420047150003925800101080000108000050640000012001920038200389996310018800102080000201600002003820038118002110910108000010001005021011816181820035080000102003920039200392003920039
8002420038150103925800101080000108000050640000012001920038200389996310018800102080000201600002003820038118002110910108000010000005021011816171720035080000102003920039200392003920039
8002420038150103925800101080000108000050640000002001920038200389996310018800102080000201600002003820038118002110910108000010000005021011816181820035080000102003920039200392003920039
8002420038150103925800101080000108000050640000012001920038200389996310018800102080000201600002003820038118002110910108000010000005020001016181420035080000102003920039200392003920039
8002420038150003925800101080000108000050640000012001920038200389996310018800102080000201600002003820038118002110910108000010000005020001516171620035080000102003920039200392003920039
8002420038150003925800101080000108000050640000012001920038200389996310018800102080000201600002003820038118002110910108000010000005020001816181720035080000102003920039200392003920039
8002420038150003925800101080000108000050640000012001920038200389996310018800102080000201600002003820038118002110910108000010000005020001816181720035080000102003920039200392003920039
8002420038150003925800101080000108000050640000012001920038200389996310018800102080000201600002003820038118002110910108000010001005020001716181820035080000102003920039200392003920039
8002420038150003925800101080000108000050640000012001920038200389996310018800102080000201600002003820038118002110910108000010000005020001716151820035080000102003920039200392003920039
8002420038150003925800101080000108000050640000012001920038200389996310018800102080000201600002003820038118002110910108000010000305020001016181520035080000102003920039200392003920039