Apple M1 Microarchitecture Research by Dougall Johnson

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

ZIP1 (vector, 4H)

Test 1: uops

Code:

  zip1 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)0308090b18191e1f3a3f4e51inst 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)a0a1a6a8a9acc2cfd5d6ddinst fetch restart (de)e0? fp/simd (ee)f5f6f7f8fd
100420371600000000611687251000100010002646800201820372037157231895100010002000203720371110011000120000073116111787100020382038203820382038
100420371600000000611687251000100010002646800201820372037157231895100010002000203720371110011000000000073116111787100020382038203820382038
1004203716000000006116872510001000100026468012018203720371572318951000100020002037203711100110000000021073116111787100020382038203820382038
100420371500000000611687251000100010002646801201820372037157231895100010002000203720371110011000000100073116111787100020382038203820382038
100420371500000000611687251000100010002646801201820372037157231895100010002000203720371110011000000003073116111787100020382038203820382038
1004203716000000006116872510001000100026468012018203720371572318951000100020002037203711100110000000021073116111787100020382038203820382038
10042037150001013200611687251000100010002646800201820372037157231895115210002000203720371110011000000000073116111787100020382038203820382038
100420371500000000611687251000100010002646800201820372037157231895100010002000208520371110011000000303073116111787100020382038203820382038
1004203715000000011031687251000100010002646800201820372037157231895100010002000203720371110011000000003073116111787100020382038203820382038
100420371600000000611687251000100010002646800201820372037157231895100010002000203720371110011000000003073116111787100020382038203820382038

Test 2: Latency 1->2

Code:

  zip1 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)0308090b18191e1f3a3f4e51inst 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)9fa0a1a6a8a9acc2c5cfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
102042003715500000150061196872510100100100001001000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000000000071021622197910100001002003820038200382003820038
102042003715500000150061196872510100100100001001000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000000000071021622197910100001002003820038200382003820038
10204200371550000000061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000000000071021622197910100001002003820038200382003820038
102042003715600000120061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000001000071021622197910100001002003820038200382003820038
1020420037156000001200103196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000000000071021622197910100001002003820038200382003820038
1020420037156000001050061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000000000071021622197910100001002003820038200382003820038
102042003715600000480061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000000000071021622198080100001002003820038200382003820038
102042003715500000210061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000000000071021622197910100001002003820038200382003820038
10204200371550000000061196872510100100100001001000050028476801200182003720037184223187451010020010000200200002003720037111020110099100100100001000000000071021622197910100001002003820038200382003820038
10204200371560000000061196872510100100100001001000050028476800200182003720037184223187451010020010000200200002003720037111020110099100100100001000000000071021622197910100001002003820038200382003820038

1000 unrolls and 10 iterations

Result (median cycles for code): 2.0037

retire (01)cycle (02)03090b1e3f4e51inst 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
1002420037155000437196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820131200382003820038
100242003715500061196872510010101000010100005028476801200542003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
100242003715500061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
100242003715500084196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
1002420037155004561196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
100242003715000061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221992910000102003820038200382003820038
100242003715000061196872510010101000010100005028476801200182003720037184443187671001020100002020000200372003711100211091010100001000640216321978510000102003820038200382003820038
100242003714900061196872510010101000010100005028476801200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
100242003715000061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038
100242003715000061196872510010101000010100005028476801200182003720037184443187671001020100002020000200372003711100211091010100001000640216221978510000102003820038200382003820038

Test 3: Latency 1->3

Code:

  zip1 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)030818191e3f4e51inst 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)9fa0a1a6a7a8acc2c5branch mispredict (cb)cdcfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
1020420037155000276119687251010010010000100100005002847680120018200372003718429718740101002001000820020016200372022611102011009910010010000100000000011171701600198020100001002003820038200382003820038
1020420037156004061196412510100100100001001000050028476801200182003720037184292418741101002001000820020016200372003711102011009910010010000100000000011171701600198010100001002003820038200382003820038
1020420037155000063119687251010010010000100100005002847680120018200372003718429718741101002001000820020016200372003711102011009910010010000100000000011171701600198010100001002003820038200382003820083
102042003715500006119687251010010010000100100005002847680120018200372003718429618741101002001000820020016200372003711102011009910010010000100000000011171701600199090100001002003820038200382003820133
102042022615500006119687251010010010000100100005002847680120018200372003718422318745101002001000020020000200372003711102011009910010010000100203000000071011611197910100001002003820038200382003820038
1020420037156000366119687251010010010000100100005002847680120018200372003718422318745101002001000020020000200372003711102011009910010010000100000000000071011611197910100001002003820038200382003820038
1020420037155000061196872510100100100001001000050028476801200182003720037184222118745101002001000020020000200372003711102011009910010010000100000000000071011611197910100001002003820038200382003820038
102042003715500006119687251010010010000100100005002847680120018200372003718422318745101002001000020020000200372003711102011009910010010000100000000000071011611197910100001002003820038200382003820038
102042003715500006119687251010010010000100100005002847680120018200372003718422318745101002001000020020000200372003711102011009910010010000100000010000071011611197910100001002003820038200382003820038
1020420037155000126119687251010010010000100100005002847680120018200372003718422318745101002021000020020000200372003711102011009910010010000100000003000080511611197910100001002003820038200382003820038

1000 unrolls and 10 iterations

Result (median cycles for code): 2.0037

retire (01)cycle (02)031e3f494e51inst 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)9fa8acc2c5branch mispredict (cb)cfd0d5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
1002420037155396101968725100101010000101000050284768002001820037200371844431876710010201000020200002003720037111002110910101000010000006400316331978510000102003820038200382003820038
100242003715506101968725100101010000101000050284768012001820037200371844431876710010201000020200002003720037111002110910101000010030006400316331978510000102003820038200382003820038
100242003715506101968725100221010000101000050284768012001820037200371844431876710010201000020200002003720037111002110910101000010000006400316331978510000102003820038200382003820038
100242003716112173251968725100101010000101000050284768002001820037200371844431876710010201000020200002003720037111002110910101000010000006400316331978510000102003820038200382003820038
100242003719306101968725100101010000101000050284768012001820037200371844431876710010201000020200002003720037111002110910101000010000006400316331978510000102003820038200382003820038
100242003715506101968725100101010000101000050284768012001820037200371844431876710010201000020200002003720037111002110910101000010000006400316331978510000102003820038200382003820038
100242003715506101968725100101010000101000050284768012001820037200371844431876710010201000020200002003720037111002110910101000010000006400316331978510000102003820038200382003820038
10024200371554810301968725100101010000101000050284768012001820037200371844431876710010201000020200002003720037111002110910101000010000006400316331978510000102003820038200382003820038
1002420037156656401968725100101010000101000050284768012001820037200371844431876710010201000020200002003720037111002110910101000010000006400316331978510000102003820038200382003820038
100242003715506101968725100101010000101000050284768012001820037200371844431876710010201000020200002003720037111002110910101000010000006400316331978510000102003820038200382003820038

Test 4: throughput

Count: 8

Code:

  zip1 v0.4h, v8.4h, v9.4h
  zip1 v1.4h, v8.4h, v9.4h
  zip1 v2.4h, v8.4h, v9.4h
  zip1 v3.4h, v8.4h, v9.4h
  zip1 v4.4h, v8.4h, v9.4h
  zip1 v5.4h, v8.4h, v9.4h
  zip1 v6.4h, v8.4h, v9.4h
  zip1 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)0318191e3a3f51inst 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? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
80204200381550000515258010010080000100800005006400000200192003820038997339996801002008000020016000020038200381180201100991001008000010000000511021611200350800001002011320193201132013920039
8020420038155000040258010010080000100800005006400000200192003820038997339996801002008000020016000020038200381180201100991001008000010000000511011611200350800001002003920039200392003920039
80204200381550000402580100100800001008000050064000012001920038200389973810023802052008009620016019020038200871180201100991001008000010000000511011611200350800001002003920395201092008820088
8020420038155110068258010010080000100800005006400000200192003820038997339996801002008000020016000020038200381180201100991001008000010000000511011611200350800001002003920039200392003920039
8020420038155000040258010010080000115800005006400000200192003820038997339996801002008000020016000020038200381180201100991001008000010000000511011611200350800001002003920039200392003920039
8020420038155000040258010010080000100800005006400001200192003820038997339996801002008000020016000020038200381180201100991001008000010000000511011611200350800001002003920039200392003920039
8020420038155000061258010010080000100800005006400001200192003820038997339996801002008000020016000020038200381180201100991001008000010000000511011611200350800001002003920039200392003920039
80204200381550000402580100100800001008000050064000002001920038200389973399968010020080000200160000200382003811802011009910010080000100000005110116112003510800001002003920039200392003920039
80204200381610012040258010010080000100800005006400000200192003820038997339996801002008000020016000020038200381180201100991001008000010000000511011611200350800001002003920039200392003920039
8020420038155000040258010010080000100800005006400001200192003820038997339996801002008000020016000020038200381180201100991001008000010000000511011611200350800001002003920039200392003920039

1000 unrolls and 10 iterations

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

retire (01)cycle (02)0307080b1e1f3f51inst 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)9fa1a7a8a9accfd5d6ddinst fetch restart (de)e0? int output thing (e9)ea? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
8002420039155100003925800101080000108000050640000200192003820038999631001880010208009720160000200382003811800211091010800001021023502051667200350080000102003920039200392003920039
800242003815500008848125800101080000108000050640000200192003820038999631001880010208000020160000200382003811800211091010800001000000502081676200350080000102003920039200392003920039
8002420038155000003925800101080000108000050640000200192003820038999631001880010208000020160000200382003811800211091010800001000000502061665200350080000102003920039200392003920039
8002420038155000003925800101080000108000050640000200192003820038999631001880010208000020160000200382003811800211091010800001000000502051656200350080000102003920039200392003920039
8002420038156000008125800101080000108000050640000200192003820038999631001880010208000020160000200382003811800211091010800001000000502071685200350080000102003920039200392003920039
8002420038155000003925800101080000108000050640000200192003820038999631001880010208000020160000200382003811800211091010800001000000502051676200350080000102003920039200392003920039
800242003815601021603925800101080095108019750640000200192003820091999681004680010208000020160000200872003811800211091010800001000000502051665200350080000102003920039200392003920039
8002420038155000003925800101080000108000050640000200192003820038999631001880010208000020160000200382003811800211091010800001000000502051656200350080000102003920039200392003920039
8002420038155000003925800101080000108000050640000200192003820038999631001880010208000020160000200382003811800211091010800001000900502051675200350080000102003920039200392003920039
8002420038155000003925800101080000108000050640000200192003820038999631001880010208000020160000200382003811800211091010800001000000502061655200350080000102003920039200392003920039