Apple M1 Microarchitecture Research by Dougall Johnson

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

TRN1 (vector, 2D)

Test 1: uops

Code:

  trn1 v0.2d, v0.2d, 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)031e3f4e51inst issue (52)~issue fp/simd (54)~dispatch fp/simd (57)huge thing fp/simd (5b)696d6edispatch 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
10042037150611687251000100010002646802018203720371572318951000100020002037203711100110000373116111787100020382038203820382038
10042037150611687251000100010002646802018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
10042037150611687251000100010002646802018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
10042037150611687251000100010002646802018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
10042037150611687251000100010002646802018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
10042037150611687251000100010002646802018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
100420371501471687251000100010002646802018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
10042037160611687251000100010002646802018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
10042037150611687251000100010002646802018203720371572318951000100020002037203711100110000073116111787100020382038203820382038
100420371575611687251000100010002646802018203720371572318951000100020002037203711100110000073116111787100020382038203820382038

Test 2: Latency 1->2

Code:

  trn1 v0.2d, v0.2d, 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): 2.0037

retire (01)cycle (02)0307080a0b18191e1f3a3f4e51inst 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)9fa0a1a6a8a9acc2cfd0d5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10204200371500000000006119687251010010010000100100005002847680120018020037200371842231874510100200100002002000020037200371110201100991001001000010000000570710021622197910100001002003820038200382003820038
102042003715000000000061196872510100100100001001000050028476801200180200372003718422318745101002001000020020000200372003711102011009910010010000100000201020710021622197910100001002003820038200382003820038
102042003715000000000061196872510100100100001001000050028476801200180200372003718422318745101002001000020020000200372003711102011009910010010000100000001170710021622197910100001002003820038200382003820038
102042003715000000000061196872510100100100001001000050028476801200180200372003718422318745101002001000020020000200372003711102011009910010010000100000460120710021622197910100001002003820038200382003820038
1020420037150000000000611968725101001001000010010000500284768012001802003720037184223187451010020010000200200002003720037111020110099100100100001000000000710021622197910100001002003820038200382003820038
1020420037150000000000611968725101001001000010010000500284768012001802003720037184223187451010020010000200200002003720037111020110099100100100001000001090710021622197910100001002003820038200382003820038
1020420037149000000000611968725101001001000010010000500284768012001802003720037184223187451010020010000200200002003720037111020110099100100100001000001030710021622197910100001002003820038200382003820038
102042003715000000000061196872510100100100001001000050028476801200180200372003718422318745101002001000020020000200372003711102011009910010010000100000001380710021622197910100001002003820038200382003820038
102042003715000000000061196872510100100100001001000050028476801200180200372003718422318745101002001000020020000200372003711102011009910010010000100000201320710021622197910100001002003820038200382003820038
1020420037150000000000611968725101001001000010010000500284768012001832003720037184223187451010020010000200200002003720037111020110099100100100001000003000710021622197910100001002003820038200382003820038

1000 unrolls and 10 iterations

Result (median cycles for code): 2.0037

retire (01)cycle (02)03181e3f4e51inst 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)9fa6a8acc5cfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
1002420037150006119687251001010100001010000502847680200182003720037184443187671001020100002020000200372003711100211091010100001000006403163319785010000102003820038200382003820038
100242003716205676119687251001010100001010000502847680200182003720037184443187671001020100002020000200372003711100211091010100001000306403163319785010000102003820038200382003820038
1002420037150006119687251001010100001010000502847680200182003720037184443187671001020100002020000200372003711100211091010100001000006403163319785010000102003820038200382003820038
1002420037150036119687251001010100001010000502847680200182003720037184443187671001020100002020000200372003711100211091010100001000016403163319785010000102003820038200382003820038
1002420037150066119687251001010100001010000502847680200182003720037184443187671001020100002020000200372003711100211091010100001000006403163319785010000102003820038200382003820038
10024200371500072619687251001010100001010000502847680200182003720037184443187671001020100002020000200372003711100211091010100001000006403163319785010000102003820038200382003820038
1002420037149006119687251001010100001010608502847680200182003720037184443187671001020100002020000200372003711100211091010100001000006403163319785010000102003820038200382003820038
1002420037150006119687251001010100001010000502847680200182003720037184443187671001020100002020000200372003711100211091010100001000006403163319785010000102003820038200382003820038
1002420037150006119687251001010100001010000502847680200182003720037184443187671001020100002020000200372003711100211091010100001000006403163319785010000102003820038200382003820038
1002420037150006119687251001010100001010000502847680200182003720037184443187671001020100002020000200372003711100211091010100001000006403163319785010000102003820038200382003820038

Test 3: Latency 1->3

Code:

  trn1 v0.2d, v1.2d, v0.2d
  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)030b1e1f3f4e51inst 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)9fa1a6a7a8acc5cfd0d5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10204200371500120611968725101001001000010010000500284768012001820037200371842231874510100200100002022000020037200841110201100991001001000010000000071001161119791100001002003820038200382003820038
102042003715000012461968725101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000071001161119791100001002003820038200382003820038
1020420037150000611968725101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000071001161119791100001002003820038200382003820038
1020420037150000611968725101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000000071001161119791100001002003820038200382003820038
1020420037150000611968725101001001000010010000500284768012001820037200371842231874510100200100002002000020037200371110201100991001001000010000002068071001161119791100001002003820038200382003820038
1020420037150000611968725101001001000010010000500284768002001820037200371842231874510100200100002002000020037200371110201100991001001000010000000071001161119791100001002003820038200382003820038
1020420037150000611968725101001001000010010000500284768002001820037200371842231874510100200100002002000020037200371110201100991001001000010000000071001161119791100001002003820038200382003820038
1020420037150090841968725101001001000010010000500284768002001820037200371842231874510100200100002002000020037200371110201100991001001000010000000071001161119791100001002003820038200382003820038
1020420037150000611968725101001001000010010000500284768002001820037200371842231874510100200100002002000020037200371110201100991001001000010000000071001161119791100001002003820038200382003820038
10204200371500004121968725101001001000010010000500284768002001820037200371842231874510100200100002002000020037200371110201100991001001000010000000071001161119791100001002003820038200382003820038

1000 unrolls and 10 iterations

Result (median cycles for code): 2.0037

retire (01)cycle (02)03080b18191e1f3f4e51inst 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
100242018015000000061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000000006402162219785010000102003820038200382003820038
100242003715000000061196872510010101000010100005028476801200182003720037184443187671001020100002020000200372003711100211091010100001000000006402162219785010000102003820038200382003820038
100242003715000000061196872510010101000010100005028476801200182003720037184443187671001020100002020000200372003711100211091010100001000000006402162219785010000102003820038200382003820038
10024200371500000026461196872510010101000010100005028476801200182003720037184443187671001020100002020000200372003711100211091010100001000000006402163219785010000102003820038200382003820038
100242003715000000061196872510010101000010100005028476801200182003720037184443187671001020100002020000200372003711100211091010100001000000006402162219785010000102003820038200382003820038
1002420037150000012061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000000006402162219785010000102003820038200382003820038
100242003714900000061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000000006402162219785010000102003820038200382003820038
100242003715000000061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000000006402162219785010000102003820038200382003820038
100242003715000000061196872510010101000010100005028476800200182003720037184443187671001020100002020000200372003711100211091010100001000000306402162219785010000102003820038200382003820038
1002420037150000018061196872510010101000010100005028476801200182003720037184443187671001020100002020000200372003711100211091010100001000000006402162219785010000102003820038200382003820038

Test 4: throughput

Count: 8

Code:

  trn1 v0.2d, v8.2d, v9.2d
  trn1 v1.2d, v8.2d, v9.2d
  trn1 v2.2d, v8.2d, v9.2d
  trn1 v3.2d, v8.2d, v9.2d
  trn1 v4.2d, v8.2d, v9.2d
  trn1 v5.2d, v8.2d, v9.2d
  trn1 v6.2d, v8.2d, v9.2d
  trn1 v7.2d, v8.2d, v9.2d
  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)03070a0b18191e1f3f51inst 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)9fa1a6a8a9acc2c5cfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
80204201601500001000402580100100800001008000050064000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511031633200350800001002003920039200392003920039
80204200381500000000402580100100800001008000050064000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511031632200350800001002003920039200392003920039
80204200381500000000402580100100800001008000050064000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511031623200350800001002003920039200392003920039
80204200381500000000402580100100800001008000050064000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000300511031633200350800001002003920039200392003920039
80204200381500000000402580100100800001008000050064000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511041623200350800001002003920039200392003920039
80204200381500000000402580100100800001008000050064000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511031632200350800001002003920039200392003920039
8020420038150000000012425801001008000010080000500640000200192003820038997339996801002008000020016000020038200381180201100991001008000010020020005110226252003511800001002003920039200392003920039
8020420038150000002408402580100111800001248000050064000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511041622200350800001002003920039200392003920039
80204200381500000000402580100100800001008000050064000020019200382003899733999680100200800002001600002003820038118020110099100100800001000010000511031623200350800001002003920039200392003920039
80204200381500000000402580100100800001008000050064000020019200382003899733999680100200800002001600002003820038118020110099100100800001000000000511031633200350800001002003920039200392003920039

1000 unrolls and 10 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)9fa8accfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
800242004715001212580010108000010800005064000020019200382003899968100448001020800002016000020038200381180021109101080000100050201216222003580000102003920039200392003920039
8002420038150039258001010800001080000506400002001920038200389996310018800102080000201600002003820038118002110910108000010005020616622003580000102003920039200392003920039
80024200381501239258001010800001080000506400002001920038200389996310018800102080000201600002003820038118002110910108000010005020216222003580000102003920039200392003920039
8002420038150039258001010800001080000506400002001920038200389996310018800102080000201600002003820038118002110910108000010005020616222003580000102003920039200392009820039
8002420038150039258001010800001080000506400002001920038200389996310018800102080000201600002003820038118002110910108000010005020216222003580000102003920039200392003920039
80024200381500704258001010800001080000506400002001920038200389996310018800102080000201600002003820038118002110910108000010005020216222003580000102003920039200392003920039
8002420038150039258001010800001080000506400002001920038200389996310018800102080000201600002003820038118002110910108000010005020216362003580000102003920039200392003920039
8002420038150039258001010800001080000506400002001920038200389996310018800102080000201600002003820038118002110910108000010005020216222003580000102003920039200392003920039
8002420038150039258001010800001080000506400002001920038200389996310018800102080000201600002003820038118002110910108000010005020616232003580000102003920039200392003920039
8002420038150039258001010800001080000506400002001920038200389996310018800102080000201600002003820038118002110910108000010005020216632003580000102003920039200392003920039