Apple M1 Microarchitecture Research by Dougall Johnson

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

FMUL (by element, 2D)

Test 1: uops

Code:

  fmul v0.2d, v0.2d, v1.d[1]
  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)a1a6a8a9accfd5d6ddinst fetch restart (de)e0? fp/simd (ee)f5f6f7f8fd
100440373006134072510001000100053190804018403740373258338951000100020004037403711100110000000073116113473100040384038403840384038
1004403730025134072510001000100053190804018403740373258338951000100020004037403711100110000000073116113473100040384038403840384038
100440373006134072510001000100053190804018403740373258338951000100020004037403711100110000000073116113473100040384038403840384038
100440373006134072510001000100053190814018403740373258338951000100020004037403711100110000000073116113473100040384038403840384038
100440373006134072510001000100053190804018403740373258338951000100020004037403711100110000000073116113473100040384038403840384038
100440373006134072510001000100053190804018403740373258338951000100020004037403711100110000010073116113473100040384038403840384038
1004403730014534072510001000100053190814018403740373258338951000100020004037403711100110000000073116113473100040384038403840384038
10044037300128340725100010001000531908040184037403732583389510001000200040374037111001100000220073116113473100040384038403840384038
1004403730306134072510001000100053190814018403740373258338951000100020004037403711100110000210073116113473100040384038403840384038
100440373006134072510001000100053190814018403740373258338951000100020004037403711100110000000073116113473100040384038403840384038

Test 2: Latency 1->2

Code:

  fmul v0.2d, v0.2d, v1.d[1]
  movi v0.16b, 1
  movi v1.16b, 2

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 4.0037

retire (01)cycle (02)0308091e1f3f4e51inst 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)9fa1a8accfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
1020440037300003570613940725101001001000012710000500570690814001840037400373810833874510100200100002002000040037400371110201100991001001000010000071021611394790100001004003840038400384008040038
102044003730000180613940725101001001000010010592500570690814001840037400373810833874510100200100002002000040037400371110201100991001001000010000071011611394790100001004003840038400384003840038
1020440037299012730613940725101001001000010010000500570690804001840037400373810833874510100200100002002000040037400371110201100991001001000010001971011611394790100001004003840038400384003840038
102044003730000930613940725101001001000010010000500570690804015840037400373810833874510100200100002002000040037400371110201100991001001000010001071011611394790100001004008640086400384003840038
1020440037300003060613940725101001001000010010000500570690804001840085400373810833874510100204100002002041240084400371110201100991001001000010000071011611394790100001004003840038400384003840038
10204402303000063521033940725101001001000010010000500570690804001840037400373810833874510100200100002002033240037400371110201100991001001000010000071011611394790100001004003840038400384003840038
10204400373000037201243940725101001001000010010000500570690804001840037400373810833874510100200100002002000040037400371110201100991001001000010000071011611396220100001004003840038400384003840038
102044003730000660103394071031010010010000100100005005706908040018400374003738108193874510100200100002002000040037400371110201100991001001000010000071011611394790100001004003840038400384003840038
10204400373000132701033940725101001001000010010000500570690814001840037400373812233874510100200100002002000040037400371110201100991001001000010000071031611394790100001004003840038400384003840038
10204400843000130180613940725101001001000011010000500570854214005340084400853810833874510253200100002002000040037400371110201100991001001000010000071011611394790100001004003840038400384003840038

1000 unrolls and 10 iterations

Result (median cycles for code): 4.0037

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)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)9fa0a1a6a8acc2c3cfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
100244003730000001800613940725100101010000101000050570690814001804003740037381303387671001020100002020000400374003711100211091010100001000000006402162239473010000104003840038400384003840038
10024400373000000330613940725100101010000101000050570690814001804003740037381303387671001020100002020000400374003711100211091010100001000000006402162239473010000104003840038400384003840038
10024400373000000690613940725100101010000101000050570690814001804003740037381303387671001020100002020000400374003711100211091010100001000000006402162239473010000104003840038400384003840038
1002440037299000027909643940725100101010000101000050570690814001834003740037381303387671001020100002020000400374003711100211091010100001000000006402162239473010000104003840038400384003840038
10024400372990000300613940725100101010000101000050570690814001804003740037381303387671001020100002020000400374003711100211091010100001000000006402162239473010000104003840038400384003840038
100244003730000003390613940725100101010000101000050570690814001804003740037381303387671001020100002020000400374003711100211091010100001000000006402162239473010000104003840038400384003840038
10024400373000000690613940725100101010000101000050570690814001804003740037381303387671001020100002020000400374003711100211091010100001000000006402162239473010000104003840038400384003840038
10024400373000000300843940725100101010000101000050570690814001804003740037381303387671001020100002020000400374003711100211091010100001000000006402162239473010000104003840038400384003840038
10024400373000000210613940725100101010000101000050570690814001804003740037381303387671001020100002020000400374003711100211091010100001000000006402162239473010000104003840038400384003840038
10024400372990000240613940725100101010000101000050570690814001804003740037381303387671001020101632020324400374003711100211091010100001000000006402162239473010000104003840038400384003840038

Test 3: Latency 1->3

Code:

  fmul v0.2d, v1.2d, v0.d[1]
  movi v0.16b, 1
  movi v1.16b, 2

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 4.0037

retire (01)cycle (02)03080b1e3f4e51inst 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? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
102044003729900186139407251010010010000100100005005706908040018400374003738108338745101002001000020020000400374003711102011009910010010000100000071011611394790100001004003840038400384003840038
10204400373000006139407251010010010000100100005005706908140018400374003738108338745101002001000020020000400374003711102011009910010010000100000071011611394790100001004003840038400384003840038
10204400372990006139407251010010010000100100005005706908040018400374003738108338745101002001000020020000400374003711102011009910010010000100000071011611394790100001004003840038400384003840038
10204400373000006139407251010010010000100100005005706908140018400374003738108338745101002001000020020000400374003711102011009910010010000100000071011611394790100001004003840038400384003840038
10204400373000006139407251010010010000100100005005706908140018400374003738108338745101002001000020020000400374003711102011009910010010000100000171011611394790100001004003840038400384003840038
10204400373000006139407251010010010000100100005005706908140018400374003738108338745101002001000020020000400374003711102011009910010010000100000071011611394790100001004003840038400384003840038
10204400373000006139407251010010010000100100005005706908140018400374018038108338745101002001000020020000400374003711102011009910010010000100000071011611394790100001004003840038400384003840038
10204400373000006139407251010010010000100100005005706908140018400374003738108338745101002001000020020000400374003711102011009910010010000100000071011611394790100001004003840038400384003840038
102044003729900396139407251010010010000100100005005706908140018400374003738108338745101002001000020020000400374003711102011009910010010000100000071011611394790100001004003840038400384003840038
10204400373000006139407251010010010000100100005005706908140018400374003738108338745101002001000020020000400374003711102011009910010010000100000071011611394790100001004003840038400384003840038

1000 unrolls and 10 iterations

Result (median cycles for code): 4.0037

retire (01)cycle (02)0309191e1f3f4e5051inst 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)9fa8acc2cfd5d6ddinst fetch restart (de)e0? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
10024400372990060613940702510010101000010100005057069084001840037400373813033876710010201000020200004003740037111002110910101000010000640316223947310000104003840038400384003840038
10024400372990000613940702510010101000010100005057069084001840037400373813033876710010201000020200004003740037111002110910101000010000640216223947310000104003840038400384003840038
10024400373000000613940702510010101000010100005057069084001840037400373813033876710010201000020200004003740037111002110910101000010102640216223947310000104003840038400384003840038
10024400373000000613940702510010101000010100005057069084001840037400373813033876710010201000020200004003740037111002110910101000010000640216223947310000104003840038400384003840038
10024400373000000613940702510010101000010100005057069084001840037400373813033876710010201000020200004003740037111002110910101000010000640216223947310000104003840038400384003840038
10024400373000000613940702510010101000010100005057069084001840037400373813033876710010201000020200004003740037111002110910101000010000640216223947310000104003840038400384003840038
10024400373000000613940702510010101000010100005057069084001840037400373813033876710010201000020200004003740037111002110910101000010000640216223947310000104003840038400384003840038
10024400373000000613940702510010101000010100005057069084001840037400373813033876710010201000020200004003740037111002110910101000010000640216223947310000104003840038400384003840038
10024400773000000613940702510010101000010100005057069084001840037400373813033876710010201000020200004003740037111002110910101000010000640216223947310000104003840038400384003840038
10024400372990000613940702510010101000010100005057069084001840037400373813033876710010201000020200004003740037111002110910101000010000640216223947310000104003840038400384003840038

Test 4: throughput

Count: 8

Code:

  fmul v0.2d, v8.2d, v9.d[1]
  fmul v1.2d, v8.2d, v9.d[1]
  fmul v2.2d, v8.2d, v9.d[1]
  fmul v3.2d, v8.2d, v9.d[1]
  fmul v4.2d, v8.2d, v9.d[1]
  fmul v5.2d, v8.2d, v9.d[1]
  fmul v6.2d, v8.2d, v9.d[1]
  fmul v7.2d, v8.2d, v9.d[1]
  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)03080b18191e1f3f51inst 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)696b6d6edispatch 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
80204200591500000004225801001008000010080000500640000200210200402004099733999880100200800002001600002004020040118020110099100100800001000000000511021611200370800001002004120041200412004120041
80204200401500000004225801001008000010080000500640000200210200402004099733999880100200800002001600002004020040118020110099100100800001000000090511011611200370800001002004120041200412004120041
80204200401500000004225801001008000010080000500640000200210200402004099733999880100200800002001600002004020040118020110099100100800001000000000511011611200370800001002004120041200412004120041
80204200401500000004225801001008000010080000500640000200210200402004099738999880100200800002001600002004020040118020110099100100800001000000000511011611200370800001002004120041200412004120105
80204200401500000004225801001008000010080000500640000200213200402004099733999880100200800002001600002004020040118020110099100100800001000000000511011611200370800001002004120041200412004120041
80204200401500000004225801001008000010080000500640000200210200402004099733999880100200800002001600002004020040118020110099100100800001000000000511011611200370800001002004120041200412004120041
8020420040150000000422580100100800001008000050064000020021020040200409973399988010020080000200160000200402004011802011009910010080000100000101380511011611200370800001002004120041200412004120041
80204200401500000004225801001008000010080000500640000200210200402004099733999880100200800002001600002004020040118020110099100100800001000000000511011611200370800001002004120041200412004120041
802042004015000003604225801001008000010080000500643320200210200402004099733999880100200800002001600002004020040118020110099100100800001000000000511011611200370800001002004120041200412004120041
80204200401500000004225801001008000010080000500640000200210200402004099733999880100200800002001600002004020040118020110099100100800001000000000511011611200370800001002004120041200412004120041

1000 unrolls and 10 iterations

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

retire (01)cycle (02)03181e3f51inst 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)9fa7a8accdcfd5d6ddinst fetch restart (de)e0? int output thing (e9)? fp/simd (ee)gpr retires (ef)f5f6f7f8fd
8002420049150096244800101080000108000050640000020021020201200409996310020800102080000201600002004020040218002110910108000010000050204162420037080000102004120041200412004120041
8002420040150004125800101080000108000050640000020021020040200409996310020800102080000201600002004020040118002110910108000010000050204162420037080000102004120041200412004120041
80024200401500041258001010800001080000506400001200210200402004099963100208001020800002016000020040200401180021109101080000100343050202164420037080000102004120041200412004120041
80024200401500041258001010800001080000506400001200210200402004099963100208001020800002016000020040200401180021109101080000100270150202162420037080000102004120041200412004120041
8002420040150004125800101080000108000050640000120021020146200409996310020800102080000201600002004020040118002110910108000010000050204164220037080000102004120041200412011420041
8002420040150008225800101080000108000050640000020021020040200409996310020800102080000201600002004020040118002110910108000010010050202162420037080000102004120041200412004120041
8002420040150004125800101080000108000050640000020021020040200409996310020800102080000201600002004020040118002110910108000010003050204164220037080000102004120041200412004120041
8002420040150004125800101080000108000050640000020021020040200409996310020800102080000201600002004020040118002110910108000010000050202162420037080000102004120041200412004120041
80024200401500083258001010800001080000506400000200210200402004099967100208001020800002016020820040200401180021109101080000100018050202162420037080000102004120041200412004420041
8002420040150115125258001010800001080314506400001200210200402004099963100208001020800002016000020040200401180021109101080000100240050202164420037080000102004120041200412004120041