Apple Microarchitecture Research by Dougall Johnson

M1/A14 P-core (Firestorm): Overview | Base Instructions | SIMD and FP Instructions
M1/A14 E-core (Icestorm):  Overview | Base Instructions | SIMD and FP Instructions

FCVTMU (scalar, H to H)

Test 1: uops

Code:

  fcvtmu h0, h0
  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: 0.001

Load/store unit issues: 0.000

SIMD/FP unit issues: 1.000

retire uop (01)cycle (02)schedule uop (52)schedule int uop (53)schedule simd uop (54)dispatch simd uop (57)ldst uops in schedulers (5b)dispatch uop (78)map simd uop (7e)map simd uop inputs (81)? int output thing (e9)? simd retires (ee)
1004303310011100010007590510001000100011000
1004303310011100010007590510001000100011000
1004303310011100010007590510001000100011000
1004303310011100010007590510001000100011000
1004303310011100010007590510001000100011000
1004303310011100010007590510001000100011000
1004303310011100010007590510001000100011000
1004303310011100010007590510001000100011000
1004303310011100010007590510001000100011000
1004303310011100010007590510001000100011000

Test 2: Latency 1->2

Code:

  fcvtmu h0, h0
  movi v0.16b, 1
  movi v1.16b, 2

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 3.0033

retire uop (01)cycle (02)schedule uop (52)schedule int uop (53)schedule simd uop (54)dispatch int uop (56)dispatch simd uop (57)int uops in schedulers (59)ldst uops in schedulers (5b)dispatch uop (78)map int uop (7c)map simd uop (7e)map int uop inputs (7f)map simd uop inputs (81)? int output thing (e9)? simd retires (ee)? int retires (ef)
1020430033101011011000010010000300768905101002001000420010006110000100
1020430033101011011000010010000300768905101002001000420010004110000100
1020430033101011011000010010000300768905101002001000420010004110000100
1020430033101011011000010010000300768905101002001000420010004110000100
1020430033101011011000010010000300768905101002001000420010004110000100
1020430033101011011000010010000300768905101002001000420010004110000100
1020430033101011011000010010000300768905101002001000420010004110000100
1020430033101011011000010010000300768905101002001000420010004110000100
1020430033101011011000010010000300768905101002001000420010004110000100
1020430033101011011000010010000300768905101002001000420010004110000100

1000 unrolls and 10 iterations

Result (median cycles for code): 3.0033

retire uop (01)cycle (02)schedule uop (52)schedule int uop (53)schedule simd uop (54)dispatch int uop (56)dispatch simd uop (57)int uops in schedulers (59)ldst uops in schedulers (5b)dispatch uop (78)map int uop (7c)map simd uop (7e)map int uop inputs (7f)map simd uop inputs (81)? int output thing (e9)? simd retires (ee)? int retires (ef)
10024300331002121100002010000707689051002020100062010000111000010
10024300331002121100002010000707689051002020100002010000111000010
10024300331002121100002010000707689051002020100002010000111000010
10024300331002121100002010000707689051002020100042010006111000010
10024300841002921100082010033707689051002020100002010000111000010
10024300331002121100002010000697699971008620100812010000111000010
10024300331002121100002010000707689051002020100002010000111000010
10024300331002121100002010000707689051002020100002010000111000010
10024300331002121100002010000707689051002020100002010000111000010
10024301341003721100162010066707689051002020100002010000111000010

Test 3: throughput

Count: 8

Code:

  fcvtmu h0, h8
  fcvtmu h1, h8
  fcvtmu h2, h8
  fcvtmu h3, h8
  fcvtmu h4, h8
  fcvtmu h5, h8
  fcvtmu h6, h8
  fcvtmu h7, h8
  movi v8.16b, 9

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code divided by count): 0.5004

retire uop (01)cycle (02)schedule uop (52)schedule int uop (53)schedule simd uop (54)schedule ldst uop (55)dispatch int uop (56)dispatch simd uop (57)int uops in schedulers (59)ldst uops in schedulers (5b)dispatch uop (78)map int uop (7c)map simd uop (7e)map int uop inputs (7f)map simd uop inputs (81)? int output thing (e9)? simd retires (ee)? int retires (ef)
80204400558010510180004010080008300320044801102008001420080012180000100
80204400358010510180004010080008300320036801082008001220080012180000100
80204400358010510180004010080008300320036801082008001220080012180000100
80204400358010510180004010080008300320036801082008001220080012180000100
80204400358010510180004010080008300320036801082008001220080012180000100
80204400358010510180004010080008300320036801082008001220080012180000100
80204400358010510180004010080008300320036801082008001220080012180000100
80204400358010510180004010080008300320036801082008001220080012180000100
80204400358010510180004010080008300320036801082008001220080012180000100
80204400358010510180004010080008300320036801082008001220080012180000100

1000 unrolls and 10 iterations

Result (median cycles for code divided by count): 0.5004

retire uop (01)cycle (02)schedule uop (52)schedule int uop (53)schedule simd uop (54)dispatch int uop (56)dispatch simd uop (57)dispatch ldst uop (58)int uops in schedulers (59)simd uops in schedulers (5a)ldst uops in schedulers (5b)dispatch uop (78)map int uop (7c)map ldst uop (7d)map simd uop (7e)map int uop inputs (7f)map simd uop inputs (81)? int output thing (e9)? simd retires (ee)? int retires (ef)
80024401928002721800062080010070032004480030200800162080000118000010
80024400358002121800002080000070032078480216200801962080000118000010
80024400358002121800002080000070032000080020200800002080000118000010
80024400358002121800002080000070032000080020200800002080000118000010
80024400358002121800002080000070032000080020200800002080200118000010
80024400358002121800002080000070032000080020200800002080000118000010
80025400718006321800422080054140587139267153251428361828171722801032080000118000010
80024400358002121800002080000070032000080020200800002080000118000010
80024400358002121800002080000070032039680119200800992080000118000010
80024400358002121800002080000070032000080020200800002080000118000010