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

SCVTF (scalar, fixed-point, D from W)

Test 1: uops

Code:

  scvtf d0, w0, #3
  mov x0, 1
  mov x1, 2

(no loop instructions)

1000 unrolls and 1 iteration

Retires: 2.000

Issues: 2.000

Integer unit issues: 0.001

Load/store unit issues: 1.000

SIMD/FP unit issues: 1.000

retire uop (01)cycle (02)schedule uop (52)schedule int uop (53)schedule simd uop (54)schedule ldst uop (55)dispatch simd uop (57)dispatch ldst uop (58)simd uops in schedulers (5a)ldst uops in schedulers (5b)dispatch uop (78)map ldst uop (7d)map simd uop (7e)map ldst uop inputs (80)map simd uop inputs (81)? int output thing (e9)? ldst retires (ed)? simd retires (ee)
20045762001110001000100010003000800020001000100010001000110001000
20045402001110001000100010003000800020001000100010001000110001000
20045372001110001000100010003000800020001000100010001000110001000
20045372001110001000100010003000800020001000100010001000110001000
20045372001110001000100010003000800020001000100010001000110001000
20045372001110001000100010003000800020001000100010001000110001000
20045372001110001000100010003000800020001000100010001000110001000
20045372001110001000100010003000800020001000100010001000110001000
20045372001110001000100010003000800020001000100010001000110001000
20045372001110001000100010003000800020001000100010001000110001000

Test 2: Latency 1->2 roundtrip

Code:

  scvtf d0, w0, #3
  fmov x0, d0
  mov x0, 1
  mov x1, 2

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 10.0030

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)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 ldst uop inputs (80)map simd uop inputs (81)? int output thing (e9)? ldst retires (ed)? simd retires (ee)? int retires (ef)
3020410003040101101012000010000100200001000130015461762578264301012001000220004200100022000410001100001000010100
3020410003040101101012000010000100200001000030015461742578264301002001000220002200100202003710002100001000010100
3020410003440101101012000010000100200001000030015461742578264301002001000220002200100022000210001100001000010100
3020410003040101101012000010000100200001000030015461742578264301002001000220002200100022000210001100001000010100
3020410003040101101012000010000100200001000030015461742578264301002001000220002200100212003910003100001000010100
3020410003040101101012000010000100200001000030015461742578264301002001000220002200100022000210001100001000010100
3020410003040101101012000010000100200001000030015461742578264301002001000220002200100022000210001100001000010100
3020410003040101101012000010000100200001000030015461742578264301002001000220002200100022000210001100001000010100
3020410003040101101012000010000100200001059632115766422629274316992061063521246215109542187210146100001000010100
3020410424140574102452022110108117213771076833015863502645348321612101082721626218109332183110144100001000010100

1000 unrolls and 10 iterations

Result (median cycles for code): 10.0030

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)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 ldst uop inputs (80)map simd uop inputs (81)? int output thing (e9)? ldst retires (ed)? simd retires (ee)? int retires (ef)
300241000694001110011200001000010200001000030154771225807603001020100002000020100002000010001100001000010010
300241000334001110011200001000010200001000030154617425782643001020100002000020100002000010001100001000010010
300241000304001110011200001000010200001000030154617425782643001020100002000020100002000010001100001000010010
300241000304001110011200001000010200001000030154617425782643001020100002000020100002000010001100001000010010
300241000304001110011200001000010200001000030154617425782643001020100002000020100002000010001100001000010010
300241000304001110011200001000010200001001730154638425786173005620100212003920100002000010001100001000010010
300241000304001110011200001000010200001000030154617425782643001020100002000020100002000010001100001000010010
300241000304001110011200001000010200001000030154617425782643001020100002000020100002000010001100001000010010
300241000304001110011200001000010200001000030154617425782643001020100002000020100002000010001100001000010010
300241000304001110011200001000010200001000030154617425782643001020100002000020100002000010001100001000010010

Test 3: throughput

Count: 8

Code:

  scvtf d0, w8, #3
  scvtf d1, w8, #3
  scvtf d2, w8, #3
  scvtf d3, w8, #3
  scvtf d4, w8, #3
  scvtf d5, w8, #3
  scvtf d6, w8, #3
  scvtf d7, w8, #3
  mov x8, 9

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code divided by count): 0.5011

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)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 ldst uop inputs (80)map simd uop inputs (81)? int output thing (e9)? ldst retires (ed)? simd retires (ee)? int retires (ef)
16020440160160115101800068000810080012800123002400366400801601242008001280012200800128001218000080000100
16020440090160115101800068000810080012800123002400366400801601242008001280012200800128001218000080000100
16020440090160115101800068000810080012800123002400366400801601242008001280012200800128001218000080000100
16020440090160115101800068000810080012800123002400366400801601242008001280012200800128001218000080000100
16020440090160115101800068000810080012800123002400366400801601242008001280012200800128001218000080000100
16020540140160175101800368003810080054800123002400366400801601242008001280012200800128001218000080000100
16020440090160115101800068000810080012800123002400366400801601242008001280012200800128001218000080000100
16020440090160115101800068000810080012800123002400366400801601242008001280012200800128001218000080000100
16020440090160115101800068000810080012800123002400366400801601242008001280012200800128001218000080000100
16020440090160115101800068000810080012800123002400366400801601242008001280012200800128001218000080000100

1000 unrolls and 10 iterations

Result (median cycles for code divided by count): 0.5006

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)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 ldst uop inputs (80)map simd uop inputs (81)? int output thing (e9)? ldst retires (ed)? simd retires (ee)? int retires (ef)
16002440540160025118000680008108001280055302402736404721601202080055800552080012800121800008000010
16002440068160025118000680008108001280000302400006400001600102080000800002080000800001800008000010
16002440044160011118000080000108000080000302400006400001600102080000800002080000800001800008000010
16002440042160011118000080000108000080000302400006400001600102080000800002080000800001800008000010
16002440042160011118000080000108000080000302400006400001600102080000800002080000800001800008000010
16002440042160011118000080000108000080000302400006400001600102080000800002080000800001800008000010
16002440042160011118000080000108000080000302400006400001600102080000800002080000800001800008000010
16002440042160011118000080000108000080000302400006400001600102080000800002080000800001800008000010
16002440042160011118000080000108000080000302400006400001600102080000800002080000800001800008000010
16002440042160011118000080000108000080000302400006400001600102080000800002080000800001800008000010