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

DMB (OSH)

Test 1: uops

Code:

  dmb osh

(no loop instructions)

1000 unrolls and 1 iteration

Retires: 1.000

Issues: 1.000

Integer unit issues: 0.001

Load/store unit issues: 1.000

SIMD/FP unit issues: 0.000

Recorded non-zero counters (first 10 of 64 runs)

retire uop (01)	cycle (02)	schedule uop (52)	schedule int uop (53)	schedule ldst uop (55)	dispatch ldst uop (58)	simd uops in schedulers (5a)	dispatch uop (78)	map ldst uop (7d)	? int output thing (e9)	? ldst retires (ed)
1004	4040	1001	1	1000	1000	4000	1000	1000	1	1000
1004	4040	1001	1	1000	1000	4000	1000	1000	1	1000
1004	4040	1001	1	1000	1000	4000	1000	1000	1	1000
1004	4040	1001	1	1000	1000	4000	1000	1000	1	1000
1004	4040	1001	1	1000	1000	4000	1000	1000	1	1000
1004	4040	1001	1	1000	1000	4000	1000	1000	1	1000
1004	4040	1001	1	1000	1000	4000	1000	1000	1	1000
1004	4040	1001	1	1000	1000	4000	1000	1000	1	1000
1004	4040	1001	1	1000	1000	4000	1000	1000	1	1000
1004	4040	1001	1	1000	1000	4000	1000	1000	1	1000

Test 2: throughput

Code:

  dmb osh

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 4.0040

Recorded non-zero counters (first 10 of 64 runs)

retire uop (01)	cycle (02)	schedule uop (52)	schedule int uop (53)	schedule ldst uop (55)	dispatch int uop (56)	dispatch ldst uop (58)	int uops in schedulers (59)	simd uops in schedulers (5a)	dispatch uop (78)	map int uop (7c)	map ldst uop (7d)	map int uop inputs (7f)	? int output thing (e9)	? ldst retires (ed)	? int retires (ef)
10204	40040	10105	101	10004	100	10006	300	40024	10106	200	10006	200	1	10000	100
10204	40040	10105	101	10004	100	10004	300	40016	10104	200	10004	200	1	10000	100
10204	40040	10105	101	10004	100	10012	300	40048	10112	200	10012	200	1	10000	100
10204	40041	10107	101	10006	100	10004	300	40016	10104	200	10004	200	1	10000	100
10204	40040	10105	101	10004	100	10004	300	40016	10104	200	10004	200	1	10000	100
10204	40040	10105	101	10004	100	10004	300	40016	10104	200	10004	200	1	10000	100
10204	40040	10105	101	10004	100	10004	300	40016	10104	200	10004	200	1	10000	100
10204	40040	10105	101	10004	100	9408	71205	66077	13791	6224	9805	200	1	10000	100
10204	40040	10105	101	10004	100	10004	300	40016	10104	200	10004	200	1	10000	100
10204	40040	10105	101	10004	100	10004	300	40016	10104	200	10004	200	1	10000	100

1000 unrolls and 10 iterations

Result (median cycles for code): 4.0040

Recorded non-zero counters (first 10 of 64 runs)

retire uop (01)	cycle (02)	schedule uop (52)	schedule int uop (53)	schedule ldst uop (55)	dispatch int uop (56)	dispatch ldst uop (58)	int uops in schedulers (59)	simd uops in schedulers (5a)	dispatch uop (78)	map int uop (7c)	map ldst uop (7d)	map int uop inputs (7f)	? int output thing (e9)	? ldst retires (ed)	? int retires (ef)
10024	40040	10015	11	10004	10	10006	30	40024	10016	20	10006	20	1	10000	10
10024	40040	10011	11	10000	10	10012	30	40048	10022	20	10012	20	1	10000	10
10024	40040	10011	11	10000	10	10000	30	40000	10010	20	10000	20	1	10000	10
10024	40040	10011	11	10000	10	10000	30	40000	10010	20	10000	20	1	10000	10
10024	40040	10011	11	10000	10	10000	30	40000	10010	20	10000	20	1	10000	10
10025	40080	10023	11	10012	10	10000	30	40000	10010	20	10000	20	1	10000	10
10024	40040	10011	11	10000	10	10000	30	40000	10010	20	10000	20	1	10000	10
10024	40040	10011	11	10000	10	10000	30	40000	10010	20	10000	20	1	10000	10
10024	40040	10011	11	10000	10	10000	30	40000	10010	20	10000	20	1	10000	10
10024	40040	10011	11	10000	10	10000	30	40000	10010	20	10000	20	1	10000	10