Firestorm: SBFX (32-bit) measurements

Code:

  sbfx w0, w0, #3, #7

Initialization

  mov x0, 1
  mov x1, 2

(no loop instructions)

1000 unrolls and 1 iteration

Retires: 1.000

Issues: 1.000

Integer unit issues: 1.000

Load/store unit issues: 0.000

SIMD/FP unit issues: 0.000

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

retire uop (01)	cycle (02)	03	mmu table walk instruction (07)	1e	3f	4d	51	schedule uop (52)	schedule int uop (53)	dispatch int uop (56)	int uops in schedulers (59)	60	6d	6e	map stall dispatch (70)	map rewind (75)	map stall (76)	dispatch uop (78)	map int uop (7c)	map int uop inputs (7f)	82	83	flush restart other nonspec (84)	85	inst all (8c)	inst int alu (97)	l1d cache writeback (a8)	ac	cf	d5	map dispatch bubble (d6)	dd	fetch restart (de)	e0	? int output thing (e9)	? int retires (ef)	f5	f6	f7	f8	fd
1004	1035	8	0	0	61	862	25	1000	1000	1000	16916	1	1035	1035	728	3	868	1000	1000	1000	1035	41	1	1	1001	1000	0	0	75	2	39	2	2	937	1000	1000	1036	1036	1036	1036	1036
1004	1035	7	0	0	61	862	25	1000	1000	1000	16916	0	1035	1035	728	3	870	1000	1000	1000	1035	41	1	1	1001	1000	0	0	73	1	41	1	1	937	1000	1000	1036	1036	1036	1036	1036
1004	1035	7	0	0	61	862	25	1000	1000	1000	16916	0	1035	1035	728	3	868	1000	1000	1000	1035	41	1	1	1001	1000	0	0	73	1	39	1	1	937	1000	1000	1036	1036	1036	1036	1036
1004	1035	8	1	0	61	862	25	1000	1000	1000	16916	1	1035	1035	728	3	868	1000	1000	1000	1035	43	1	1	1001	1000	0	0	73	1	41	1	1	937	1000	1000	1036	1036	1036	1036	1036
1004	1035	8	0	0	61	862	25	1000	1000	1000	16916	1	1035	1035	728	3	868	1000	1000	1000	1035	41	1	1	1001	1000	0	0	73	1	41	1	1	937	1000	1000	1036	1036	1036	1036	1036
1004	1035	8	0	0	61	862	25	1000	1000	1000	16911	1	1035	1035	728	3	868	1000	1000	1000	1035	41	1	1	1001	1000	0	0	73	1	41	1	1	937	1000	1000	1036	1036	1036	1036	1036
1004	1035	8	0	0	75	863	25	1000	1000	1000	16916	1	1035	1035	728	3	870	1000	1000	1000	1035	43	1	1	1001	1000	0	0	73	1	41	1	1	937	1000	1000	1036	1036	1036	1036	1036
1004	1035	8	1	0	157	863	25	1000	1000	1000	16916	1	1035	1035	728	3	868	1000	1000	1000	1035	41	1	1	1001	1000	0	0	73	1	41	1	1	937	1000	1000	1036	1036	1036	1036	1036
1004	1035	9	0	0	61	862	25	1000	1000	1000	16916	1	1035	1035	728	3	870	1000	1000	1000	1035	43	1	1	1001	1000	0	6	73	1	41	1	1	937	1000	1000	1036	1036	1036	1036	1036
1004	1035	7	0	0	61	862	25	1000	1000	1000	16916	1	1035	1035	729	3	868	1000	1000	1000	1035	41	1	1	1001	1000	0	0	73	1	41	1	1	937	1000	1000	1036	1036	1036	1036	1036

Test 2: Latency 1->2

Code:

  sbfx w0, w0, #3, #7

Initialization

  mov x0, 1
  mov x1, 2

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 1.0035

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

retire uop (01)	cycle (02)	03	l2 tlb miss data (0b)	3f	4d	51	schedule uop (52)	schedule int uop (53)	dispatch int uop (56)	int uops in schedulers (59)	60	69	6a	6d	6e	map stall dispatch (70)	map rewind (75)	map stall (76)	dispatch uop (78)	map int uop (7c)	map int uop inputs (7f)	82	83	flush restart other nonspec (84)	85	inst all (8c)	inst branch (8d)	inst branch taken (90)	inst branch cond (94)	inst int alu (97)	9f	l1d tlb access (a0)	l1d cache writeback (a8)	ac	cf	d5	map dispatch bubble (d6)	dd	fetch restart (de)	e0	? int output thing (e9)	? int retires (ef)	f5	f6	f7	f8	fd
10204	10035	75	0	61	9877	25	10100	10100	10100	88664	0	49	6955	10035	10035	8580	3	8722	10100	10200	10200	10035	41	1	1	10201	100	99	100	10100	100	0	0	0	710	1	37	1	1	9941	10000	10100	10036	10036	10036	10036	10036
10204	10035	75	0	671	9877	25	10100	10100	10100	88664	0	49	6955	10035	10035	8580	3	8722	10100	10200	10200	10035	41	1	1	10201	100	99	100	10100	100	0	0	0	710	1	37	1	1	9941	10000	10100	10036	10036	10036	10036	10036
10204	10035	75	0	61	9877	25	10100	10100	10100	88664	1	49	6955	10035	10035	8580	3	8722	10100	10200	10200	10035	41	1	1	10201	100	99	100	10100	100	0	0	0	710	1	37	1	1	9941	10000	10100	10036	10036	10036	10036	10036
10204	10035	75	0	61	9877	25	10100	10100	10100	88664	0	49	6955	10035	10035	8580	3	8722	10100	10200	10200	10035	41	1	1	10201	100	99	100	10100	100	0	0	0	710	1	37	1	1	9941	10000	10100	10036	10036	10036	10036	10036
10204	10035	75	0	807	9877	25	10100	10100	10100	88664	0	49	6955	10035	10035	8580	3	8722	10100	10200	10200	10035	41	1	1	10201	100	99	100	10100	100	0	0	0	710	1	37	1	1	9941	10000	10100	10036	10036	10036	10036	10036
10204	10035	75	0	82	9877	25	10100	10100	10100	88664	0	49	6955	10035	10035	8580	3	8722	10100	10200	10200	10035	41	1	1	10201	100	99	100	10100	100	1	0	0	710	1	37	1	1	9941	10000	10100	10036	10036	10036	10036	10036
10204	10035	75	0	628	9877	25	10100	10100	10100	88664	0	49	6955	10035	10035	8580	3	8722	10100	10200	10200	10035	41	1	1	10201	100	99	100	10100	100	0	0	0	710	1	37	1	1	9941	10000	10100	10036	10036	10036	10036	10036
10204	10035	75	0	61	9877	25	10100	10100	10100	88664	0	49	6955	10035	10035	8580	3	8722	10100	10200	10200	10035	41	1	1	10201	100	99	100	10100	100	0	0	0	710	1	37	1	1	9941	10000	10100	10036	10036	10036	10036	10036
10204	10035	75	0	82	9877	25	10100	10100	10100	88664	0	49	6955	10035	10035	8580	3	8722	10100	10200	10200	10035	41	1	1	10201	100	99	100	10100	100	0	0	0	710	1	37	1	1	9941	10000	10100	10036	10036	10036	10036	10036
10204	10035	75	0	71	9877	25	10100	10100	10100	88664	0	49	6955	10035	10035	8580	3	8722	10100	10200	10200	10035	41	1	1	10201	100	99	100	10100	100	0	0	0	710	1	37	1	1	9941	10000	10100	10036	10036	10036	10036	10036

1000 unrolls and 10 iterations

Result (median cycles for code): 1.0035

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

retire uop (01)	cycle (02)	03	mmu table walk data (08)	1e	3f	4d	51	schedule uop (52)	schedule int uop (53)	dispatch int uop (56)	int uops in schedulers (59)	60	69	6a	6d	6e	map stall dispatch (70)	map rewind (75)	map stall (76)	dispatch uop (78)	map int uop (7c)	map int uop inputs (7f)	82	83	flush restart other nonspec (84)	85	inst all (8c)	inst branch (8d)	inst branch taken (90)	inst branch cond (94)	inst int alu (97)	9f	l1d tlb miss (a1)	l1d cache writeback (a8)	a9	ac	cf	d5	map dispatch bubble (d6)	dd	fetch restart (de)	e0	? int output thing (e9)	? int retires (ef)	f5	f6	f7	f8	fd
10024	10035	75	0	39	103	9863	25	10010	10010	10010	88784	0	49	6955	10035	10035	8602	3	8740	10010	10020	10020	10035	41	1	1	10021	10	9	10	10010	10	0	21	0	0	640	2	41	2	2	9940	10000	10010	10036	10036	10036	10036	10036
10024	10035	75	0	0	105	9863	25	10010	10010	10010	88784	0	49	6955	10035	10035	8602	3	8740	10010	10020	10020	10035	41	1	1	10021	10	9	10	10010	10	0	0	0	2635	640	2	41	2	2	9940	10000	10010	10036	10036	10036	10036	10036
10024	10035	75	0	0	61	9863	25	10010	10010	10010	88784	0	49	6955	10035	10035	8602	3	8740	10010	10196	10191	10035	41	1	1	10021	10	9	10	10010	10	0	0	0	0	640	2	41	2	2	9940	10000	10010	10036	10036	10036	10036	10036
10024	10035	75	0	0	61	9863	25	10010	10010	10010	88784	1	49	6955	10035	10035	8602	3	8740	10010	10020	10020	10035	41	1	1	10021	10	9	10	10010	10	0	0	0	0	640	2	41	2	2	9940	10000	10010	10036	10036	10036	10036	10036
10024	10035	75	0	0	61	9863	25	10010	10010	10010	88784	1	49	6955	10035	10035	8602	3	8740	10010	10020	10020	10035	41	1	1	10021	10	9	10	10010	10	0	0	0	0	640	2	41	2	2	9940	10000	10010	10036	10036	10036	10036	10036
10024	10035	75	0	0	61	9863	25	10010	10010	10010	88784	1	49	6955	10035	10035	8602	3	8740	10010	10020	10020	10035	41	1	1	10021	10	9	10	10010	10	0	0	0	0	640	2	41	2	2	9940	10000	10010	10036	10036	10036	10036	10036
10024	10035	75	0	0	61	9863	25	10010	10010	10010	88784	1	49	6955	10035	10035	8602	3	8740	10010	10020	10020	10035	41	1	1	10021	10	9	10	10010	10	0	1	0	0	640	2	41	2	2	9940	10000	10010	10036	10036	10036	10036	10036
10024	10035	75	0	0	84	9863	25	10010	10010	10010	88784	1	49	6955	10035	10035	8602	3	8740	10010	10020	10020	10035	41	1	1	10021	10	9	10	10010	10	0	0	0	0	640	2	41	2	2	9940	10000	10010	10036	10036	10036	10036	10036
10024	10035	75	0	0	61	9863	25	10010	10010	10010	88784	1	49	6955	10035	10035	8602	3	8740	10010	10020	10020	10035	41	1	1	10021	10	9	10	10010	10	0	0	0	0	640	2	41	2	2	9940	10000	10010	10036	10036	10036	10036	10036
10024	10035	75	0	0	61	9863	25	10010	10010	10010	88784	1	49	6955	10035	10035	8602	3	8740	10010	10020	10020	10035	41	1	1	10021	10	9	10	10010	10	0	0	0	0	640	2	41	2	2	9940	10000	10010	10036	10036	10036	10036	10036

Test 3: throughput

Count: 8

Code:

  sbfx w0, w8, #3, #7
  sbfx w1, w8, #3, #7
  sbfx w2, w8, #3, #7
  sbfx w3, w8, #3, #7
  sbfx w4, w8, #3, #7
  sbfx w5, w8, #3, #7
  sbfx w6, w8, #3, #7
  sbfx w7, w8, #3, #7

Initialization

  mov x8, 9

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code divided by count): 0.1675

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

retire uop (01)	cycle (02)	03	1e	3f	51	schedule uop (52)	schedule int uop (53)	dispatch int uop (56)	int uops in schedulers (59)	69	6a	6d	6e	map stall dispatch (70)	int prf full (71)	map rewind (75)	map stall (76)	dispatch uop (78)	map int uop (7c)	map int uop inputs (7f)	82	83	flush restart other nonspec (84)	85	inst all (8c)	inst branch (8d)	inst branch taken (90)	inst branch cond (94)	inst int alu (97)	9f	l1d tlb access (a0)	l1d tlb miss (a1)	l1d cache writeback (a8)	ac	branch cond mispred nonspec (c5)	branch mispred nonspec (cb)	cd	cf	d5	map dispatch bubble (d6)	dd	fetch restart (de)	e0	? int output thing (e9)	? int retires (ef)	f5	f6	f7	f8	fd
80204	13413	100	0	93	27	80136	80136	80148	400710	49	10310	13390	13390	3326	0	6	3336	80148	80264	80264	13390	39	1	1	80201	100	99	100	80100	100	0	0	3	1373	1	1	1	5119	1	16	2	3	13387	80036	80100	13391	13391	13517	13391	13391
80204	13390	101	0	28	27	80136	80136	80148	400710	49	10310	13390	13390	3326	0	6	3336	80148	80264	80264	13390	39	1	1	80201	100	99	100	80100	100	0	0	0	0	1	1	1	5120	1	16	2	2	13387	80036	80100	13391	13391	13391	13391	13391
80204	13390	100	0	70	27	80136	80136	80148	400710	49	10310	13390	13390	3326	0	6	3336	80148	80264	80264	13390	39	1	1	80201	100	99	100	80100	100	0	0	4	0	1	1	1	5120	1	16	1	2	13387	80036	80100	13391	13391	13391	13391	13391
80204	13390	100	0	28	27	80136	80136	80148	400710	49	10310	13390	13390	3326	0	6	3336	80148	80264	80264	13390	39	1	1	80201	100	99	100	80100	100	0	0	0	0	1	1	1	5119	0	16	0	0	13440	80036	80100	13391	13391	13391	13391	13391
80204	13390	100	0	28	27	80136	80136	80148	400710	49	10310	13390	13390	3326	0	6	3336	80148	80264	80264	13390	39	1	1	80201	100	99	100	80100	100	2	0	0	0	1	1	1	5120	0	16	2	2	13387	80036	80100	13391	13391	13391	13391	13391
80204	13390	101	0	28	27	80136	80136	80148	400710	49	10310	13390	13390	3326	0	6	3336	80148	80264	80264	13390	39	1	1	80201	100	99	100	80100	100	0	0	0	0	1	1	1	5120	0	16	2	1	13387	80036	80100	13391	13391	13391	13391	13391
80204	13390	100	0	28	27	80136	80136	80148	400710	49	10310	13390	13390	3326	0	6	3336	80148	80264	80264	13390	39	1	1	80201	100	99	100	80100	100	0	0	0	0	1	1	1	5120	0	16	2	2	13387	80036	80100	13391	13391	13391	13391	13391
80204	13390	100	0	28	27	80136	80136	80148	400710	49	10310	13390	13390	3326	0	6	3336	80148	80264	80264	13390	39	1	1	80201	100	99	100	80100	100	0	0	0	0	1	1	1	5120	0	16	2	2	13387	80036	80100	13391	13391	13391	13391	13391
80204	13390	100	0	49	27	80136	80136	80408	400710	49	10310	13390	13390	3326	0	6	3336	80148	80264	80264	13390	39	1	1	80201	100	99	100	80100	100	0	0	0	0	1	1	1	5120	0	16	2	2	13387	80036	80100	13391	13391	13391	13391	13391
80204	13390	100	0	28	27	80136	80136	80148	400710	49	10310	13390	13390	3326	0	6	3336	80148	80264	80264	13390	39	1	1	80201	100	99	100	80100	100	0	0	0	0	1	1	1	5120	0	16	1	2	13387	80036	80100	13391	13391	13391	13391	13391

1000 unrolls and 10 iterations

Result (median cycles for code divided by count): 0.1671

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

retire uop (01)	cycle (02)	03	mmu table walk instruction (07)	l2 tlb miss instruction (0a)	1e	3a	3f	51	schedule uop (52)	schedule int uop (53)	dispatch int uop (56)	int uops in schedulers (59)	60	67	69	6a	6d	6e	map stall dispatch (70)	map rewind (75)	map stall (76)	dispatch uop (78)	map int uop (7c)	map int uop inputs (7f)	82	83	flush restart other nonspec (84)	85	inst all (8c)	inst branch (8d)	inst branch taken (90)	91	inst branch cond (94)	inst int alu (97)	9f	ac	cf	d5	map dispatch bubble (d6)	dd	fetch restart (de)	e0	? int output thing (e9)	? int retires (ef)	f5	f6	f7	f8	fd
80024	13390	100	1	1	237	3	45	25	80010	80010	80010	400050	1	0	49	10291	13371	13371	3330	3	3348	80010	80020	80020	13371	39	1	1	80021	10	9	13355	10	80010	10	0	5030	20	19	17	17	13368	80000	80010	13372	13372	13372	13372	13372
80024	13371	100	1	1	0	3	110	25	80010	80010	80010	400050	0	0	49	10291	13371	13371	3330	3	3348	80010	80020	80020	13371	39	1	1	80021	10	9	13355	10	80010	10	0	5030	16	19	17	19	13368	80000	80010	13372	13372	13372	13372	13372
80024	13371	100	1	1	27	3	45	25	80010	80010	80010	400050	0	0	49	10291	13371	13371	3330	3	3348	80010	80020	80020	13371	39	1	1	80021	10	9	0	10	80010	10	0	5030	16	19	17	15	13368	80000	80010	13372	13372	13372	13372	13372
80024	13371	100	1	1	237	3	45	25	80010	80010	80010	400050	1	0	49	10291	13371	13371	3330	3	3348	80010	80020	80020	13371	39	1	1	80021	10	9	0	10	80010	10	0	5030	17	19	15	17	13368	80000	80010	13372	13372	13372	13372	13372
80024	13371	100	1	1	0	3	45	25	80010	80010	80010	400050	0	0	49	10291	13371	13371	3330	3	3348	80010	80020	80020	13371	39	1	1	80021	10	9	0	10	80010	10	0	5030	16	19	17	17	13368	80000	80010	13372	13372	13372	13372	13372
80024	13371	100	1	1	195	3	45	25	80010	80010	80010	400050	1	0	49	10291	13371	13371	3330	3	3348	80010	80020	80020	13371	39	1	1	80021	10	9	0	10	80010	10	0	5029	15	19	17	19	13368	80000	80010	13372	13372	13372	13372	13372
80024	13371	100	1	1	0	3	45	25	80010	80010	80010	400050	0	0	49	10291	13371	13371	3330	3	3348	80010	80020	80020	13371	39	1	1	80021	10	9	0	10	80010	10	0	5030	8	19	16	14	13368	80000	80010	13372	13372	13372	13372	13372
80024	13371	100	1	1	0	3	45	25	80010	80010	80010	400050	0	0	49	10291	13371	13371	3330	3	3348	80010	80285	80020	13371	39	1	1	80021	10	9	0	10	80010	10	0	5030	17	19	17	17	13368	80000	80010	13372	13372	13372	13372	13372
80024	13371	100	1	1	12	3	45	25	80010	80136	80010	400050	1	0	49	10291	13371	13371	3330	3	3348	80010	80020	80020	13371	39	1	1	80021	10	9	0	10	80010	10	0	5030	16	19	13	17	13368	80000	80010	13372	13372	13372	13372	13372
80024	13371	100	1	1	0	3	45	25	80010	80010	80010	400050	1	0	49	10291	13371	13371	3330	3	3348	80010	80020	80020	13371	39	1	1	80021	10	9	0	10	80010	10	0	5029	18	19	15	17	13368	80000	80010	13372	13372	13372	13372	13372

SBFX (32-bit)

Test 1: uops

1000 unrolls and 1 iteration

Test 2: Latency 1->2

100 unrolls and 100 iterations

1000 unrolls and 10 iterations

Test 3: throughput

100 unrolls and 100 iterations

1000 unrolls and 10 iterations