Firestorm: UCVTF (scalar, integer, H from H) measurements

Code:

  ucvtf h0, h0

Initialization

  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

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

retire uop (01)	cycle (02)	03	mmu table walk data (08)	09	l2 tlb miss data (0b)	18	19	1e	1f	3f	4e	51	schedule uop (52)	schedule simd uop (54)	dispatch simd uop (57)	ldst uops in schedulers (5b)	60	69	6d	6e	map stall dispatch (70)	map rewind (75)	map stall (76)	dispatch uop (78)	map simd uop (7e)	map simd uop inputs (81)	82	83	flush restart other nonspec (84)	85	inst all (8c)	inst simd alu (9a)	l1d tlb access (a0)	l1d tlb miss (a1)	st unit uop (a7)	l1d cache writeback (a8)	a9	ac	c2	cf	d5	map dispatch bubble (d6)	d9	dd	fetch restart (de)	e0	? simd retires (ee)	f5	f6	f7	f8	fd
1004	3037	23	0	0	0	0	0	0	0	61	2547	25	1000	1000	1000	398160	1	3018	3037	3037	2414	7	2898	1150	1000	1163	3084	3073	2	1	1001	1000	0	0	0	0	0	2846	2	94	1	24	0	1	1	2645	1000	3038	3038	3086	3074	3085
1004	3084	24	1	0	0	1	0	144	88	1002	2538	44	1008	1008	1150	399512	0	3054	3084	3085	2418	6	2907	1000	1162	1161	3084	3084	2	1	1001	1000	0	0	1	12	0	474	1	73	1	40	0	1	1	2664	1000	3085	3038	3084	3086	3038
1004	3085	24	1	0	1	1	1	9	88	803	2547	44	1008	1008	1150	399512	0	3018	3084	3085	2414	8	2914	1150	1162	1000	3085	3037	2	1	1001	1000	2	2	0	4	2	223	4	93	1	24	0	1	1	2645	1000	3086	3084	3085	3086	3085
1004	3084	23	0	0	0	0	1	132	88	750	2538	49	1000	1008	1107	399512	1	3018	3037	3037	2414	3	2895	1000	1000	1000	3037	3037	1	1	1001	1000	0	0	0	3	0	0	0	73	1	16	0	1	1	2629	1000	3038	3038	3038	3038	3038
1004	3037	24	0	0	1	0	0	0	0	104	2547	25	1000	1000	1000	398160	1	3018	3037	3037	2414	3	2895	1000	1000	1000	3037	3037	1	1	1001	1000	2	2	0	3	0	3990	0	91	1	24	0	1	1	2658	1000	3086	3074	3086	3085	3086
1004	3037	25	1	0	1	1	1	132	88	1347	2547	25	1008	1000	1150	399512	1	3090	3037	3084	2422	8	2914	1139	1000	1096	3084	3084	2	1	1001	1000	0	0	0	3	4	5395	2	94	1	24	0	1	1	2724	1000	3120	3121	3122	3121	3122
1004	3168	25	0	1	1	1	1	264	176	899	2538	43	1008	1013	1150	398160	0	3054	3167	3083	2418	10	2914	1150	1162	1168	3083	3083	2	1	1001	1000	0	0	0	0	0	4075	2	100	1	24	0	1	1	2674	1000	3086	3131	3121	3085	3084
1004	3132	25	1	0	0	1	1	132	176	61	2547	25	1000	1000	1000	398160	0	3018	3037	3037	2414	3	2895	1000	1000	1000	3037	3037	1	1	1001	1000	0	0	1	18	2	2860	0	73	1	16	0	1	1	2629	1000	3038	3038	3038	3038	3038
1004	3037	24	0	0	0	0	0	6	0	926	2538	44	1016	1000	1000	398160	1	3018	3037	3037	2414	3	2895	1000	1000	1000	3037	3037	1	1	1001	1000	0	0	0	7	0	72	0	73	1	16	0	1	1	2629	1000	3038	3038	3038	3038	3038
1004	3037	24	0	0	0	0	0	0	0	61	2547	25	1000	1000	1000	398160	1	3018	3037	3037	2414	3	2895	1000	1000	1000	3037	3037	1	1	1001	1000	2	0	0	2	2	220	0	73	1	16	0	1	1	2629	1000	3038	3038	3132	3132	3133

Test 2: Latency 1->2

Code:

  ucvtf h0, h0

Initialization

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

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 3.0037

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

retire uop (01)	cycle (02)	03	1e	1f	3f	4e	51	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)	60	63	69	6d	6e	map stall dispatch (70)	map rewind (75)	map stall (76)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	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)	inst simd alu (9a)	9f	l1d cache writeback (a8)	ac	cf	d5	map dispatch bubble (d6)	dd	fetch restart (de)	e0	? simd retires (ee)	? int retires (ef)	f5	f6	f7	f8	fd
10204	30037	233	0	0	280	29547	25	10100	100	10000	100	10000	500	4277160	1	0	30018	30037	30037	28264	3	28745	10100	200	10000	200	10000	30037	30037	1	1	10201	100	99	100	100	10000	100	0	0	710	2	16	1	1	29633	10000	100	30085	30038	30038	30038	30038
10204	30037	232	0	0	61	29547	25	10100	114	10008	113	10150	500	4277160	1	0	30018	30037	30037	28264	3	28745	10100	200	10000	200	10000	30037	30037	1	1	10201	100	99	100	100	10000	100	0	3	710	1	16	1	1	29633	10000	100	30087	30038	30038	30038	30038
10204	30037	241	276	88	61	29547	25	10100	100	10000	100	10000	500	4277160	1	0	30018	30037	30037	28264	3	28745	10100	200	10000	200	10000	30037	30037	1	1	10201	100	99	100	100	10000	100	0	0	710	1	24	1	1	29633	10000	100	30038	30038	30038	30038	30038
10204	30037	233	12	0	61	29547	25	10100	100	10000	100	10000	500	4277160	1	0	30018	30037	30037	28264	3	28745	10100	200	10000	200	10000	30037	30037	1	1	10201	100	99	100	100	10000	100	0	0	710	1	16	1	1	29633	10000	100	30038	30038	30038	30038	30038
10204	30037	233	12	0	61	29547	25	10100	100	10000	100	10000	500	4277160	1	0	30018	30037	30037	28264	3	28745	10100	200	10000	200	10000	30037	30037	1	1	10201	100	99	100	100	10000	100	0	0	710	1	16	1	1	29633	10000	100	30038	30038	30038	30038	30038
10204	30037	233	0	0	210	29547	25	10130	100	10000	134	10000	522	4277160	1	0	30018	30037	30037	28264	3	28745	10100	200	10333	200	10498	30037	30037	1	1	10201	100	99	100	100	10000	100	0	3	710	1	16	1	1	29633	10000	100	30038	30038	30038	30038	30038
10204	30037	233	0	0	61	29547	25	10100	100	10000	100	10000	500	4277160	1	0	30018	30037	30037	28264	3	28745	10100	200	10000	200	10000	30037	30037	1	1	10201	100	99	100	100	10000	100	0	0	710	1	16	1	1	29633	10000	100	30038	30038	30038	30038	30038
10204	30037	233	0	0	202	29547	25	10100	100	10000	100	10000	500	4277160	1	0	30018	30037	30037	28264	3	28745	10100	200	10000	200	10000	30037	30037	1	1	10201	100	99	100	100	10000	100	0	0	710	1	16	1	1	29633	10000	100	30038	30038	30038	30038	30038
10204	30037	233	27	0	61	29547	25	10100	100	10000	100	10000	500	4277160	1	0	30018	30037	30232	28264	3	28745	10100	200	10000	200	10000	30037	30037	1	1	10201	100	99	100	100	10000	100	0	0	710	1	16	1	1	29633	10000	100	30038	30038	30038	30038	30038
10204	30037	233	0	0	61	29547	25	10100	100	10000	100	10000	500	4277160	1	0	30018	30037	30037	28264	3	28745	10100	200	10000	200	10000	30037	30037	1	1	10201	100	99	100	100	10000	100	0	0	710	1	16	1	1	29633	10000	100	30038	30038	30038	30038	30038

1000 unrolls and 10 iterations

Result (median cycles for code): 3.0037

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

retire uop (01)	cycle (02)	03	3f	4e	51	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)	69	6d	6e	map stall dispatch (70)	map rewind (75)	map stall (76)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	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)	inst simd alu (9a)	9f	l1d cache writeback (a8)	ac	c2	cf	d5	map dispatch bubble (d6)	dd	fetch restart (de)	e0	ea	? simd retires (ee)	? int retires (ef)	f5	f6	f7	f8	fd
10024	30037	233	109	29547	25	10010	10	10000	10	10000	50	4277160	30018	30037	30037	28286	3	28767	10010	20	10000	20	10000	30037	30037	1	1	10021	10	9	10	10	10000	10	0	0	0	640	3	16	3	3	29629	0	10000	10	30038	30038	30038	30038	30038
10024	30037	233	726	29547	25	10010	10	10000	10	10000	50	4277160	30018	30037	30037	28286	3	28767	10010	20	10000	20	10000	30037	30037	1	1	10021	10	9	10	10	10000	10	0	3	2	640	3	16	3	3	29629	0	10000	10	30038	30038	30038	30038	30038
10024	30037	233	61	29547	25	10010	10	10000	10	10000	50	4277160	30018	30037	30037	28286	3	28767	10010	20	10000	20	10000	30037	30037	1	1	10021	10	9	10	10	10000	10	0	0	0	640	3	16	3	3	29629	0	10000	10	30038	30038	30038	30038	30038
10024	30037	233	61	29547	25	10010	10	10000	10	10000	50	4277160	30018	30037	30037	28286	3	28767	10010	20	10000	20	10000	30037	30037	1	1	10021	10	9	10	10	10000	10	0	0	0	640	3	16	3	3	29629	0	10000	10	30038	30038	30038	30038	30038
10024	30037	232	61	29547	25	10010	10	10000	10	10000	50	4277160	30018	30037	30037	28286	3	28767	10010	20	10000	20	10000	30037	30037	1	1	10021	10	9	10	10	10000	10	0	0	0	640	3	16	3	3	29629	0	10000	10	30086	30038	30038	30038	30038
10024	30084	233	61	29547	25	10010	10	10000	10	10000	50	4277160	30018	30037	30037	28286	3	28767	10010	20	10000	20	10000	30037	30037	1	1	10021	10	9	10	10	10000	10	0	0	0	640	3	16	3	3	29629	0	10000	10	30038	30038	30038	30038	30038
10024	30037	233	61	29547	25	10010	10	10000	10	10000	50	4277160	30018	30037	30037	28286	3	28767	10010	20	10000	20	10000	30037	30037	1	1	10021	10	9	10	10	10000	10	0	0	0	640	3	16	3	3	29629	0	10000	10	30038	30038	30038	30038	30038
10024	30037	233	61	29547	25	10010	10	10000	10	10000	50	4277160	30018	30037	30037	28286	3	28767	10010	20	10000	20	10000	30037	30037	1	1	10021	10	9	10	10	10000	10	0	0	0	640	3	16	3	3	29629	0	10000	10	30038	30038	30038	30038	30038
10024	30037	232	61	29547	25	10010	10	10000	10	10000	50	4277160	30018	30037	30037	28286	3	28767	10010	20	10000	20	10000	30037	30037	1	1	10021	10	9	10	10	10000	10	0	0	0	640	3	16	3	3	29629	0	10000	10	30038	30038	30038	30038	30038
10024	30037	232	61	29547	25	10010	10	10000	10	10000	50	4277160	30018	30037	30037	28286	3	28767	10010	20	10000	20	10000	30037	30037	1	1	10021	10	9	10	10	10000	10	0	0	0	640	3	16	3	3	29629	0	10000	10	30038	30038	30228	30038	30038

Test 3: throughput

Count: 8

Code:

  ucvtf h0, h8
  ucvtf h1, h8
  ucvtf h2, h8
  ucvtf h3, h8
  ucvtf h4, h8
  ucvtf h5, h8
  ucvtf h6, h8
  ucvtf h7, h8

Initialization

  movi v8.16b, 9

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code divided by count): 0.2506

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

retire uop (01)	cycle (02)	03	l2 tlb miss data (0b)	1e	3f	51	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)	60	69	6d	6e	map stall dispatch (70)	map rewind (75)	map stall (76)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	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)	inst simd alu (9a)	9f	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)	? simd retires (ee)	? int retires (ef)	f5	f6	f7	f8	fd
80204	20049	155	0	0	30	25	80108	100	80008	100	80020	500	640132	0	20020	20039	20039	9977	6	9990	80120	200	80032	200	80032	20039	20039	1	1	80201	100	99	100	100	80000	100	0	3	1	1	1	5118	3	16	0	0	20036	0	80000	100	20040	20040	20040	20040	20040
80204	20039	155	0	0	1060	25	80108	100	80008	100	80020	500	640132	1	20020	20039	20039	9977	6	9990	80120	200	80032	200	80032	20039	20039	1	1	80201	100	99	100	100	80000	100	0	9	1	1	1	5118	0	16	0	0	20036	0	80000	100	20040	20040	20040	20040	20040
80204	20039	155	0	0	30	25	80108	100	80008	100	80020	500	640132	1	20020	20039	20039	9977	6	9990	80120	200	80032	200	80032	20039	20039	1	1	80201	100	99	100	100	80000	100	0	0	1	1	1	5118	0	16	0	0	20036	2	80000	100	20040	20040	20040	20040	20040
80204	20039	155	0	0	158	25	80108	100	80008	100	80020	500	640132	0	20020	20039	20039	9977	6	9990	80120	200	80032	200	80032	20039	20039	1	1	80201	100	99	100	100	80000	100	0	0	1	1	1	5118	0	16	0	0	20036	0	80000	100	20040	20040	20040	20040	20040
80204	20039	161	0	0	72	25	80108	100	80008	100	80020	500	640132	1	20020	20039	20039	9977	6	9990	80120	200	80032	200	80032	20039	20039	1	1	80201	100	99	100	100	80000	100	0	0	1	1	1	5118	0	16	0	0	20036	0	80000	100	20040	20040	20040	20040	20040
80204	20039	155	0	0	30	25	80108	100	80008	100	80020	500	640132	0	20020	20039	20039	9977	6	9990	80120	200	80032	200	80032	20039	20039	1	1	80201	100	99	100	100	80000	100	0	0	1	1	1	5118	0	16	0	0	20036	0	80000	100	20040	20040	20040	20040	20040
80204	20039	155	0	0	146	25	80108	100	80008	100	80020	500	640132	0	20020	20039	20039	9977	6	9990	80120	200	80032	200	80032	20039	20039	1	1	80201	100	99	100	100	80000	100	0	0	1	1	1	5118	0	16	0	0	20036	0	80000	100	20040	20040	20040	20040	20040
80204	20039	156	0	0	72	25	80108	100	80008	100	80020	500	640132	1	20020	20039	20039	9977	6	9990	80120	200	80032	200	80032	20039	20039	1	1	80201	100	99	100	100	80000	100	0	0	1	1	1	5118	0	16	0	0	20036	1	80000	100	20040	20040	20040	20040	20040
80204	20039	155	0	0	30	25	80108	100	80008	100	80020	500	640132	0	20020	20039	20039	9977	6	9990	80120	200	80032	200	80032	20039	20039	1	1	80201	100	99	100	100	80000	100	0	0	1	1	1	5118	0	16	1	0	20036	0	80000	100	20040	20040	20040	20040	20040
80204	20039	156	0	0	116	25	80108	100	80008	100	80020	500	640132	0	20020	20039	20039	9977	6	9990	80120	200	80032	200	80032	20039	20039	1	1	80201	100	99	100	100	80000	100	0	0	1	1	1	5118	0	16	0	0	20036	0	80000	100	20040	20040	20040	20040	20040

1000 unrolls and 10 iterations

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

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

retire uop (01)	cycle (02)	03	1e	3a	3f	51	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)	60	69	6d	6e	map stall dispatch (70)	simd prf full (72)	map rewind (75)	map stall (76)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	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)	inst simd alu (9a)	9f	l1d cache writeback (a8)	a9	ac	cf	d5	map dispatch bubble (d6)	dd	fetch restart (de)	e0	ea	? simd retires (ee)	? int retires (ef)	f5	f6	f7	f8	fd
80024	20051	155	0	0	125	25	80010	10	80000	10	80000	50	640000	1	20020	20039	20039	9996	0	3	10019	80010	20	80000	20	80000	20039	20039	1	1	80021	10	9	10	10	80000	10	0	0	0	5020	17	16	18	18	20036	0	80000	10	20040	20040	20040	20040	20040
80024	20039	155	0	0	56	25	80010	10	80000	10	80000	50	640000	1	20020	20039	20039	9996	0	3	10019	80010	20	80000	20	80000	20039	20039	1	1	80021	10	9	10	10	80000	10	0	0	0	5020	18	16	18	18	20036	0	80000	10	20040	20040	20040	20040	20040
80024	20039	155	0	0	259	42	80108	10	80000	10	80000	50	640000	0	20020	20039	20039	9996	0	3	10019	80010	20	80000	20	80000	20039	20039	1	1	80021	10	9	10	10	80000	10	0	0	0	5020	18	16	9	18	20036	0	80000	10	20040	20040	20040	20040	20040
80024	20039	156	0	0	73	25	80010	10	80000	10	80000	50	640000	0	20020	20039	20039	9996	0	3	10019	80010	20	80000	20	80000	20039	20039	1	1	80021	10	9	10	10	80000	10	0	0	0	5020	18	16	15	7	20036	0	80000	10	20040	20040	20040	20040	20040
80024	20039	161	0	0	68	25	80010	10	80000	10	80000	50	640000	0	20020	20039	20039	9996	0	3	10019	80010	20	80000	20	80000	20039	20039	1	1	80021	10	9	10	10	80000	10	0	0	0	5020	9	16	18	9	20036	0	80000	10	20040	20040	20040	20040	20040
80024	20039	155	0	0	69	25	80010	10	80000	10	80000	50	640000	0	20020	20039	20039	9996	0	3	10019	80010	20	80000	20	80000	20039	20039	1	1	80021	10	9	10	10	80000	10	0	0	0	5020	9	16	9	18	20036	0	80000	10	20040	20040	20040	20040	20040
80024	20039	155	0	0	68	25	80010	10	80000	10	80000	50	640000	1	20020	20039	20039	9996	0	3	10019	80010	20	80000	20	80000	20039	20039	1	1	80021	10	9	10	10	80000	10	0	0	0	5020	9	16	18	9	20036	0	80000	10	20091	20040	20040	20040	20040
80024	20039	156	0	0	1090	25	80010	10	80000	10	80000	50	640000	0	20020	20039	20039	9996	0	3	10019	80010	20	80000	20	80000	20039	20039	1	1	80021	10	9	10	10	80000	10	0	0	0	5020	9	16	18	9	20036	0	80000	10	20040	20040	20040	20040	20040
80024	20039	155	0	0	68	25	80010	10	80000	10	80000	50	640000	0	20020	20039	20039	9996	0	3	10019	80010	20	80000	20	80000	20039	20039	1	1	80021	10	9	10	10	80000	10	0	0	0	5020	7	16	8	18	20036	0	80000	10	20040	20040	20040	20040	20040
80024	20039	156	0	0	89	25	80010	10	80000	10	80000	50	640000	1	20020	20039	20039	9996	0	3	10019	80010	20	80000	20	80000	20039	20039	1	1	80021	10	9	10	10	80000	10	0	0	0	5020	7	16	18	7	20036	0	80000	10	20040	20040	20040	20040	20040

UCVTF (scalar, integer, H from H)

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