Icestorm: SCVTF (scalar, integer, H from W) measurements

Code:

  scvtf h0, w0

Initialization

  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

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

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)
2004	615	2001	1	1000	1000	1000	1000	3000	8000	2000	1000	1000	1000	1000	1	1000	1000
2004	537	2001	1	1000	1000	1000	1000	3000	8000	2000	1000	1000	1000	1000	1	1000	1000
2004	537	2001	1	1000	1000	1000	1000	3000	8000	2000	1000	1000	1000	1000	1	1000	1000
2004	537	2001	1	1000	1000	1000	1000	3000	8000	2000	1000	1000	1000	1000	1	1000	1000
2004	537	2001	1	1000	1000	1000	1000	3000	8000	2000	1000	1000	1000	1000	1	1000	1000
2004	537	2001	1	1000	1000	1000	1000	3000	8000	2000	1000	1000	1000	1000	1	1000	1000
2004	537	2001	1	1000	1000	1000	1000	3000	8000	2000	1000	1000	1000	1000	1	1000	1000
2004	537	2001	1	1000	1000	1000	1000	3000	8000	2000	1000	1000	1000	1000	1	1000	1000
2004	537	2001	1	1000	1000	1000	1000	3000	8000	2000	1000	1000	1000	1000	1	1000	1000
2004	537	2001	1	1000	1000	1000	1000	3000	8000	2000	1000	1000	1000	1000	1	1000	1000

Test 2: Latency 1->2 roundtrip

Code:

  scvtf h0, w0
  fmov x0, d0

Initialization

  mov x0, 1
  mov x1, 2

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 10.0030

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

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)
30204	100030	40101	10101	20000	10000	100	20000	10001	300	1546176	2578264	30101	200	10002	20004	200	10002	20004	10001	10000	10000	10100
30204	100030	40101	10101	20000	10000	100	20000	10017	300	1546384	2578617	30146	200	10021	20039	200	10002	20004	10001	10000	10000	10100
30204	100030	40101	10101	20000	10000	100	20000	10000	300	1546174	2578264	30100	200	10002	20002	200	10002	20002	10001	10000	10000	10100
30204	100030	40101	10101	20000	10000	100	20000	10000	300	1546174	2578264	30100	200	10002	20002	200	10002	20002	10001	10000	10000	10100
30204	100030	40101	10101	20000	10000	100	20000	10000	300	1546174	2578264	30100	200	10002	20002	200	10002	20002	10001	10000	10000	10100
30204	100030	40101	10101	20000	10000	100	20000	10000	300	1546174	2578264	30100	200	10002	20002	200	10002	20002	10001	10000	10000	10100
30204	100030	40101	10101	20000	10000	100	20000	10000	300	1546174	2578264	30100	200	10002	20002	200	10002	20002	10001	10000	10000	10100
30204	100030	40101	10101	20000	10000	100	20000	10001	300	1546172	2578264	30101	200	10002	20004	200	10002	20002	10001	10000	10000	10100
30204	100030	40101	10101	20000	10000	100	20000	10000	300	1546174	2578264	30100	200	10002	20002	200	10002	20002	10001	10000	10000	10100
30204	100030	40101	10101	20000	10000	100	20000	10000	300	1546174	2578264	30100	200	10002	20002	200	10002	20002	10001	10000	10000	10100

1000 unrolls and 10 iterations

Result (median cycles for code): 10.0030

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

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)
30024	100030	40011	10011	20000	10000	10	20000	10000	30	1546168	2578264	30010	20	10000	20000	20	10000	20000	10001	10000	10000	10010
30024	100030	40011	10011	20000	10000	10	20000	10000	30	1546174	2578264	30010	20	10000	20000	20	10021	20039	10003	10000	10000	10010
30024	100037	40011	10011	20000	10000	10	20000	10000	30	1546174	2578264	30010	20	10000	20000	20	10000	20000	10001	10000	10000	10010
30024	100030	40011	10011	20000	10000	10	20000	10000	30	1546174	2578264	30010	20	10000	20000	20	10000	20000	10001	10000	10000	10010
30024	100030	40011	10011	20000	10000	10	20000	10000	30	1546174	2578264	30010	20	10000	20000	20	10000	20000	10001	10000	10000	10010
30024	100030	40011	10011	20000	10000	10	20000	10000	30	1546174	2578264	30010	20	10000	20000	20	10000	20000	10001	10000	10000	10010
30024	100030	40011	10011	20000	10000	10	20000	10000	30	1546174	2578264	30010	20	10000	20000	20	10000	20000	10001	10000	10000	10010
30024	100030	40011	10011	20000	10000	10	20000	10019	30	1546386	2578617	30058	20	10022	20042	20	10000	20000	10001	10000	10000	10010
30024	100030	40011	10011	20000	10000	10	20000	10017	30	1546496	2578799	30056	20	10021	20039	20	10000	20000	10001	10000	10000	10010
30024	100030	40011	10011	20000	10000	10	20000	10000	30	1546174	2578264	30010	20	10000	20000	20	10000	20000	10001	10000	10000	10010

Test 3: throughput

Count: 8

Code:

  scvtf h0, w8
  scvtf h1, w8
  scvtf h2, w8
  scvtf h3, w8
  scvtf h4, w8
  scvtf h5, w8
  scvtf h6, w8
  scvtf h7, w8

Initialization

  mov x8, 9

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

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

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

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)
160204	40137	160115	101	80006	80008	100	80012	80012	300	240036	640080	160124	200	80012	80012	200	80012	80012	1	80000	80000	100
160204	40092	160115	101	80006	80008	100	80012	80012	300	240036	640080	160124	200	80012	80012	200	80012	80012	1	80000	80000	100
160204	40090	160115	101	80006	80008	100	80012	80012	300	240036	640080	160124	200	80012	80012	200	80012	80012	1	80000	80000	100
160204	40090	160115	101	80006	80008	100	80012	80012	300	240036	640080	160124	200	80012	80012	200	80012	80012	1	80000	80000	100
160204	40090	160115	101	80006	80008	100	80012	80012	300	240036	640080	160124	200	80012	80012	200	80012	80012	1	80000	80000	100
160204	40093	160115	101	80006	80008	100	80012	80011	300	253138	654787	160120	200	80012	80012	200	80012	80012	1	80000	80000	100
160204	40103	160115	101	80006	80008	100	80012	80012	300	248160	651396	160124	200	80012	80012	200	80012	80012	1	80000	80000	100
160204	40090	160115	101	80006	80008	100	80012	80012	300	240036	640080	160124	200	80012	80012	200	80012	80012	1	80000	80000	100
160204	40090	160115	101	80006	80008	100	80012	80012	300	240036	640080	160124	200	80012	80012	200	80012	80012	1	80000	80000	100
160204	40090	160115	101	80006	80008	100	80012	80012	300	240036	640080	160124	200	80012	80012	200	80012	80012	1	80000	80000	100

1000 unrolls and 10 iterations

Result (median cycles for code divided by count): 0.5008

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

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)
160025	40577	160087	11	80037	80039	10	80055	80012	30	244979	653013	160034	20	80012	80012	20	80055	80055	1	80000	80000	10
160024	40076	160011	11	80000	80000	10	80000	80000	30	249271	649255	160010	20	80000	80000	20	80000	80000	1	80000	80000	10
160024	40048	160011	11	80000	80000	10	80000	80000	30	254236	667847	160010	20	80000	80000	20	80000	80000	1	80000	80000	10
160024	40042	160011	11	80000	80000	10	80000	80000	30	240000	640000	160010	20	80000	80000	20	80000	80000	1	80000	80000	10
160024	40068	160011	11	80000	80000	10	80000	80000	30	240000	640000	160010	20	80000	80000	20	80000	80000	1	80000	80000	10
160024	40048	160011	11	80000	80000	10	80000	80000	30	247433	655433	160010	20	80000	80000	20	80000	80000	1	80000	80000	10
160024	40042	160011	11	80000	80000	10	80000	80000	30	240000	640000	160010	20	80000	80000	20	80000	80000	1	80000	80000	10
160024	40055	160011	11	80000	80000	10	80000	80000	30	240276	640276	160010	20	80000	80000	20	80000	80000	1	80000	80000	10
160025	40095	160085	11	80036	80038	10	80054	80000	30	240000	640000	160010	20	80000	80000	20	80000	80000	1	80000	80000	10
160024	40050	160011	11	80000	80000	10	80000	80000	30	246743	646743	160010	20	80000	80000	20	80000	80000	1	80000	80000	10

SCVTF (scalar, integer, H from W)

Test 1: uops

1000 unrolls and 1 iteration

Test 2: Latency 1->2 roundtrip

100 unrolls and 100 iterations

1000 unrolls and 10 iterations

Test 3: throughput

100 unrolls and 100 iterations

1000 unrolls and 10 iterations