Icestorm: SADALP (vector, 8H) measurements

Code:

  sadalp v0.8h, v1.16b

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: 0.001

Load/store unit issues: 0.000

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)	dispatch simd uop (57)	ldst uops in schedulers (5b)	dispatch uop (78)	map simd uop (7e)	map simd uop inputs (81)	? int output thing (e9)	? simd retires (ee)
1004	3033	1001	1	1000	1000	75905	1000	1000	2000	1	1000
1004	3033	1001	1	1000	1000	75905	1000	1000	2000	1	1000
1004	3033	1001	1	1000	1000	75905	1000	1000	2000	1	1000
1004	3033	1001	1	1000	1000	75905	1000	1000	2000	1	1000
1004	3033	1001	1	1000	1000	75905	1000	1000	2000	1	1000
1004	3033	1001	1	1000	1000	75905	1000	1000	2000	1	1000
1004	3033	1001	1	1000	1000	75905	1000	1000	2000	1	1000
1004	3033	1001	1	1000	1000	75905	1000	1000	2000	1	1000
1004	3033	1001	1	1000	1000	75905	1000	1000	2000	1	1000
1004	3033	1001	1	1000	1000	75905	1000	1000	2000	1	1000

Test 2: Latency 1->1

Code:

  sadalp v0.8h, v1.16b

Initialization

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

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 3.0033

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)	dispatch int uop (56)	dispatch simd uop (57)	int uops in schedulers (59)	ldst uops in schedulers (5b)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	? int output thing (e9)	? simd retires (ee)	? int retires (ef)
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10006	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	769232	10131	200	10045	200	20008	1	10000	100

1000 unrolls and 10 iterations

Result (median cycles for code): 3.0033

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)	dispatch int uop (56)	dispatch simd uop (57)	int uops in schedulers (59)	ldst uops in schedulers (5b)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	? int output thing (e9)	? simd retires (ee)	? int retires (ef)
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10006	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20088	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	769247	10051	20	10044	20	20008	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10

Test 3: Latency 1->2

Code:

  sadalp v0.8h, v0.16b

Initialization

  movi v0.16b, 1

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code): 3.0033

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)	dispatch int uop (56)	dispatch simd uop (57)	int uops in schedulers (59)	ldst uops in schedulers (5b)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	? int output thing (e9)	? simd retires (ee)	? int retires (ef)
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100
10204	30033	10101	101	10000	100	10000	300	768905	10100	200	10004	200	20008	1	10000	100

1000 unrolls and 10 iterations

Result (median cycles for code): 3.0033

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)	dispatch int uop (56)	dispatch simd uop (57)	int uops in schedulers (59)	ldst uops in schedulers (5b)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	? int output thing (e9)	? simd retires (ee)	? int retires (ef)
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10024	30033	10021	21	10000	20	10000	70	768905	10020	20	10000	20	20000	11	10000	10
10025	30066	10027	21	10006	20	10031	70	768905	10020	20	10000	20	20000	11	10000	10

Test 4: throughput

Count: 8

Code:

  movi v0.16b, 0
  sadalp v0.8h, v8.16b
  movi v1.16b, 0
  sadalp v1.8h, v8.16b
  movi v2.16b, 0
  sadalp v2.8h, v8.16b
  movi v3.16b, 0
  sadalp v3.8h, v8.16b
  movi v4.16b, 0
  sadalp v4.8h, v8.16b
  movi v5.16b, 0
  sadalp v5.8h, v8.16b
  movi v6.16b, 0
  sadalp v6.8h, v8.16b
  movi v7.16b, 0
  sadalp v7.8h, v8.16b

Initialization

  movi v8.16b, 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)	dispatch int uop (56)	dispatch simd uop (57)	int uops in schedulers (59)	ldst uops in schedulers (5b)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	? int output thing (e9)	? simd retires (ee)	? int retires (ef)
160204	40490	80109	101	80008	100	80012	300	320052	80112	200	80012	200	160024	1	160000	100
160204	40106	80110	101	80009	100	80013	300	320052	80112	200	80012	200	160024	1	160000	100
160205	40122	80147	101	80046	100	80050	300	320052	80112	200	80012	200	160024	1	160000	100
160204	40096	80110	101	80009	100	80013	300	320052	80112	200	80012	200	160024	1	160000	100
160204	40086	80109	101	80008	100	80012	300	320052	80112	200	80012	200	160024	1	160000	100
160204	40086	80109	101	80008	100	80012	300	320052	80112	200	80012	200	160024	1	160000	100
160204	40086	80109	101	80008	100	80012	300	320052	80112	200	80012	200	160024	1	160000	100
160204	40086	80109	101	80008	100	80012	300	320052	80112	200	80012	200	160024	1	160000	100
160204	40086	80109	101	80008	100	80012	300	320052	80112	200	80012	200	160026	1	160000	100
160204	40286	80110	101	80009	100	80013	300	320056	80113	200	80013	200	160024	1	160000	100

1000 unrolls and 10 iterations

Result (median cycles for code divided by count): 0.5053

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)	dispatch int uop (56)	dispatch simd uop (57)	int uops in schedulers (59)	ldst uops in schedulers (5b)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	? int output thing (e9)	? simd retires (ee)	? int retires (ef)
160024	43771	80020	11	80009	10	80013	30	320000	80010	20	80000	20	160000	1	160000	10
160024	41209	80011	11	80000	10	80000	30	320000	80010	20	80000	20	160098	1	160000	10
160024	40482	80011	11	80000	10	80000	30	320000	80010	20	80000	20	160000	1	160000	10
160024	40438	80011	11	80000	10	80000	30	320000	80010	20	80000	20	160000	1	160000	10
160024	40398	80011	11	80000	10	80000	30	320000	80010	20	80000	20	160000	1	160000	10
160024	40396	80011	11	80000	10	80000	30	320000	80010	20	80000	20	160000	1	160000	10
160024	40405	80011	11	80000	10	80000	30	320000	80010	20	80000	20	160000	1	160000	10
160024	40410	80011	11	80000	10	80000	30	320000	80010	20	80000	20	160000	1	160000	10
160024	40369	80011	11	80000	10	80000	30	320000	80010	20	80000	20	160000	1	160000	10
160024	40396	80011	11	80000	10	80000	30	320000	80010	20	80000	20	160000	1	160000	10

Test 5: throughput

Count: 16

Code:

  sadalp v0.8h, v16.16b
  sadalp v1.8h, v16.16b
  sadalp v2.8h, v16.16b
  sadalp v3.8h, v16.16b
  sadalp v4.8h, v16.16b
  sadalp v5.8h, v16.16b
  sadalp v6.8h, v16.16b
  sadalp v7.8h, v16.16b
  sadalp v8.8h, v16.16b
  sadalp v9.8h, v16.16b
  sadalp v10.8h, v16.16b
  sadalp v11.8h, v16.16b
  sadalp v12.8h, v16.16b
  sadalp v13.8h, v16.16b
  sadalp v14.8h, v16.16b
  sadalp v15.8h, v16.16b

Initialization

  movi v16.16b, 17

(fused SUBS/B.cc loop)

100 unrolls and 100 iterations

Result (median cycles for code divided by count): 0.5002

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)	dispatch int uop (56)	dispatch simd uop (57)	int uops in schedulers (59)	ldst uops in schedulers (5b)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	? int output thing (e9)	? simd retires (ee)	? int retires (ef)
160204	80066	160107	101	160006	100	160010	300	640036	160108	200	160012	200	320024	1	160000	100
160204	80035	160107	101	160006	100	160010	300	640036	160108	200	160012	200	320024	1	160000	100
160204	80035	160105	101	160004	100	160008	300	640036	160108	200	160012	200	320024	1	160000	100
160204	80035	160105	101	160004	100	160008	300	640036	160108	200	160012	200	320024	1	160000	100
160204	80035	160105	101	160004	100	160008	300	640036	160108	200	160012	200	320024	1	160000	100
160204	80035	160105	101	160004	100	160008	300	640036	160108	200	160012	200	320024	1	160000	100
160204	80035	160105	101	160004	100	160008	300	640044	160110	200	160014	200	320024	1	160000	100
160204	80035	160105	101	160004	100	160008	300	640044	160110	200	160014	200	320024	1	160000	100
160204	80035	160107	101	160006	100	160010	300	640036	160108	200	160012	200	320024	1	160000	100
160204	80035	160105	101	160004	100	160008	300	640036	160108	200	160012	200	320024	1	160000	100

1000 unrolls and 10 iterations

Result (median cycles for code divided by count): 0.5002

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)	dispatch int uop (56)	dispatch simd uop (57)	int uops in schedulers (59)	ldst uops in schedulers (5b)	dispatch uop (78)	map int uop (7c)	map simd uop (7e)	map int uop inputs (7f)	map simd uop inputs (81)	? int output thing (e9)	? simd retires (ee)	? int retires (ef)
160025	80161	160053	11	160042	10	160054	30	640044	160020	20	160014	20	320000	1	160000	10
160024	80035	160011	11	160000	10	160000	30	640000	160010	20	160000	20	320000	1	160000	10
160024	80035	160011	11	160000	10	160000	30	640000	160010	20	160000	20	320000	1	160000	10
160024	80035	160011	11	160000	10	160000	30	640000	160010	20	160000	20	320000	1	160000	10
160024	80035	160011	11	160000	10	160000	30	640000	160010	20	160000	20	320000	1	160000	10
160024	80035	160011	11	160000	10	160000	30	640000	160010	20	160000	20	320000	1	160000	10
160024	80035	160011	11	160000	10	160000	30	640000	160010	20	160000	20	320000	1	160000	10
160025	80070	160051	11	160040	10	160052	30	640000	160010	20	160000	20	320024	1	160000	10
160024	80035	160011	11	160000	10	160000	30	640000	160010	20	160000	20	320000	1	160000	10
160024	80184	160108	11	160097	10	160097	30	640000	160010	20	160000	20	320000	1	160000	10

SADALP (vector, 8H)

Test 1: uops

1000 unrolls and 1 iteration

Test 2: Latency 1->1

100 unrolls and 100 iterations

1000 unrolls and 10 iterations

Test 3: Latency 1->2

100 unrolls and 100 iterations

1000 unrolls and 10 iterations

Test 4: throughput

100 unrolls and 100 iterations

1000 unrolls and 10 iterations

Test 5: throughput

100 unrolls and 100 iterations

1000 unrolls and 10 iterations