[CRPYT] [Python] Use numpy array instead of standard list

[misterdz-mh]

TEAM 487

[dedece35]

TO DISCUSS INSIDE CORE-TEAM : issue to close ? no existing PR but really in project Kanban
(https://github.com/orgs/green-code-initiative/projects/1?pane=issue&itemId=24907316)

[jhertout]

We perform some basic measures for this rule and the result were not satisfying. If there is an advantage to use numpy it must be detailed. I think we should close this one.

Here are the things we tested :

Use numpy array instead of standard list

A - Construction d'une liste

1 - Non compliant

list(range(1000000))

2 - Compliant

np.arange(1000000)

B - Itération d'une liste

1 - Non compliant

Liste utilisé :

my_list = list(range(1000000))

Code testé :

for item in my_list:
    tmp = item

2 - Compliant

Liste utilisé :

my_list = np.arange(1000000)

Code testé :

for item in my_list:
    tmp = item

C - Récupération d'un élément d'une liste

1 - Non compliant

Liste utilisé :

my_list = list(range(1000000))

Code testé :

tmp = my_list[500000]

2 - Compliant

Liste utilisé :

my_list = np.arange(1000000)

Code testé :

tmp = my_list[500000]

And the results of the measures :

Use numpy array instead of standard list

A - Construction d'une liste

Itération : 500

1 - Non compliant

Numéro d'itération	Consommation en Joules	Temps d'exécution en secondes	Rapport J/s
1	226.73	14.211546971	15.95392820
2	225.80	13.973445392	16.15922155
3	226.30	14.076505224	16.07643348
4	228.87	14.204236174	16.11279882
5	228.34	14.237142466	16.03833076
6	227.08	14.153487727	16.04410194
7	228.23	14.236500496	16.03132736
8	225.82	14.057398904	16.06413829
9	228.24	14.183727774	16.09167939
10	227.67	14.150479407	16.08920754
11	227.71	14.100659046	16.14889057
12	228.29	14.250680476	16.01958589
13	226.63	14.118486547	16.05200382
14	225.51	14.022863767	16.08159387
15	224.61	14.024847034	16.01514793
16	225.64	14.026694030	16.08647051
17	227.28	14.123749114	16.09204455
18	227.58	14.073604854	16.17069701
19	229.07	14.165816004	16.17061805
20	229.36	14.147023417	16.21259774

2 - Compliant

Numéro d'itération	Consommation en Joules	Temps d'exécution en secondes	Rapport J/s
1	6.58	0.298146448	22.06969106
2	6.67	0.295191261	22.59551985
3	6.82	0.297514261	22.92327089
4	6.82	0.297843323	22.89794490
5	6.67	0.297768124	22.39997992
6	6.70	0.295827283	22.64835052
7	6.72	0.299124966	22.46552699
8	6.62	0.295482665	22.40402156
9	6.77	0.296678858	22.81928697
10	6.69	0.294887133	22.68664601
11	6.64	0.294725764	22.52941822
12	6.73	0.299843733	22.44502472
13	6.65	0.294155281	22.60710729
14	6.72	0.297277877	22.60511299
15	6.69	0.294785327	22.69448099
16	6.68	0.296182546	22.55365851
17	6.70	0.295367076	22.68363857
18	6.67	0.295390171	22.58030447
19	6.71	0.296991203	22.59326179
20	6.65	0.293984978	22.62020340

---

B - Itération d'une liste

Itération : 200

1 - Non compliant

Numéro d'itération	Consommation en Joules	Temps d'exécution en secondes	Rapport J/s
1	143.41	8.173361236	17.54602492
2	144.98	8.303619395	17.45985613
3	131.78	7.501407735	17.56736930
4	132.06	7.502569370	17.60196987
5	137.56	7.903276506	17.40543936
6	137.19	7.734243464	17.73799863
7	140.43	7.920118769	17.73079471
8	145.29	8.167104089	17.78965939
9	145.37	8.309516583	17.49439916
10	134.33	7.706901237	17.42983280
11	136.23	7.700913954	17.69010805
12	152.16	8.639646176	17.61183234
13	138.73	7.906341411	17.54667459
14	138.64	7.866312787	17.62452164
15	133.13	7.501496175	17.74712629
16	135.72	7.694811850	17.63785816
17	135.94	7.650249324	17.76935551
18	133.72	7.510285395	17.80491592
19	153.59	8.690937717	17.67243133
20	146.30	8.323642714	17.57643918

2 - Compliant

Numéro d'itération	Consommation en Joules	Temps d'exécution en secondes	Rapport J/s
1	237.36	12.779763873	18.57311311
2	241.67	13.081704894	18.47389174
3	239.73	12.787814427	18.74675311
4	242.82	13.071383557	18.57645741
5	235.85	12.769037699	18.47045999
6	239.07	12.778712854	18.70845700
7	237.21	12.786104169	18.55217170
8	246.36	13.305249354	18.51600022
9	248.10	13.335602493	18.60433378
10	247.69	13.342719391	18.56368201
11	239.13	12.811708422	18.66495803
12	242.88	13.038682697	18.62764863
13	240.70	13.075853225	18.40797658
14	241.07	13.077014954	18.43463518
15	236.63	12.816065218	18.46354524
16	236.28	12.758304294	18.51970250
17	255.53	13.712045731	18.63543959
18	238.95	12.925029380	18.48738544
19	243.28	13.134115617	18.52275456
20	251.90	13.538669159	18.60596466

---

C - Récupération d'un élément d'une liste

Itération : 100.000.000

1 - Non compliant

Numéro d'itération	Consommation en Joules	Temps d'exécution en secondes	Rapport J/s
1	109.24	6.259988527	17.45051121
2	99.44	5.734942290	17.33932007
3	102.04	5.846834486	17.45217865
4	106.62	6.102100406	17.47267217
5	101.69	5.814824555	17.48805987
6	99.86	5.725344921	17.44174392
7	107.80	6.033074307	17.86817044
8	107.30	6.150831456	17.44479600
9	101.69	5.840219330	17.41201729
10	107.77	6.184935986	17.42459424
11	100.78	5.734545254	17.57419211
12	103.52	5.782912192	17.90101536
13	103.35	5.850249311	17.66591379
14	103.95	5.905713920	17.60159760
15	103.34	5.919728349	17.45688212
16	100.43	5.731974073	17.52101435
17	103.11	5.907553667	17.45392523
18	103.12	5.792451412	17.80247992
19	104.91	6.026638367	17.40771448
20	100.95	5.748075928	17.56239849

2 - Compliant

Numéro d'itération	Consommation en Joules	Temps d'exécution en secondes	Rapport J/s
1	189.52	10.145179463	18.68079324
2	190.63	10.217838957	18.65658685
3	219.37	12.441805554	17.63168529
4	188.30	10.103760791	18.63662490
5	191.76	10.268583140	18.67443613
6	194.17	10.378052149	18.70967665
7	183.96	10.007252852	18.38266732
8	191.92	10.279490444	18.67018613
9	188.44	10.096896236	18.66316099
10	196.46	10.530679005	18.65596699
11	186.98	10.006633914	18.68560413
12	196.30	10.466818655	18.75450473
13	195.05	10.440584045	18.68190506
14	186.40	10.033753463	18.57729519
15	186.02	10.157204095	18.31409492
16	190.14	10.163845529	18.70748620
17	198.29	10.631342053	18.65145519
18	187.58	10.059887267	18.64633221
19	186.93	9.986243065	18.71875126
20	190.74	10.181448440	18.73407316

We can see that the results for the iteration and the access of an element of the list are not so efficient in term of energy. The rule cannot be integrated in this state.

[glalloue]

Related to previous results, this rule is not relevant.
This issue will be closed and can be reopened if new results show Numpy could be better than a standard list.

A new rules list will be created to list studied rules but refused to archive it and keep results somewhere

(seen with @jhertout and @dedece35 )

[dedece35]

as discussed, I affect this issue to @jhertout to initiate a doc to reference Rules refused or deleted (with reference to issue / PR dealing with discussion about the decision)
next @jhertout will clos this issue

• initiate doc for refused / deleted rules : check Rules.md: Refused rules #234
• mark as done the ticket of this rule on Kanban project
• close this issue