Chapel versus Classic Fortran

How the programs are written matters! Always look at the source code.

If the fastest programs are * hand-written vector instructions, does the host language matter? You might be more interested in the less optimised programs — more seconds, less gz source code.

fannkuch-redux
	elapsed				Mean or 95% CI
source	secs	mem	gz	make	cpu secs
Chapel #2	8.48	9,699	737	21.01	33.63—33.69
Classic Fortran	39.02	3,191	608	2.95	38.97—39.21
Classic Fortran #3	10.15	12,329	1164	3.36	40.49
n-body
	elapsed				Mean or 95% CI
source	secs	mem	gz	make	cpu secs
Chapel #3	3.61	9,720	983	20.46	3.58
Classic Fortran #6	4.22	2,146	1530	2.93	4.23
Chapel #4	4.68	9,568	1027	20.35	4.66
Classic Fortran #2	5.39	2,146	1507	2.88	5.39
Classic Fortran	5.45	2,142	1400	3.14	5.45
Chapel #2	5.74	9,929	977	20.76	5.72
Classic Fortran #5	5.97	2,150	1669	3.04	5.99—6.10
Classic Fortran #4	6.18	2,146	1183	2.96	6.23
Classic Fortran #3	6.33	2,146	1311	3.03	6.36—6.40
spectral-norm
	elapsed				Mean or 95% CI
source	secs	mem	gz	make	cpu secs
Classic Fortran #2	2.68	4,239	520	3.16	2.67
Classic Fortran #3	0.72	5,259	644	3.06	2.86—2.91
Chapel #2	0.79	9,916	348	20.86	2.91
Chapel	1.50	10,109	322	20.65	5.78
mandelbrot
	elapsed				Mean or 95% CI
source	secs	mem	gz	make	cpu secs
Chapel #3	1.18	42,295	582	21.29	4.52
Classic Fortran #7	1.44	79,483	965	2.98	5.64
Classic Fortran #6	3.27	79,585	669	2.68	10.80
Chapel	6.68	41,964	447	20.87	26.49
Classic Fortran #4	10.40	79,577	626	2.70	26.84—26.88
Classic Fortran #5	45.35	45,691	528	2.56	45.36—45.41
fasta
	elapsed				Mean or 95% CI
source	secs	mem	gz	make	cpu secs
Chapel #6	0.84	10,310	1108	21.24	0.81
Classic Fortran #4	2.06	2,142	1355	3.21	2.06
Chapel #5	1.01	11,284	1391	22.88	3.54—3.60
Chapel #3	4.07	9,597	1219	22.91	4.05
Classic Fortran #3	6.98	2,146	1217	2.98	7.05—7.16
Classic Fortran	7.19	2,142	1193	2.78	7.23—7.31
k-nucleotide
	elapsed				Mean or 95% CI
source	secs	mem	gz	make	cpu secs
Chapel #4	5.02	197,980	1074	28.32	16.99—17.77
Chapel #3	5.99	190,841	1044	28.35	20.72—21.61
Classic Fortran		Empty
reverse-complement
	elapsed				Mean or 95% CI
source	secs	mem	gz	make	cpu secs
Chapel #9	1.00	1,087,742	1260	21.34	2.82—2.88
Chapel #5	1.83	1,442,632	554	20.72	3.23—3.45
Chapel #3	1.84	1,669,767	709	20.76	3.27—3.41
Classic Fortran	4.24	1,012,376	1026	2.89	4.25
Classic Fortran #2	45.32	790,282	778	3.25	45.50—45.79
binary-trees
	elapsed				Mean or 95% CI
source	secs	mem	gz	make	cpu secs
Classic Fortran #2	2.22	141,066	1155	2.78	6.99—7.03
Chapel #5	2.06	1,825,448	646	19.88	7.24
Chapel #3	8.61	314,032	494	20.69	26.00—26.04
Chapel #4	7.29	495,673	489	20.08	26.60—26.71
Classic Fortran	32.31	540,062	780	2.66	32.35—32.48
pidigits
	elapsed				Mean or 95% CI
source	secs	mem	gz	make	cpu secs
* Classic Fortran #4	0.74	5,399	987	2.35	0.74
* Chapel #4	0.80	11,416	515	19.88	0.78
Chapel #2	0.80	11,543	423	20.04	0.78
regex-redux
	elapsed				Mean or 95% CI
source	secs	mem	gz	make	cpu secs
* Classic Fortran	0.98	179,618	2229	3.02	3.11—3.15
Chapel #3	3.17	392,372	503	22.24	4.07—4.13
Chapel #2	3.58	391,442	498	21.16	4.51
Chapel	chpl version 2.3.0 built with LLVM version 19.1.1 Copyright 2020-2024 Hewlett Packard Enterprise Development LP Copyright 2004-2019 Cray Inc.
Classic Fortran	Fortran Compiler Classic 2021.12.0 20240222