50 commits

Author	SHA1	Message	Date
Stefano Pigozzi	518040a414	Allow including or excluding init_array via the POLYBENCH_INCLUDE_INIT macro	2022-11-28 15:44:40 +01:00
Stefano Pigozzi	bf873d846c	Cleanup and format a lot of the atax.cu file	2022-11-28 15:43:05 +01:00
Stefano Pigozzi	2f476affee	Exclude again init_array from the benchmark	2022-11-28 15:26:52 +01:00
Stefano Pigozzi	f4a903371a	Run format document to indent code using tabs	2022-11-28 15:26:10 +01:00
Stefano Pigozzi	f0394d1b3b	Indent M_PI definition	2022-11-28 15:23:27 +01:00
Stefano Pigozzi	2ab3f9b06b	Try fixing the makefile	2022-11-28 15:08:26 +01:00
Stefano Pigozzi	118b18a2a1	Move --silent to the bench script	2022-11-28 14:46:43 +01:00
Stefano Pigozzi	fbfa6f3b47	Use tabs in Makefile	2022-11-28 14:41:35 +01:00
Stefano Pigozzi	43deb504c9	Remove OpenMP pragmas	2022-11-28 14:38:13 +01:00
Stefano Pigozzi	a2a070bb3a	Configure makefile to use nvcc	2022-11-28 14:37:37 +01:00
Stefano Pigozzi	28479dfb4b	Cleanup comments	2022-11-17 19:53:46 +01:00
Stefano Pigozzi	44fe50bd4a	Reduce EXTRALARGE_DATASET to 12000	2022-11-17 19:53:39 +01:00
Stefano Pigozzi	bffa050239	Merge branch 'master' of github.com:Steffo99/unimore-hpc-1	2022-11-17 19:14:17 +01:00
Stefano Pigozzi	097efddbe3	Remove -g3 CFLAG	2022-11-17 19:13:26 +01:00
Fabio Zanichelli	0ba75336e6	Spostato polybench_start per cronometrare anche le inizializzazioni	2022-11-17 17:33:59 +01:00
Stefano Pigozzi	a86d078546	Run bench with all dataset sizes	2022-11-17 14:58:32 +01:00
Stefano Pigozzi	c91361ba88	Add indicator of progress for single runs	2022-11-17 03:02:20 +01:00
Stefano Pigozzi	7cd8707bb9	Make some optimizations toggleable, so results can be compared easily	2022-11-17 02:59:31 +01:00
Stefano Pigozzi	60a061991b	main: Remove commented duplicate polybench_start_instruments	2022-11-17 02:07:57 +01:00
Stefano Pigozzi	20e653ea70	Fix EXTRALARGE_DATASET so it does not overflow anymore	2022-11-17 02:03:18 +01:00
Stefano Pigozzi	e49e89817d	kernel_atax: Format for loops	2022-11-17 01:59:27 +01:00
Stefano Pigozzi	1850c42a9f	kernel_atax: Remove nested parallelization ... Seems to improve the execution time on my PC 0.0340s → 0.0317s	2022-11-17 01:59:09 +01:00
Stefano Pigozzi	a16813dc01	main: Remove blank lines	2022-11-17 01:57:25 +01:00
Stefano Pigozzi	0eb63cb684	init_array: Format second for loop	2022-11-17 01:57:13 +01:00
Stefano Pigozzi	ac1ec275d7	print_array: Format for loop	2022-11-17 01:56:52 +01:00
Stefano Pigozzi	c5c79e00c4	print_array: Add (obvious) comment	2022-11-17 01:56:27 +01:00
Stefano Pigozzi	1e45a5adca	print_array: Remove newline after every 20 elements ... The terminal will handle wrapping as necessary.	2022-11-17 01:56:08 +01:00
Stefano Pigozzi	a23fd895e9	init_array: Parallelize the second loop ... The performance hit is gone? 0.0437s → 0.0342s	2022-11-17 01:54:13 +01:00
FABIO ZANICHELLI	9dc24a3367	Aggiunta una reduction (al momento fa poco, magari con acceleratore va meglio), tolto un *4 perche Jetson ha 4 core CPU)	2022-11-16 15:01:58 -05:00
Stefano Pigozzi	d89c501b59	kernel_atax: Parallelizing the second loop gives a nice speedup	2022-11-16 18:05:12 +01:00
Stefano Pigozzi	9c153bb89f	Hide insignificant digits in the bench target	2022-11-16 18:03:24 +01:00
Stefano Pigozzi	c104caa1a6	Use THREAD_COUNT instead of a fixed amount of threads	2022-11-16 17:51:27 +01:00
Stefano Pigozzi	e23d565fd2	Create bench target for calculating the average of 9 runs	2022-11-16 17:39:09 +01:00
Stefano Pigozzi	8cb741d24d	Create bench target	2022-11-16 17:09:27 +01:00
Stefano Pigozzi	f655df0eb7	Make the compilation silent so only the execution time is output	2022-11-16 17:06:16 +01:00
Stefano Pigozzi	cf79cb0924	Improve makefile flags	2022-11-16 17:04:50 +01:00
Stefano Pigozzi	7fc2506cc7	kernel_atax: Remove useless tmp array, use a local variable instead ... Local improvement: 30ms Co-authored-by: Caterina Gazzotti <gazzotti.caterina@gmail.com>	2022-11-16 14:21:57 +01:00
Stefano Pigozzi	aefa038d44	Revert "Aggiunge due for parallelizzati" ... This reverts commit 5b8b4f84b2.	2022-11-16 14:15:09 +01:00
Gattopandacorno	5b8b4f84b2	aggiunge due for parallelizzati	2022-11-16 11:09:30 +01:00
Stefano Pigozzi	0a91ad019a	Add CFLAGS to the Makefile	2022-11-16 01:43:59 +01:00
Stefano Pigozzi	e891286291	kernel_atax: Add comments to the second iteration	2022-11-16 01:43:49 +01:00
Stefano Pigozzi	57fe10ad6a	kernel_atax: Try parallelizing the first cycle ... Doesn't seem to create a big difference	2022-11-16 01:43:22 +01:00
Stefano Pigozzi	d5664a9393	print_array: Improve wording	2022-11-16 01:42:53 +01:00
Stefano Pigozzi	18466da754	init_array: Leave a comment with my experiments on the second iteration	2022-11-16 01:42:37 +01:00
Stefano Pigozzi	7fcff5d127	init_array: Actually, do optimize the first iteration	2022-11-16 01:41:29 +01:00
Stefano Pigozzi	f48d351890	init_array: Do not optimize, costs are greater than benefits	2022-11-15 19:46:23 +01:00
Stefano Pigozzi	1c834ba294	Delete blank line	2022-11-14 15:40:03 +01:00
Stefano Pigozzi	46dedaef45	Start commenting some lines ... Co-authored-by: Caterina Gazzotti <gazzotti.caterina@gmail.com> Co-authored-by: Fabio Zanichelli <274956@studenti.unimore.it>	2022-11-11 16:47:48 +00:00
Stefano Pigozzi	5bbe60d482	Create workaround for unavailable M_PI	2022-11-11 16:42:39 +00:00
Alessandro Capotondi	e11b42a518	init commit	2022-11-11 13:23:45 +01:00