From: Raphael M Zinsly <rzinsly@linux.ibm.com>
To: libc-alpha@sourceware.org
Subject: Re: [PATCH v3 1/2] powerpc: Add optimized strncpy for POWER9
Date: Tue, 29 Sep 2020 12:22:33 -0300 [thread overview]
Message-ID: <1930c5ea-6a2c-d1bd-3265-36bdc3d56652@linux.ibm.com> (raw)
In-Reply-To: <20200929152103.18564-1-rzinsly@linux.ibm.com>
Benchtest output:
generic_strncpy __strncpy_power9
__strncpy_power8 __strncpy_power7 __strncpy_ppc
Length 16, n 16, alignment 1/ 1: 7.12492 2.55148 2.55079 5.71316
9.05306
Length 16, n 16, alignment 1/ 1: 6.71919 2.51696 2.56187 5.92056
9.43297
Length 16, n 16, alignment 1/ 2: 6.65909 2.53904 2.55074 5.6362 9.30194
Length 16, n 16, alignment 2/ 1: 6.50693 2.51671 2.82125 5.9298 9.18833
Length 2, n 4, alignment 7/ 2: 7.68477 2.27528 5.07192 4.8353 8.4619
Length 4, n 2, alignment 2/ 7: 6.03804 1.6644 2.32355 2.8178 6.27398
Length 2, n 4, alignment 7/ 2: 7.68944 2.31009 4.25078 4.83038
8.59554
Length 4, n 2, alignment 2/ 7: 6.04246 1.66875 2.31775 2.73826 6.8358
Length 16, n 16, alignment 2/ 2: 6.50729 2.51669 2.83075 5.91498
9.60274
Length 16, n 16, alignment 2/ 2: 6.3441 2.51684 2.82829 5.64233 9.29031
Length 16, n 16, alignment 2/ 4: 6.33989 2.51866 2.82089 5.59129
9.50426
Length 16, n 16, alignment 4/ 2: 7.88012 2.51145 3.44369 5.91774
9.50185
Length 4, n 8, alignment 6/ 4: 7.88965 2.27023 5.2189 4.67992 9.06714
Length 8, n 4, alignment 4/ 6: 5.8805 1.69238 2.67589 2.95865 7.70634
Length 4, n 8, alignment 6/ 4: 7.69107 2.29724 5.2196 4.68409 9.07751
Length 8, n 4, alignment 4/ 6: 6.33989 1.69501 2.67179 2.95862
7.75311
Length 16, n 16, alignment 3/ 3: 6.58968 2.51681 3.14295 5.92364
8.86981
Length 16, n 16, alignment 3/ 3: 6.76256 2.51385 3.14379 5.91558
9.02347
Length 16, n 16, alignment 3/ 6: 6.76734 2.53841 3.08182 5.90924
8.98558
Length 16, n 16, alignment 6/ 3: 6.67014 2.51618 4.16905 5.94761
7.81751
Length 8, n 16, alignment 5/ 6: 7.70082 2.30026 4.59182 5.41689
10.5428
Length 16, n 8, alignment 6/ 5: 5.63868 1.87873 2.32929 4.5053 5.78866
Length 8, n 16, alignment 5/ 6: 7.40013 2.2999 4.23768 5.41724 10.1649
Length 16, n 8, alignment 6/ 5: 5.63858 1.87872 2.32768 4.58045
6.02812
Length 16, n 16, alignment 4/ 4: 7.37003 2.5167 3.50594 5.91125 8.93866
Length 16, n 16, alignment 4/ 4: 7.51015 2.51684 3.58684 5.91127
8.60509
Length 16, n 16, alignment 4/ 0: 7.42056 2.51149 3.38179 5.92321
8.86607
Length 16, n 16, alignment 0/ 4: 6.6704 1.87853 2.44519 5.91475 7.68788
Length 16, n 32, alignment 4/ 0: 11.0276 3.0727 6.01877 6.9094 11.4447
Length 32, n 16, alignment 0/ 4: 6.90919 1.87852 2.45708 5.91217 6.7671
Length 16, n 32, alignment 4/ 0: 9.76588 3.07257 5.92168 6.81253
11.8936
Length 32, n 16, alignment 0/ 4: 6.90342 1.88296 2.44527 5.91673
7.68469
Length 16, n 16, alignment 5/ 5: 6.90186 2.51712 3.91963 5.91852
9.46308
Length 16, n 16, alignment 5/ 5: 6.58716 2.51626 3.94884 5.91303
9.59648
Length 16, n 16, alignment 5/ 2: 6.92421 2.52057 3.80827 5.91558 9.3486
Length 16, n 16, alignment 2/ 5: 6.50526 2.53369 2.82035 5.91729 9.065
Length 32, n 64, alignment 3/ 2: 14.0395 3.79978 6.41657 11.19 13.9713
Length 64, n 32, alignment 2/ 3: 9.85699 2.75331 3.21559 8.23056
11.4077
Length 32, n 64, alignment 3/ 2: 14.0923 3.8037 6.38851 11.4514 15.9838
Length 64, n 32, alignment 2/ 3: 9.4437 2.75344 3.21249 8.21276 13.9496
Length 16, n 16, alignment 6/ 6: 6.33989 2.51408 4.38486 5.91681
7.37203
Length 16, n 16, alignment 6/ 6: 6.76503 2.51645 4.26454 5.9103 7.87574
Length 16, n 16, alignment 6/ 4: 6.51654 2.51654 4.24635 5.91578
7.17827
Length 16, n 16, alignment 4/ 6: 7.28735 2.53335 3.54029 5.92337
8.63075
Length 64, n 128, alignment 2/ 4: 15.4973 4.98808 7.34157 11.5113
16.7688
Length 128, n 64, alignment 4/ 2: 11.6235 3.54914 4.80814 10.3103
11.6194
Length 64, n 128, alignment 2/ 4: 15.4979 5.02559 7.28236 11.5045
22.1309
Length 128, n 64, alignment 4/ 2: 11.6138 3.53841 4.80527 10.3293
19.5239
Length 16, n 16, alignment 7/ 7: 6.84212 2.51109 5.0585 5.7457 7.2307
Length 16, n 16, alignment 7/ 7: 6.86215 2.50957 5.06541 5.91726
8.55044
Length 16, n 16, alignment 7/ 6: 6.97428 2.51876 5.05053 5.92637
7.07715
Length 16, n 16, alignment 6/ 7: 7.01347 2.53448 4.38004 5.93278
7.86288
Length 128, n 256, alignment 1/ 6: 17.9407 7.92071 9.38384 16.9419
20.6065
Length 256, n 128, alignment 6/ 1: 13.3609 4.7983 7.967 12.5699 14.9996
Length 128, n 256, alignment 1/ 6: 17.9371 7.69161 9.36672 16.739 38.9048
Length 256, n 128, alignment 6/ 1: 13.3632 4.87671 7.80194 12.7028
33.9017
Length 8, n 16, alignment 0/ 0: 7.4529 2.29963 3.62737 4.22665 7.50268
Length 32, n 16, alignment 0/ 0: 6.86674 1.87853 2.45092 2.41528
7.30161
Length 8, n 16, alignment 7/ 2: 7.40103 2.29399 3.75703 5.43637
8.45285
Length 32, n 16, alignment 7/ 2: 7.72683 2.35278 5.04996 5.93629
7.18881
Length 16, n 32, alignment 0/ 0: 9.87066 3.17511 4.89448 4.41405
10.3408
Length 64, n 32, alignment 0/ 0: 8.06217 2.32926 2.94508 2.71275
8.11769
Length 16, n 32, alignment 6/ 4: 9.50052 3.07627 6.37858 9.46793
10.1393
Length 64, n 32, alignment 6/ 4: 9.7197 2.75154 4.47331 7.73667 9.26558
Length 32, n 64, alignment 0/ 0: 10.9157 3.79013 4.83041 4.97713
11.5486
Length 128, n 64, alignment 0/ 0: 9.28057 3.15788 3.5178 4.23091 11.0874
Length 32, n 64, alignment 5/ 6: 14.0472 3.8515 7.26431 10.1343 12.8115
Length 128, n 64, alignment 5/ 6: 11.5493 3.5659 5.05553 9.1005 13.4053
Length 64, n 128, alignment 0/ 0: 12.0056 4.94615 6.45436 7.06235
14.4743
Length 256, n 128, alignment 0/ 0: 7.87506 4.49546 6.4492 5.38877 12.1437
Length 64, n 128, alignment 4/ 0: 12.4174 4.99773 7.73749 11.1452
16.1494
Length 256, n 128, alignment 4/ 0: 12.2601 4.88446 6.95948 13.3726
16.7583
Length 128, n 256, alignment 0/ 0: 13.9215 7.51155 7.87942 8.79876
20.4226
Length 512, n 256, alignment 0/ 0: 10.5798 6.77319 8.79757 9.03297
20.0197
Length 128, n 256, alignment 3/ 2: 18.0213 7.57884 9.89436 18.7839
20.5445
Length 512, n 256, alignment 3/ 2: 16.9909 7.07957 9.9271 23.2621 25.2442
Length 256, n 512, alignment 0/ 0: 17.6825 12.3074 13.3245 13.9381
28.7687
Length 1024, n 512, alignment 0/ 0: 16.3837 10.8306 16.6999 16.6797
38.0562
Length 256, n 512, alignment 2/ 4: 23.1953 13.0445 14.324 26.8918 30.2049
Length 1024, n 512, alignment 2/ 4: 25.4059 12.0938 17.2483 41.4883
47.2025
Length 512, n 1024, alignment 0/ 0: 21.029 17.1782 19.4815 21.0035 43.2361
Length 2048, n 1024, alignment 0/ 0: 28.5154 19.3221 36.9624 35.482 68.4792
Length 512, n 1024, alignment 1/ 6: 32.4103 17.9272 21.5421 46.6099 55.059
Length 2048, n 1024, alignment 1/ 6: 43.0516 21.6315 37.8787 77.7889
83.4195
next prev parent reply other threads:[~2020-09-29 15:22 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-29 15:21 Raphael Moreira Zinsly
2020-09-29 15:21 ` [PATCH v3 2/2] powerpc: Add optimized stpncpy " Raphael Moreira Zinsly
2020-09-29 15:23 ` Raphael M Zinsly
2020-09-30 13:42 ` Adhemerval Zanella
2020-09-30 14:21 ` Raphael M Zinsly
2020-09-30 14:46 ` Adhemerval Zanella
2020-11-12 17:12 ` Tulio Magno Quites Machado Filho
2020-09-29 15:22 ` Raphael M Zinsly [this message]
2020-10-15 15:20 ` [PATCH v3 1/2] powerpc: Add optimized strncpy " Lucas A. M. Magalhaes
2020-11-12 17:09 ` Tulio Magno Quites Machado Filho
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1930c5ea-6a2c-d1bd-3265-36bdc3d56652@linux.ibm.com \
--to=rzinsly@linux.ibm.com \
--cc=libc-alpha@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).