Message-ID: <20120828121743.27185dq47e19rtwk@www.81.fi>
Date:	Tue, 28 Aug 2012 12:17:43 +0300
From:	Jussi Kivilinna <jussi.kivilinna@...et.fi>
To:	Borislav Petkov <bp@...en8.de>
Cc:	Johannes Goetzfried 
	<Johannes.Goetzfried@...ormatik.stud.uni-erlangen.de>,
	linux-crypto@...r.kernel.org,
	Herbert Xu <herbert@...dor.apana.org.au>,
	Tilo Müller 
	<tilo.mueller@...ormatik.uni-erlangen.de>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] crypto: twofish - add x86_64/avx assembler
 implementation

Quoting Borislav Petkov <bp@...en8.de>:

> On Wed, Aug 22, 2012 at 10:20:03PM +0300, Jussi Kivilinna wrote:
>> Actually it does look better, at least for encryption. Decryption  
>> had different
>> ordering for test, which appears to be bad on bulldozer as it is on
>> sandy-bridge.
>>
>> So, yet another patch then :)
>
> Here you go:

Thanks!

With this patch twofish-avx is faster than twofish-3way for 256, 1k  
and 8k tests.

size    old-vs-new      new-vs-3way     old-vs-3way
         ecb-enc ecb-dec ecb-enc ecb-dec ecb-enc ecb-dec
256     1.10x   1.11x   1.01x   1.01x   0.92x   0.91x
1k      1.11x   1.12x   1.08x   1.07x   0.97x   0.96x
8k      1.11x   1.13x   1.10x   1.08x   0.99x   0.97x

-Jussi

>
> [  153.736745]
> [  153.736745] testing speed of async ecb(twofish) encryption
> [  153.745806] test 0 (128 bit key, 16 byte blocks): 4832343  
> operations in 1 seconds (77317488 bytes)
> [  154.752525] test 1 (128 bit key, 64 byte blocks): 2049979  
> operations in 1 seconds (131198656 bytes)
> [  155.755195] test 2 (128 bit key, 256 byte blocks): 620439  
> operations in 1 seconds (158832384 bytes)
> [  156.761694] test 3 (128 bit key, 1024 byte blocks): 173900  
> operations in 1 seconds (178073600 bytes)
> [  157.768282] test 4 (128 bit key, 8192 byte blocks): 22366  
> operations in 1 seconds (183222272 bytes)
> [  158.774815] test 5 (192 bit key, 16 byte blocks): 4850741  
> operations in 1 seconds (77611856 bytes)
> [  159.781498] test 6 (192 bit key, 64 byte blocks): 2046772  
> operations in 1 seconds (130993408 bytes)
> [  160.788163] test 7 (192 bit key, 256 byte blocks): 619915  
> operations in 1 seconds (158698240 bytes)
> [  161.794636] test 8 (192 bit key, 1024 byte blocks): 173442  
> operations in 1 seconds (177604608 bytes)
> [  162.801242] test 9 (192 bit key, 8192 byte blocks): 22083  
> operations in 1 seconds (180903936 bytes)
> [  163.807793] test 10 (256 bit key, 16 byte blocks): 4862951  
> operations in 1 seconds (77807216 bytes)
> [  164.814449] test 11 (256 bit key, 64 byte blocks): 2050036  
> operations in 1 seconds (131202304 bytes)
> [  165.821121] test 12 (256 bit key, 256 byte blocks): 620349  
> operations in 1 seconds (158809344 bytes)
> [  166.827621] test 13 (256 bit key, 1024 byte blocks): 173917  
> operations in 1 seconds (178091008 bytes)
> [  167.834218] test 14 (256 bit key, 8192 byte blocks): 22362  
> operations in 1 seconds (183189504 bytes)
> [  168.840798]
> [  168.840798] testing speed of async ecb(twofish) decryption
> [  168.849968] test 0 (128 bit key, 16 byte blocks): 4889899  
> operations in 1 seconds (78238384 bytes)
> [  169.855439] test 1 (128 bit key, 64 byte blocks): 2052293  
> operations in 1 seconds (131346752 bytes)
> [  170.862113] test 2 (128 bit key, 256 byte blocks): 616979  
> operations in 1 seconds (157946624 bytes)
> [  171.868631] test 3 (128 bit key, 1024 byte blocks): 172773  
> operations in 1 seconds (176919552 bytes)
> [  172.875244] test 4 (128 bit key, 8192 byte blocks): 22224  
> operations in 1 seconds (182059008 bytes)
> [  173.881777] test 5 (192 bit key, 16 byte blocks): 4893653  
> operations in 1 seconds (78298448 bytes)
> [  174.888451] test 6 (192 bit key, 64 byte blocks): 2048078  
> operations in 1 seconds (131076992 bytes)
> [  175.895131] test 7 (192 bit key, 256 byte blocks): 619204  
> operations in 1 seconds (158516224 bytes)
> [  176.901651] test 8 (192 bit key, 1024 byte blocks): 172569  
> operations in 1 seconds (176710656 bytes)
> [  177.908253] test 9 (192 bit key, 8192 byte blocks): 21888  
> operations in 1 seconds (179306496 bytes)
> [  178.914781] test 10 (256 bit key, 16 byte blocks): 4921751  
> operations in 1 seconds (78748016 bytes)
> [  179.917481] test 11 (256 bit key, 64 byte blocks): 2051219  
> operations in 1 seconds (131278016 bytes)
> [  180.920147] test 12 (256 bit key, 256 byte blocks): 618536  
> operations in 1 seconds (158345216 bytes)
> [  181.926637] test 13 (256 bit key, 1024 byte blocks): 172886  
> operations in 1 seconds (177035264 bytes)
> [  182.933249] test 14 (256 bit key, 8192 byte blocks): 22222  
> operations in 1 seconds (182042624 bytes)
> [  183.939803]
> [  183.939803] testing speed of async cbc(twofish) encryption
> [  183.953902] test 0 (128 bit key, 16 byte blocks): 5195403  
> operations in 1 seconds (83126448 bytes)
> [  184.962487] test 1 (128 bit key, 64 byte blocks): 1912010  
> operations in 1 seconds (122368640 bytes)
> [  185.969150] test 2 (128 bit key, 256 byte blocks): 540125  
> operations in 1 seconds (138272000 bytes)
> [  186.975650] test 3 (128 bit key, 1024 byte blocks): 140631  
> operations in 1 seconds (144006144 bytes)
> [  187.982411] test 4 (128 bit key, 8192 byte blocks): 17737  
> operations in 1 seconds (145301504 bytes)
> [  188.988782] test 5 (192 bit key, 16 byte blocks): 5182287  
> operations in 1 seconds (82916592 bytes)
> [  189.995435] test 6 (192 bit key, 64 byte blocks): 1912356  
> operations in 1 seconds (122390784 bytes)
> [  191.002093] test 7 (192 bit key, 256 byte blocks): 540991  
> operations in 1 seconds (138493696 bytes)
> [  192.008600] test 8 (192 bit key, 1024 byte blocks): 140791  
> operations in 1 seconds (144169984 bytes)
> [  193.015197] test 9 (192 bit key, 8192 byte blocks): 17609  
> operations in 1 seconds (144252928 bytes)
> [  194.021740] test 10 (256 bit key, 16 byte blocks): 5191521  
> operations in 1 seconds (83064336 bytes)
> [  195.028534] test 11 (256 bit key, 64 byte blocks): 1906226  
> operations in 1 seconds (121998464 bytes)
> [  196.035069] test 12 (256 bit key, 256 byte blocks): 540479  
> operations in 1 seconds (138362624 bytes)
> [  197.041579] test 13 (256 bit key, 1024 byte blocks): 140654  
> operations in 1 seconds (144029696 bytes)
> [  198.048164] test 14 (256 bit key, 8192 byte blocks): 17741  
> operations in 1 seconds (145334272 bytes)
> [  199.054717]
> [  199.054717] testing speed of async cbc(twofish) decryption
> [  199.064019] test 0 (128 bit key, 16 byte blocks): 4783914  
> operations in 1 seconds (76542624 bytes)
> [  200.069414] test 1 (128 bit key, 64 byte blocks): 1954641  
> operations in 1 seconds (125097024 bytes)
> [  201.076079] test 2 (128 bit key, 256 byte blocks): 604230  
> operations in 1 seconds (154682880 bytes)
> [  202.082586] test 3 (128 bit key, 1024 byte blocks): 167613  
> operations in 1 seconds (171635712 bytes)
> [  203.089199] test 4 (128 bit key, 8192 byte blocks): 21451  
> operations in 1 seconds (175726592 bytes)
> [  204.095716] test 5 (192 bit key, 16 byte blocks): 4795759  
> operations in 1 seconds (76732144 bytes)
> [  205.102390] test 6 (192 bit key, 64 byte blocks): 1953134  
> operations in 1 seconds (125000576 bytes)
> [  206.109055] test 7 (192 bit key, 256 byte blocks): 599761  
> operations in 1 seconds (153538816 bytes)
> [  207.115564] test 8 (192 bit key, 1024 byte blocks): 166437  
> operations in 1 seconds (170431488 bytes)
> [  208.122184] test 9 (192 bit key, 8192 byte blocks): 20789  
> operations in 1 seconds (170303488 bytes)
> [  209.128728] test 10 (256 bit key, 16 byte blocks): 4794873  
> operations in 1 seconds (76717968 bytes)
> [  210.135375] test 11 (256 bit key, 64 byte blocks): 1953978  
> operations in 1 seconds (125054592 bytes)
> [  211.142039] test 12 (256 bit key, 256 byte blocks): 604269  
> operations in 1 seconds (154692864 bytes)
> [  212.148556] test 13 (256 bit key, 1024 byte blocks): 167571  
> operations in 1 seconds (171592704 bytes)
> [  213.155143] test 14 (256 bit key, 8192 byte blocks): 21453  
> operations in 1 seconds (175742976 bytes)
> [  214.161698]
> [  214.161698] testing speed of async ctr(twofish) encryption
> [  214.175571] test 0 (128 bit key, 16 byte blocks): 4581950  
> operations in 1 seconds (73311200 bytes)
> [  215.184354] test 1 (128 bit key, 64 byte blocks): 1944709  
> operations in 1 seconds (124461376 bytes)
> [  216.191166] test 2 (128 bit key, 256 byte blocks): 594086  
> operations in 1 seconds (152086016 bytes)
> [  217.197536] test 3 (128 bit key, 1024 byte blocks): 163216  
> operations in 1 seconds (167133184 bytes)
> [  218.204149] test 4 (128 bit key, 8192 byte blocks): 21075  
> operations in 1 seconds (172646400 bytes)
> [  219.210813] test 5 (192 bit key, 16 byte blocks): 4705554  
> operations in 1 seconds (75288864 bytes)
> [  220.217330] test 6 (192 bit key, 64 byte blocks): 1963988  
> operations in 1 seconds (125695232 bytes)
> [  221.224004] test 7 (192 bit key, 256 byte blocks): 581953  
> operations in 1 seconds (148979968 bytes)
> [  222.230513] test 8 (192 bit key, 1024 byte blocks): 162790  
> operations in 1 seconds (166696960 bytes)
> [  223.237126] test 9 (192 bit key, 8192 byte blocks): 20706  
> operations in 1 seconds (169623552 bytes)
> [  224.243642] test 10 (256 bit key, 16 byte blocks): 4437112  
> operations in 1 seconds (70993792 bytes)
> [  225.250324] test 11 (256 bit key, 64 byte blocks): 1963735  
> operations in 1 seconds (125679040 bytes)
> [  226.256990] test 12 (256 bit key, 256 byte blocks): 596765  
> operations in 1 seconds (152771840 bytes)
> [  227.263498] test 13 (256 bit key, 1024 byte blocks): 163385  
> operations in 1 seconds (167306240 bytes)
> [  228.270232] test 14 (256 bit key, 8192 byte blocks): 20950  
> operations in 1 seconds (171622400 bytes)
> [  229.276657]
> [  229.276657] testing speed of async ctr(twofish) decryption
> [  229.285975] test 0 (128 bit key, 16 byte blocks): 4571340  
> operations in 1 seconds (73141440 bytes)
> [  230.291288] test 1 (128 bit key, 64 byte blocks): 1949949  
> operations in 1 seconds (124796736 bytes)
> [  231.297951] test 2 (128 bit key, 256 byte blocks): 591529  
> operations in 1 seconds (151431424 bytes)
> [  232.304470] test 3 (128 bit key, 1024 byte blocks): 163609  
> operations in 1 seconds (167535616 bytes)
> [  233.311073] test 4 (128 bit key, 8192 byte blocks): 20975  
> operations in 1 seconds (171827200 bytes)
> [  234.317581] test 5 (192 bit key, 16 byte blocks): 4639461  
> operations in 1 seconds (74231376 bytes)
> [  235.324307] test 6 (192 bit key, 64 byte blocks): 1963173  
> operations in 1 seconds (125643072 bytes)
> [  236.330929] test 7 (192 bit key, 256 byte blocks): 585030  
> operations in 1 seconds (149767680 bytes)
> [  237.337445] test 8 (192 bit key, 1024 byte blocks): 162872  
> operations in 1 seconds (166780928 bytes)
> [  238.344050] test 9 (192 bit key, 8192 byte blocks): 20728  
> operations in 1 seconds (169803776 bytes)
> [  239.350603] test 10 (256 bit key, 16 byte blocks): 4443427  
> operations in 1 seconds (71094832 bytes)
> [  240.357259] test 11 (256 bit key, 64 byte blocks): 1965011  
> operations in 1 seconds (125760704 bytes)
> [  241.363914] test 12 (256 bit key, 256 byte blocks): 590193  
> operations in 1 seconds (151089408 bytes)
> [  242.370422] test 13 (256 bit key, 1024 byte blocks): 163370  
> operations in 1 seconds (167290880 bytes)
> [  243.377018] test 14 (256 bit key, 8192 byte blocks): 20969  
> operations in 1 seconds (171778048 bytes)
> [  244.383546]
> [  244.383546] testing speed of async lrw(twofish) encryption
> [  244.398118] test 0 (256 bit key, 16 byte blocks): 3582956  
> operations in 1 seconds (57327296 bytes)
> [  245.406230] test 1 (256 bit key, 64 byte blocks): 1618011  
> operations in 1 seconds (103552704 bytes)
> [  246.412911] test 2 (256 bit key, 256 byte blocks): 502411  
> operations in 1 seconds (128617216 bytes)
> [  247.419427] test 3 (256 bit key, 1024 byte blocks): 140501  
> operations in 1 seconds (143873024 bytes)
> [  248.422071] test 4 (256 bit key, 8192 byte blocks): 18166  
> operations in 1 seconds (148815872 bytes)
> [  249.424613] test 5 (320 bit key, 16 byte blocks): 3576354  
> operations in 1 seconds (57221664 bytes)
> [  250.431245] test 6 (320 bit key, 64 byte blocks): 1626817  
> operations in 1 seconds (104116288 bytes)
> [  251.437908] test 7 (320 bit key, 256 byte blocks): 504222  
> operations in 1 seconds (129080832 bytes)
> [  252.444407] test 8 (320 bit key, 1024 byte blocks): 140962  
> operations in 1 seconds (144345088 bytes)
> [  253.451020] test 9 (320 bit key, 8192 byte blocks): 17955  
> operations in 1 seconds (147087360 bytes)
> [  254.457555] test 10 (384 bit key, 16 byte blocks): 3558173  
> operations in 1 seconds (56930768 bytes)
> [  255.464210] test 11 (384 bit key, 64 byte blocks): 1630951  
> operations in 1 seconds (104380864 bytes)
> [  256.470866] test 12 (384 bit key, 256 byte blocks): 504089  
> operations in 1 seconds (129046784 bytes)
> [  257.477383] test 13 (384 bit key, 1024 byte blocks): 141065  
> operations in 1 seconds (144450560 bytes)
> [  258.483979] test 14 (384 bit key, 8192 byte blocks): 18168  
> operations in 1 seconds (148832256 bytes)
> [  259.490542]
> [  259.490542] testing speed of async lrw(twofish) decryption
> [  259.499858] test 0 (256 bit key, 16 byte blocks): 3557489  
> operations in 1 seconds (56919824 bytes)
> [  260.505175] test 1 (256 bit key, 64 byte blocks): 1630277  
> operations in 1 seconds (104337728 bytes)
> [  261.511865] test 2 (256 bit key, 256 byte blocks): 503750  
> operations in 1 seconds (128960000 bytes)
> [  262.518383] test 3 (256 bit key, 1024 byte blocks): 140698  
> operations in 1 seconds (144074752 bytes)
> [  263.524988] test 4 (256 bit key, 8192 byte blocks): 18124  
> operations in 1 seconds (148471808 bytes)
> [  264.531487] test 5 (320 bit key, 16 byte blocks): 3579978  
> operations in 1 seconds (57279648 bytes)
> [  265.538179] test 6 (320 bit key, 64 byte blocks): 1632251  
> operations in 1 seconds (104464064 bytes)
> [  266.544843] test 7 (320 bit key, 256 byte blocks): 502180  
> operations in 1 seconds (128558080 bytes)
> [  267.551350] test 8 (320 bit key, 1024 byte blocks): 139727  
> operations in 1 seconds (143080448 bytes)
> [  268.557964] test 9 (320 bit key, 8192 byte blocks): 17731  
> operations in 1 seconds (145252352 bytes)
> [  269.564481] test 10 (384 bit key, 16 byte blocks): 3570236  
> operations in 1 seconds (57123776 bytes)
> [  270.571162] test 11 (384 bit key, 64 byte blocks): 1623126  
> operations in 1 seconds (103880064 bytes)
> [  271.577828] test 12 (384 bit key, 256 byte blocks): 504857  
> operations in 1 seconds (129243392 bytes)
> [  272.584346] test 13 (384 bit key, 1024 byte blocks): 140801  
> operations in 1 seconds (144180224 bytes)
> [  273.586961] test 14 (384 bit key, 8192 byte blocks): 18139  
> operations in 1 seconds (148594688 bytes)
> [  274.589525]
> [  274.589525] testing speed of async xts(twofish) encryption
> [  274.603741] test 0 (256 bit key, 16 byte blocks): 3098851  
> operations in 1 seconds (49581616 bytes)
> [  275.612164] test 1 (256 bit key, 64 byte blocks): 1577161  
> operations in 1 seconds (100938304 bytes)
> [  276.618836] test 2 (256 bit key, 256 byte blocks): 525612  
> operations in 1 seconds (134556672 bytes)
> [  277.625459] test 3 (256 bit key, 1024 byte blocks): 150507  
> operations in 1 seconds (154119168 bytes)
> [  278.632105] test 4 (256 bit key, 8192 byte blocks): 19633  
> operations in 1 seconds (160833536 bytes)
> [  279.638587] test 5 (384 bit key, 16 byte blocks): 3092237  
> operations in 1 seconds (49475792 bytes)
> [  280.645261] test 6 (384 bit key, 64 byte blocks): 1576545  
> operations in 1 seconds (100898880 bytes)
> [  281.651795] test 7 (384 bit key, 256 byte blocks): 526516  
> operations in 1 seconds (134788096 bytes)
> [  282.658305] test 8 (384 bit key, 1024 byte blocks): 150782  
> operations in 1 seconds (154400768 bytes)
> [  283.664935] test 9 (384 bit key, 8192 byte blocks): 19632  
> operations in 1 seconds (160825344 bytes)
> [  284.671425] test 10 (512 bit key, 16 byte blocks): 3164770  
> operations in 1 seconds (50636320 bytes)
> [  285.678254] test 11 (512 bit key, 64 byte blocks): 1586822  
> operations in 1 seconds (101556608 bytes)
> [  286.684781] test 12 (512 bit key, 256 byte blocks): 527705  
> operations in 1 seconds (135092480 bytes)
> [  287.691290] test 13 (512 bit key, 1024 byte blocks): 150918  
> operations in 1 seconds (154540032 bytes)
> [  288.697885] test 14 (512 bit key, 8192 byte blocks): 19640  
> operations in 1 seconds (160890880 bytes)
> [  289.704422]
> [  289.704422] testing speed of async xts(twofish) decryption
> [  289.713733] test 0 (256 bit key, 16 byte blocks): 3082480  
> operations in 1 seconds (49319680 bytes)
> [  290.719098] test 1 (256 bit key, 64 byte blocks): 1571464  
> operations in 1 seconds (100573696 bytes)
> [  291.725752] test 2 (256 bit key, 256 byte blocks): 528360  
> operations in 1 seconds (135260160 bytes)
> [  292.732271] test 3 (256 bit key, 1024 byte blocks): 150115  
> operations in 1 seconds (153717760 bytes)
> [  293.738874] test 4 (256 bit key, 8192 byte blocks): 19513  
> operations in 1 seconds (159850496 bytes)
> [  294.745427] test 5 (384 bit key, 16 byte blocks): 3087055  
> operations in 1 seconds (49392880 bytes)
> [  295.752083] test 6 (384 bit key, 64 byte blocks): 1572391  
> operations in 1 seconds (100633024 bytes)
> [  296.754760] test 7 (384 bit key, 256 byte blocks): 527241  
> operations in 1 seconds (134973696 bytes)
> [  297.757259] test 8 (384 bit key, 1024 byte blocks): 150210  
> operations in 1 seconds (153815040 bytes)
> [  298.763871] test 9 (384 bit key, 8192 byte blocks): 19504  
> operations in 1 seconds (159776768 bytes)
> [  299.770425] test 10 (512 bit key, 16 byte blocks): 3157185  
> operations in 1 seconds (50514960 bytes)
> [  300.777072] test 11 (512 bit key, 64 byte blocks): 1579551  
> operations in 1 seconds (101091264 bytes)
> [  301.783745] test 12 (512 bit key, 256 byte blocks): 526692  
> operations in 1 seconds (134833152 bytes)
> [  302.790244] test 13 (512 bit key, 1024 byte blocks): 150220  
> operations in 1 seconds (153825280 bytes)
> [  303.796840] test 14 (512 bit key, 8192 byte blocks): 19498  
> operations in 1 seconds (159727616 bytes)
>
> --
> Regards/Gruss,
> Boris.
>
>



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists