Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 6 Oct 2006 09:55:02 -0400
From:      "Bucky Jordan" <bjordan@lumeta.com>
To:        <soralx@cydem.org>, <freebsd-hardware@freebsd.org>
Cc:        dking@ketralnis.com
Subject:   RE: Quiet computer
Message-ID:  <78ED28FACE63744386D68D8A9D1CF5D4209C44@MAIL.corp.lumeta.com>
In-Reply-To: <200610052145.54292.soralx@cydem.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Here's openssl speed for the Woodcrest (Intel 5160 3 Ghz, dual core, 4mb
shared cache per socket) if anyone's interested. (FreeBSD 6.1 amd64,
this is a Dell Poweredge 2950)

OpenSSL 0.9.7e-p1 25 Oct 2004
built on: Sun May  7 02:04:05 UTC 2006
options:bn(64,64) md2(int) rc4(ptr,int) des(ptr,risc2,4,int)
aes(partial) blowfish(idx)
compiler: cc
available timing options: USE_TOD HZ=3D128 [sysconf value]
timing function used: getrusage
The 'numbers' are in 1000s of bytes per second processed.
type             16 bytes     64 bytes    256 bytes   1024 bytes   8192
bytes
md2               1982.25k     4223.76k     5887.59k     6529.65k
6744.10k
mdc2              9192.48k    10607.07k    11026.26k    11144.61k
11178.08k
md4              21764.77k    73114.34k   200145.53k   353657.49k
453376.98k
md5              18431.16k    58731.43k   146746.50k   234291.88k
283300.29k
hmac(md5)        20466.63k    62714.97k   152306.47k   238210.06k
284042.86k
sha1             19202.25k    55284.85k   125863.44k   184984.41k
213893.52k
rmd160           14812.46k    39928.95k    81317.52k   109759.01k
122228.05k
rc4             329572.77k   341938.54k   346376.43k   399439.55k
404077.57k
des cbc          55494.37k    58391.56k    59583.38k    59580.46k
59940.61k
des ede3         21785.29k    22214.20k    22248.75k    22240.27k
22346.94k
idea cbc             0.00         0.00         0.00         0.00
0.00
rc2 cbc          28855.43k    29362.38k    29437.11k    29400.29k
29410.44k
rc5-32/12 cbc   137268.12k   146253.58k   150532.22k   150499.88k
151679.10k
blowfish cbc     96832.32k   102509.13k   103935.48k   104272.80k
104929.04k
cast cbc         75312.95k    79105.50k    80213.31k    80461.91k
80519.39k
aes-128 cbc     152031.80k   156145.92k   157391.51k   157771.15k
158187.02k
aes-192 cbc     133868.42k   137744.69k   138928.19k   139084.59k
139197.70k
aes-256 cbc     118072.93k   122630.84k   123532.91k   123685.57k
123870.41k
                  sign    verify    sign/s verify/s
rsa  512 bits   0.0002s   0.0000s   5052.7  59340.8
rsa 1024 bits   0.0007s   0.0000s   1523.3  23025.8
rsa 2048 bits   0.0037s   0.0001s    269.3   8265.9
rsa 4096 bits   0.0245s   0.0004s     40.8   2540.3
                  sign    verify    sign/s verify/s
dsa  512 bits   0.0001s   0.0001s   8161.5   7012.2
dsa 1024 bits   0.0003s   0.0004s   3087.2   2595.8
dsa 2048 bits   0.0010s   0.0012s    957.8    807.8


> -----Original Message-----
> From: owner-freebsd-hardware@freebsd.org [mailto:owner-freebsd-
> hardware@freebsd.org] On Behalf Of soralx@cydem.org
> Sent: Friday, October 06, 2006 12:46 AM
> To: freebsd-hardware@freebsd.org
> Cc: dking@ketralnis.com
> Subject: Re: Quiet computer
>=20
>=20
> > I'll try to do a ubench tonight.
> >
> > I'd love to test the speed of the crypto, but I'm not exactly sure
> > how, nor even test if it's actually being used e.g. by openssl. All
> > I've confirmed is whether /dev/crypto exists :) I'd like to make
sure
> > that at the very least, Apache from ports, and OpenSSH and OpenSSL
in
> > the base system are using it, and ideally that OpenSSL from ports is
> > using it too. I'm working on getting IPsec up and running, and I
have
> > FAST_IPSEC in the kernel, so it *should* use it, but again, I can't
> > think of an easy way to confirm this other than watching the device
> > node for opens/reads/writes
>=20
> just do `openssl speed`
> if you see some damn large numbers, you know 'tis hw crypto working :)
> also, test the speed of the chip's true RNG
>=20
> here are some numbers for comparison (Intel Northwood 2.4GHz):
>=20
> OpenSSL 0.9.7e 25 Oct 2004
> built on: Fri Mar 24 14:39:51 PST 2006
> options:bn(64,32) md2(int) rc4(idx,int) des(ptr,risc1,16,long)
> aes(partial) blowfish(idx)
> compiler: cc
> available timing options: USE_TOD HZ=3D128 [sysconf value]
> timing function used: getrusage
> The 'numbers' are in 1000s of bytes per second processed.
> type             16 bytes     64 bytes    256 bytes   1024 bytes
8192
> bytes
> md2               1083.57k     2403.96k     3470.45k     3857.17k
> 4012.49k
> mdc2              3369.63k     4140.27k     4382.06k     4442.89k
> 4504.70k
> md4               7496.84k    25970.13k    77299.58k   147729.23k
> 193925.60k
> md5               6296.26k    21208.61k    60141.29k   103789.67k
> 141599.07k
> hmac(md5)         8946.89k    27790.12k    72816.20k   114649.47k
> 143806.72k
> sha1              6163.63k    19126.65k    44285.72k    68503.02k
> 82858.09k
> rmd160            4775.30k    13618.17k    29234.02k    40202.74k
> 46639.83k
> rc4              79872.90k    87853.64k    90752.84k    91046.33k
> 90689.96k
> des cbc          40637.00k    40933.07k    41313.21k    40968.95k
> 40949.12k
> des ede3         14860.39k    15206.22k    15465.58k    15398.36k
> 15453.31k
> idea cbc             0.00         0.00         0.00         0.00
> 0.00
> rc2 cbc          10668.86k    10772.21k    10747.81k    10773.98k
> 10643.61k
> rc5-32/12 cbc    85925.70k    85319.97k    85797.86k    85211.97k
> 85898.78k
> blowfish cbc     74931.29k    81573.23k    82962.56k    83240.87k
> 81787.52k
> cast cbc         54535.14k    58286.62k    59041.53k    58917.05k
> 58862.35k
> aes-128 cbc      48317.69k    45316.48k    45234.67k    45561.34k
> 45518.08k
> aes-192 cbc      42265.50k    39788.11k    40072.84k    40091.84k
> 39944.97k
> aes-256 cbc      37967.32k    35824.60k    35768.69k    35125.65k
> 35716.55k
>                   sign    verify    sign/s verify/s
> rsa  512 bits   0.0012s   0.0001s    868.2   8914.5
> rsa 1024 bits   0.0056s   0.0003s    178.3   3244.6
> rsa 2048 bits   0.0332s   0.0010s     30.1   1011.1
> rsa 4096 bits   0.2197s   0.0034s      4.6    295.3
>                   sign    verify    sign/s verify/s
> dsa  512 bits   0.0009s   0.0011s   1056.9    893.7
> dsa 1024 bits   0.0028s   0.0034s    351.9    294.6
> dsa 2048 bits   0.0092s   0.0113s    109.0     88.6
>=20
>=20
> > > You might wants to consider using a low-power 'brick'-type AC->DC
> > > PSU (~90W?) and a DC-DC voltage converter that plugs directly into
> > > a mainboard's power connector. This will be quite noiseless.
> >
> > Do you have any recommendations? Ideally it would mount on the case
> > (<http://www.logicsupply.com/product_info.php/products_id/134>) like
> > the current power supply does but I'm open to other options
>=20
> something like this (random example):
> http://www.logicsupply.com/product_info.php/products_id/596
>=20
> check your mainboard's manual -- I'm not sure if its 12V current
> requrement is less than this power module can supply (remember,
> in contrast to Pentium IV, C3 feeds off 5V, I believe).
>=20
> you might want to get higher power version of this for greater
> efficiency or if you use more than one slow HDD (unless they're
> laptop HDDs that operate at 5V). I wouldn't use the highest-power
> modules meant for fast processors (such as Opteron, etc), as
> they're probably operating at high switching speeds of 2-4 MHz
> (holy crap!) =3D=3D unnecessary EMR [acoustic noise isn't the only
> bad noise to be worried about ;)]. But that's just my opinion
> (completely unsupported).
>=20
> Note that I haven't got a slightest idea on the reliability of this
> setup, although I see less failure modes for the power module than
> ATX PSU. As usual, no guarantees: it may behave badly from day 1,
> drink all your beer, ride your bike, flush the toilet while you're
> showering, and it will surely bite your wallet :)
>=20
> if you're good friends with your hands, then it should be very
> easy to integrate a 'brick' power supply into the case (and get
> rid of that noisey ATX PSU) :)
>=20
> [SorAlx]  ridin' VN1500-B2
> _______________________________________________
> freebsd-hardware@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-hardware
> To unsubscribe, send any mail to "freebsd-hardware-
> unsubscribe@freebsd.org"



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?78ED28FACE63744386D68D8A9D1CF5D4209C44>