headers
zooko | 1 Feb 2008 15:38
Gravatar

add Poly1305?


Dear maintainers of Crypto++:

Would there be any interest in adding Poly1305 MAC?  I was just  
looking at DJB's timings.html [1], which shows that while VMAC is  
faster than Poly1305 on amd64 (e.g. [2] vs [3]), VMAC is slower than  
Poly1305 on the same machine in 32-bit mode ([4] vs [5]),
VMAC is slower than Poly1305 on an UltraSPARC III ([6] vs [7]) and  
VMAC is dramatically slower than Poly1305 on an old PowerPC G4 ([8]  
vs [9]).

This makes me interested in having Poly1305 available within the  
Crypto++ class framework.

Regards,

Zooko

[1] http://cr.yp.to/streamciphers/timings.html
[2] http://cr.yp.to/streamciphers/timings/graphs/katana:aes-256- 
vmac128:256,128:2048.png
[3] http://cr.yp.to/streamciphers/timings/graphs/katana:aes-256- 
poly1305:256,128:2048.png
[4] http://cr.yp.to/streamciphers/timings/graphs/katana-x86:aes-256- 
vmac128:256,128:2048.png
[5] http://cr.yp.to/streamciphers/timings/graphs/katana-x86:aes-256- 
poly1305:256,128:2048.png
[6] http://cr.yp.to/streamciphers/timings/graphs/icarus:aes-256- 
vmac128:256,128:2048.png
[7] http://cr.yp.to/streamciphers/timings/graphs/icarus:aes-256-
(Continue reading)

Permalink | Reply | 6 comments |
headers
Wei Dai | 2 Feb 2008 02:21

Re: add Poly1305?


DJB used an old implementation of VMAC in those timings, which wasn't 
optimized for 32-bit platforms at all. The fastest VMAC implementation 
currently available is in Crypto++ 5.5 and later. If you compare 
http://www.cryptopp.com/benchmarks-p4.html and 
http://groups.google.com/group/sci.crypt/msg/622a03847bc41f9b, it's faster 
than Poly1305 on Pentium 4. I don't have a direct comparison on other 32-bit 
platforms, but I think VMAC should be fairly competitive.

--------------------------------------------------
From: "zooko" <zooko <at> zooko.com>
Sent: Friday, February 01, 2008 6:38 AM
To: "Crypto++ Users" <cryptopp-users <at> googlegroups.com>
Subject: add Poly1305?

>
> Dear maintainers of Crypto++:
>
> Would there be any interest in adding Poly1305 MAC?  I was just
> looking at DJB's timings.html [1], which shows that while VMAC is
> faster than Poly1305 on amd64 (e.g. [2] vs [3]), VMAC is slower than
> Poly1305 on the same machine in 32-bit mode ([4] vs [5]),
> VMAC is slower than Poly1305 on an UltraSPARC III ([6] vs [7]) and
> VMAC is dramatically slower than Poly1305 on an old PowerPC G4 ([8]
> vs [9]).
>
> This makes me interested in having Poly1305 available within the
> Crypto++ class framework.
>
> Regards,
(Continue reading)

Permalink | Reply | 5 comments |
headers
Wei Dai | 2 Feb 2008 03:15

Re: hashing with a key


HMAC should be at least as strong as the underlying hash function, but 
apparently it isn't much stronger. See the papers listed at 
http://www.cryptolounge.org/wiki/HMAC.

For the highest security, I suggest using HMAC/SHA-512 or HMAC/Whirlpool.

Do not hash the HMAC output with standard SHA. That doesn't help at all.

--------------------------------------------------
From: "mihailo" <mihailovelimirovic <at> gmail.com>
Sent: Wednesday, January 30, 2008 10:38 AM
To: "Crypto++ Users" <cryptopp-users <at> googlegroups.com>
Subject: Re: hashing with a key

>
> Of course, it is possible to hash the result with the standard SHA
> algorithm. Or, it is also possible to encrypt data with the given key
> and use hash over encrypted data.
>
> All in all, problem is solved.
>
> Thanks for useful advice.
>
> On Jan 30, 5:59 pm, mihailo <mihailovelimiro... <at> gmail.com> wrote:
>> Thanks, Jeffrey.
>>
>> Will it be much easier to crack than SHA-1 or SHA-2 ?
>>
>> On Jan 30, 5:45 pm, "Jeffrey Walton" <noloa... <at> gmail.com> wrote:
(Continue reading)

Permalink | Reply | 0 comments |
headers
Wei Dai | 2 Feb 2008 03:47

Re: New to Visual Studio 2005


I can't figure out what the problem might be from your error log. I haven't 
seen this myself, and since no one else has reported it, I think it may be 
something specific to your machine. Try compiling Crypto++ on another 
machine and see if that works. Also, you may be able to work around this by 
defining _HAS_ITERATOR_DEBUGGING to 0 in the Crypto++ project files. See 
http://msdn2.microsoft.com/en-us/library/aa985939(VS.80).aspx.

--------------------------------------------------
From: "TimB" <tbergmann <at> yahoo.com>
Sent: Wednesday, January 30, 2008 6:42 AM
To: "Crypto++ Users" <cryptopp-users <at> googlegroups.com>
Subject: New to Visual Studio 2005

>
> I installed MSVC 2005 SP1 with Crypto++ 5.5.2 and first attempt to
> compile without any changes. I am getting the following output errors.
> The problem still eludes me. It must be something simple some of you
> may have run across.
>
>
> Output file:
>
> ------ Rebuild All started: Project: cryptlib, Configuration: Debug
> Win32 ------
> Deleting intermediate and output files for project 'cryptlib',
> configuration 'Debug|Win32'
> Performing Custom Build Step
> Compiling...
> pch.cpp
(Continue reading)

Permalink | Reply | 0 comments |
headers
Wei Dai | 2 Feb 2008 09:42

Re: Bug in AdditiveCipherTemplate


Thanks for reporting this. I've checked in a fix to SVN.

--------------------------------------------------
From: "ASBai" <baiyang <at> gmail.com>
Sent: Thursday, January 31, 2008 2:12 PM
To: "Crypto++ Users" <cryptopp-users <at> googlegroups.com>
Subject: Bug in AdditiveCipherTemplate

>
> IN  CTR or OFB mode, AdditiveCipherTemplate's IsRandomAccess method
> will cause an infinite recusive call.
> > 

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the "Crypto++ Users" Google Group.
To unsubscribe, send an email to cryptopp-users-unsubscribe <at> googlegroups.com.
More information about Crypto++ and this group is available at http://www.cryptopp.com.
-~----------~----~----~----~------~----~------~--~---
Permalink | Reply | 0 comments |
headers
litle | 3 Feb 2008 17:04
Picon

crypto++ in VisualC++ MFC


how can i  integrate crypto++ with MFC in VisualC++ 6.0 version
I am doing a project  with it so i need a immediate help for this
manual

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the "Crypto++ Users" Google Group.
To unsubscribe, send an email to cryptopp-users-unsubscribe <at> googlegroups.com.
More information about Crypto++ and this group is available at http://www.cryptopp.com.
-~----------~----~----~----~------~----~------~--~---
Permalink | Reply | 2 comments |
headers
Will | 3 Feb 2008 20:38
Picon

AIX xlc++ compiler for Crypto++ ?


Is there anyone ever successfully compile this library on AIX with xlc+
+ compiler ? I got a lot of troubles on trying this. Looks like it is
difficult to do this in a short time.
It will be very appreciated if someone can give me any inforamtion on
this.  Thanks.

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the "Crypto++ Users" Google Group.
To unsubscribe, send an email to cryptopp-users-unsubscribe <at> googlegroups.com.
More information about Crypto++ and this group is available at http://www.cryptopp.com.
-~----------~----~----~----~------~----~------~--~---
Permalink | Reply | 2 comments |
headers
StefSchultz | 5 Feb 2008 07:28
Picon
Favicon

CFB Mode - different results


Testing the CFB Mode I've noticed that it gives me some strange
results. I tested the same mode with the same IV and Key in a PHP and
Python script:

//----------------------------------
//init key & iv
$key = "xxxxxxxxxxxxxxxxxxxxxxxx";
$iv = "xxxxxxxxxxxxxxxx";

//print key
echo ('key:           ' . $key . "\n");
echo ('iv:            ' . $iv . "\n");

//print iv
echo ('----------------------------------------' . "\n");

//original
$original = 'hello';
echo ('original:      ' . $original . "\n");

//encode
$encoded = mcrypt_cfb(MCRYPT_RIJNDAEL_128, $key, $original,
MCRYPT_ENCRYPT, $iv);
echo ('encoded (hex): ' . str2hex($encoded) . "\n");

//decode
$decoded = mcrypt_cfb(MCRYPT_RIJNDAEL_128, $key, $encoded,
MCRYPT_DECRYPT, $iv);
echo ('decoded:       ' . $decoded . "\n");
(Continue reading)

Permalink | Reply | 2 comments |
headers
Bala | 5 Feb 2008 10:07
Picon

Re: AIX xlc++ compiler for Crypto++ ?


I think a lot of people are facing difficulty in compiling cryptopp in
AIX. It would be of great help if cryptopp developers can take a look
at it.

Though i use in solaris and windows, i have commented out in AIX just
because i couldnt compile in AIX.

On Feb 4, 12:38 am, Will <wwen... <at> gmail.com> wrote:
> Is there anyone ever successfully compile this library on AIX with xlc+
> + compiler ? I got a lot of troubles on trying this. Looks like it is
> difficult to do this in a short time.
> It will be very appreciated if someone can give me any inforamtion on
> this.  Thanks.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the "Crypto++ Users" Google Group.
To unsubscribe, send an email to cryptopp-users-unsubscribe <at> googlegroups.com.
More information about Crypto++ and this group is available at http://www.cryptopp.com.
-~----------~----~----~----~------~----~------~--~---
Permalink | Reply | 1 comment |
headers
Jeffrey Walton | 5 Feb 2008 20:23
Picon

Re: CFB Mode - different results


Hi Stef,

> stfEncryptor.Put(reinterpret_cast<const unsigned
>    char*>(plaintext.c_str()), plaintext.length() + 1);
Do the scripts include the trailing '\0' of 'Hello'?

Jeff

On 2/5/08, StefSchultz <stef.schultz <at> yahoo.com> wrote:
>
> Testing the CFB Mode I've noticed that it gives me some strange
> results. I tested the same mode with the same IV and Key in a PHP and
> Python script:
>
> //----------------------------------
> //init key & iv
> $key = "xxxxxxxxxxxxxxxxxxxxxxxx";
> $iv = "xxxxxxxxxxxxxxxx";
>
> //print key
> echo ('key:           ' . $key . "\n");
> echo ('iv:            ' . $iv . "\n");
>
> //print iv
> echo ('----------------------------------------' . "\n");
>
> //original
> $original = 'hello';
> echo ('original:      ' . $original . "\n");
(Continue reading)

Permalink | Reply | 1 comment |

Gmane