Re: [viff-devel] Say hello to viff.boost

2010-08-11 Thread Marcel Keller
it are the same. Best regards, Marcel Marcel Keller wrote: Hi friends of VIFF, I've implemented Share and ShareList in C, based on a C implementation of Deferred. Using the C extension, benchmark.py and aes.py show a speed up between 50 and 100 percent. The code is in my repository: http://hg.viff.dk

Re: [viff-devel] Value overflow in Toft07

2010-07-06 Thread Marcel Keller
Dear Lars, thanks for pointing it out. It is now fixed in the official repository. Best regards, Marcel Lars Krapf wrote: Hello VIFF-team I would like to suggest the following patch to viff/comparison.py: 159c159 l = int(self.options.security_parameter +

Re: [viff-devel] VIFF and random numbers

2010-07-06 Thread Marcel Keller
Thomas P Jakobsen wrote: The urandom is os-specific: This function returns random bytes from an OS-specific randomness source. The returned data should be unpredictable enough for cryptographic applications, though its exact quality depends on the OS implementation. On a UNIX-like system this

Re: [viff-devel] VIFF reactor

2010-04-29 Thread Marcel Keller
Hi Joel, Is it still necessary to run `viff.reactor.install()` as described in http://www.mail-archive.com/viff-devel@viff.dk/msg00657.html in order to utilize the VIFF reactor? - If so, would it be possible to fix that? I don't see a good way to that, for the following reasons: - To change

[viff-devel] OrlandiRuntime implementation

2009-11-04 Thread Marcel Keller
Hi Claudio and Jesper, In the code review of the OrlandiRuntime we found two points, we want to discuss with you. Step 3 of the triple generation protocol says: Coin-flip a subset \fancy_T \subset \fancy_M of size \lambda(2d + 1)M. The current implementation loops over the elements in

Re: [viff-devel] OrlandiRuntime implementation

2009-11-04 Thread Marcel Keller
Claudio Orlandi wrote: On Wed, Nov 4, 2009 at 5:46 AM, Marcel Keller mkel...@cs.au.dk wrote: Hi Claudio and Jesper, In the code review of the OrlandiRuntime we found two points, we want to discuss with you. Step 3 of the triple generation protocol says: Coin-flip a subset \fancy_T \subset

Re: [viff-devel] OrlandiRuntime implementation

2009-11-04 Thread Marcel Keller
Claudio Orlandi schrieb: On Wed, Nov 4, 2009 at 10:15 AM, Marcel Keller mkel...@cs.au.dk wrote: Claudio Orlandi wrote: On Wed, Nov 4, 2009 at 5:46 AM, Marcel Keller mkel...@cs.au.dk wrote: Hi Claudio and Jesper, In the code review of the OrlandiRuntime we found two points, we want to discuss

Re: [viff-devel] Noisy preprocessing

2009-10-29 Thread Marcel Keller
make it an option. On 28/10/2009, at 20.01, Marcel Keller wrote: Hi Janus, do you still need the timing output in the update callback in Runtime.preprocess()? It makes benchmarking the usual runtimes very noisy because the update callback is called many times there. Best regards, Marcel

[viff-devel] Noisy preprocessing

2009-10-28 Thread Marcel Keller
Hi Janus, do you still need the timing output in the update callback in Runtime.preprocess()? It makes benchmarking the usual runtimes very noisy because the update callback is called many times there. Best regards, Marcel ___ viff-devel mailing list

[viff-devel] Orlandi preprocessing

2009-10-22 Thread Marcel Keller
Hi Janus, I remember you saying today that the preprocessing in the OrlandiRuntime is more efficient per item the more items are requested. Is that correct? I ask because in my optimizations, I limited the items being preprocessed per call in order to save memory. I would of course drop that

Re: [viff-devel] Optimizing preprocessing

2009-10-21 Thread Marcel Keller
Martin Geisler wrote: Janus Dam Nielsen janus.niel...@alexandra.dk writes: Hi Marcel, I am not opposed to your suggestion. However I would like to point out that in VIFF you compute on shares and not field elements! Well, we've actually made the outer runtime interfaces in such a way that

Re: [viff-devel] AES slides from SPEED-CC

2009-10-16 Thread Marcel Keller
Hi, There are two talks about how to implement AES efficiently, this one http://www.hyperelliptic.org/SPEED/slides09/kasper-aes_speedcc09_slides.pdf describes on slide 9 how one will typically combine SubBytes, ShiftRows, and MixColumns into one operation operating on diagonals. I don't

Re: [viff-devel] Optimizing preprocessing

2009-10-09 Thread Marcel Keller
Hi Janus, I am not opposed to your suggestion. However I would like to point out that in VIFF you compute on shares and not field elements!. Computing directly on the field elements is hacking the abstractions of VIFF. I don't think so. Since I work with VIFF, it treats field elements as

[viff-devel] Say hello to viff.reactor

2009-05-27 Thread Marcel Keller
Hi friends of VIFF, I've now implemented a special reactor for VIFF, based on the standard SelectReactor. The new reactor should make non-trivial programs considerably faster, e.g. computation of 10 AES-blocks in parallel from over 6 seconds to 2.3 seconds per block (3 players, passive

[viff-devel] Non-default reactors

2009-05-18 Thread Marcel Keller
Hi friends of VIFF, Who is working with non-default reactors, such as gtk2reactor? How do you deal with the fact that VIFF might block the reactor for a long time which might render the GUI non-responsive? I'm currently working on a special reactor for VIFF to get more control over the

[viff-devel] How to avoid the boring math

2009-04-24 Thread Marcel Keller
Hi friends of VIFF, I've now played a little bit with artificial bandwidth limits and artificial delays in order to be able to better analyze protocols. (Actually, my motivation was to get the running time for AES matching my analysis. It worked.) I found out that setting a delay of 1

[viff-devel] Two-threaded VIFF

2009-04-21 Thread Marcel Keller
Hi friends of VIFF, I've finally completed the patch for a two-threaded VIFF where most of the VIFF code runs in a separate thread. The patch is against the tip of my repository: http://hg.viff.dk/mkeller It turned out be not so straight-forward as I thought. I had to use a recursion, as in

Re: [viff-devel] Confusing behaviour?

2009-03-23 Thread Marcel Keller
I think the problem is the following: Every players sends its id on a new connection, and the connection is considered to be set up when this id arrives. So it may occur that a player has received the ids from all other players and wants to send its id to other player(s). But nothing is sent

Re: [viff-devel] Mystery of the quadratic running time solved?

2009-03-10 Thread Marcel Keller
Hi Ivan, I just wanted to say that I think it would be great if you would implement a version of your proposed two-threaded solution. I do not have a firm grasp of all the programming details, but it does seem that the overall idea is converging, and that some time soon the best way to judge

Re: [viff-devel] Mystery of the quadratic running time solved?

2009-03-08 Thread Marcel Keller
You're talking about this two-threaded solution as if it is something that exists and will solve all our problems... No, for now, it's just an imagination in my mind, a proposal for the next meeting, and a strong feeling that it's the right way to do it. But I still haven't seen it, and I

Re: [viff-devel] Mystery of the quadratic running time solved?

2009-03-06 Thread Marcel Keller
Indeed we did not know (well I didn't) back then that the data was not sent immediately by Twisted, and I was starting to think yesterday whether the hack would make a difference. Lucky for us, it apparently does :) That is not the only problem. To free the memory of the shares and to send