SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
Polter
Posts: 2526
Joined: April 29th, 2008, 4:55 pm

Requests for micro-help

March 15th, 2012, 10:41 pm

26.5 in N3337: http://www.open-std.org/jtc1/sc22/wg21/ ... 3337.pdf// N3337 is the latest C++ draft specification. The only changes since N3291 are editorial.There's no variate_generator, but at least the bind(d,e) method (mentioned in another thread) is supported by intent:QuoteIn addition to a few utilities, four categories of entities are described: uniform random number generators,random number engines, random number engine adaptors, and random number distributions. These catego-rizations are applicable to types that satisfy the corresponding requirements, to objects instantiated fromsuch types, and to templates producing such types when instantiated. [Note: These entities are specified insuch a way as to permit the binding of any uniform random number generator object e as the argument toany random number distribution object d, thus producing a zero-argument function object such as given bybind(d,e). -- end note ] If I understand correctly, the method described above "as given by bind(d,e)" is d(g) in Table 118.
Last edited by Polter on March 14th, 2012, 11:00 pm, edited 1 time in total.
 
User avatar
ExSan
Posts: 4548
Joined: April 12th, 2003, 10:40 am

Requests for micro-help

March 15th, 2012, 10:57 pm

First page------> Date: 2012-01-16 it is an updated document !
 
User avatar
Polter
Posts: 2526
Joined: April 29th, 2008, 4:55 pm

Requests for micro-help

March 15th, 2012, 11:10 pm

outrun, what are you trying to achieve?If you're trying to get uniform random numbers in a given range, why not use [rand.dist.uni]? If you're trying to obtain a random number engine that changes the values of an arbitrary existing random engine according to a given algo, as in adapting it, perhaps [rand.req.adapt] is what you're looking for:QuoteA random number engine adaptor (commonly shortened to adaptor) a of type A is a random number enginethat takes values produced by some other random number engine, and applies an algorithm to those valuesin order to deliver a sequence of values with different randomness properties.BTW, one more doc: http://en.cppreference.com/w/cpp/numeric/random
Last edited by Polter on March 15th, 2012, 11:00 pm, edited 1 time in total.
 
User avatar
tomodachi
Posts: 17
Joined: February 11th, 2012, 8:37 pm

Requests for micro-help

March 16th, 2012, 2:52 am

QuoteOriginally posted by: outrunCan someone find good docs on the new C++ <random> interface / standard.specifically the* random engine * distributions class interface (constructors, methods). *and*what way's can we combine engines with distribution to generate samples? In boost we have variate_generator, but that will probably not be.Thanks!That is puzzling to me too. I can't make sense of Stroustoup's FAQ. In his section on Random number generation he uses bind, just like you do (except with uniform initialization).What are the perfomance consequences of the various choices? Dr. Cuch seems to be the only one contributing code lately that can be downloaded, compiled and tested.I'm just glad everyone is back on track and getting along. Computers never get mad, they just do exactly what you tell them to do. I hate that.
 
User avatar
Cuchulainn
Posts: 59713
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Requests for micro-help

March 16th, 2012, 8:56 am

We can use RNG at various levels, e.g. variate_generator() + writing your own seeds and so on.Let's take the role of someone like an 'electrician' (like myself) who just want to wire up a RNG into a simulator. All I need is double getRNGF() and all nitty-gritty is taken care of by the supplier of the RNG module. Defaults are fine for the moment; Of course, I can open Padora's box and customise to suit special needs. It's a
Last edited by Cuchulainn on March 15th, 2012, 11:00 pm, edited 1 time in total.
 
User avatar
Cuchulainn
Posts: 59713
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Requests for micro-help

March 16th, 2012, 8:59 am

Quote[...] seems to be the only one contributing code lately that can be downloaded, compiled and tested.It was Outrun who delivered this project into the world. So, let's not forget that.
 
User avatar
tomodachi
Posts: 17
Joined: February 11th, 2012, 8:37 pm

Requests for micro-help

March 16th, 2012, 11:31 am

QuoteOriginally posted by: outrunthanks tomodachi, I'll check the FAQ too: good pointer!I don't get your "I'm just glad everyone is back on track and getting along." nor your "Dr Cuch seems to be the onyl one..." (just check the svn to see the activity). I think that's a bit trying to polarizing things, no need for that. You're not by any change Mr. K.L., right?I am Mr. K.L! How did you know that?It seems I've stepped on some sort of live wire. All I can say is I meant what I said and am honestly a little puzzled by what I am seeing. Let's get back to coding!http://qfcl.wilmott.com/p/sandbox/timeline/all/ does not seem very active. I see one other person contributing to http://qfcl.wilmott.com/p/random/timeline/all/ recently. Is there other activity I am missing? I would like to add Dr. Cuch's contribution together with the vs9 build from SJ and my vs10 build. I applied for permission to do that, but have not heard back yet. You seem to have some pull around here. How do we make that happen?
 
User avatar
Cuchulainn
Posts: 59713
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Requests for micro-help

March 16th, 2012, 12:56 pm

QuoteOriginally posted by: outrunGood idea. I haven't seen your request. Have you created an account?Then the use of the svn:1) Is the idea to create a dir in the sandbox where the three of you can work together? 2) does Cuch want to create a new project, add a new line in this list? Is so then he should come up with a name and description. Cuch would be project owner, and you contributors?this choice needs to be made first by cuch.I think a project is best as you suggest. So, an MC and PDE dirs that people can add to?Sounds fine by me. Only issue is that I use top-down approach to implementing loosely coupled components. If someone comes up with an even better design (with working code) that's 100% fine by me. But software developers seem to like Boost Function big time. That seems to be the message. But this is well known by now. It works, and qfcl can help improve it. BTW I have a upcoming MC103 that I would like to use with Milo's RNG. What do you think.BTW any PDE/FDM folk out there?
Last edited by Cuchulainn on March 15th, 2012, 11:00 pm, edited 1 time in total.
ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...


GZIP: On