About the new 'ThreadSelf' keyword: - I am surprised that under Windows the 'thread id' returned by 'ThreadSelf()' has the same value as the 'thread handle' returned by 'ThreadCreate()' (as under Linux). - If this is always true, then why use 2 different naming (the true naming should be 'thread ha...

This is what I am using to randomly choose a key value. Neat, I didn't think about using rdtsc. I was just thinking about this problem. To seed the generators having something that changes every time. For the Squares PRNG seed I was thinking of using below with a default key only because it appears...

That sounds good, adding 'fbmath.bi'

I just noticed there is also:
- 'fbio.bi' for IsRedirected
- 'fbthread.bi' for ThreadDetach

what is this kind of thread-safe RANDOMIZE? There is still only one global state and only one generator can be selected at any one time across all threads. I would expect that the user would want to call the RANDOMIZE function before starting any threads. However, if for some reason RANDOMIZE were ...

Shouldn't the return value be a ulong instead of uinteger? Following-up to this thread and bug ticket #924 RGB() & RGBA() macros return incorrect data-type Proposed is to return the ULONG type for RGB and RGBA macros. Timings depend on back end and optimization levels. For example, - 64-bit gcc...

With commented sub procs lines --> 2 error messages ... When uncommenting sub procs lines --> only 1 error message That bug was tough to find though seems obvious now. parser-quirk-thread.bas:cThreadCallFunc() was using a wrong function to check if the procedure has a return type and was referring ...

With these fast PRNG's, I think the general pattern is; Some state vector S: S = {x0, x1, ... } Some transition function f() to a new state S': S' = f(S) S'' = f(S') = f(f(S)) Maybe some extraction of the pseudo random number with bit-twiddling function g(): rnd := g(S) In either f() or g(), the mod...

Well, I think we are on the right track with this change overall. 'fbmath.bi' needs some work to make it a little more user friendly and some updates to the documentation will help advertise it. fxm, sorry, I didn't mean to turn this in to a development discussion. For sure, we can update the follow...

Sorry about that. Sometime in last months Travis-CI changed the default build environment - previously Ubuntu 14.04 Trusty which uses gcc 4.8.4. - currently Ubuntu 16.04 Xenial which uses gcc 5.4.0 I have only been testing on Ubuntu 16.04 x86_64 lately. An easy fix though and should have it pushed s...

I've rebased and pushed changes to https://github.com/jayrm/fbc/tree/emscripten including the patches posted in this topic from angros47 I tested on windows x86_64 only. One issue and work-around (hack) is below. Prerequisites - git installed - development environment for fbc, rtlib, & gfxlib2 -...

I've merged new changes with https://github.com/freebasic/fbc/pull/267 It adds the new header ./inc/fbmath.bi that has the new PRNG's. I'm hoping that with the source in fb itself there will be contributors from the community that can expand on the very simple starting point. I'll update this thread...

Naming of "rnd64()" looks good to me. I can add xoroshiro128 to fbmath.bi header also. Maybe tonight. At first I was seeing truly mind boggling timings with xoroshiro128. Like it defied reality. ...Evidently, if seeded with {0,0}, seems that gcc optimizes xoroshiro128 out of existence. :) ...

The gcc optimizations didn't work out the way I thought. So I need to regroup and figure out my exit strategy. fbc 1.07 is neither thread-safe nor has rnd32. So my base line test timings are limited. On -O2 production code built-in RND() - it doesn't really matter much if it is gas or gcc because it...

Thanks paul. I'm kind of happy with last rewrite. It has some generic qualities so it will be easy to add in additional generators. But the performance is still bad due the function pointers, by like a factor of 10. namespace fbc extern "rtlib" type RndMSWS32 extends FB_RNDSTATE declare co...

@coderjeff: You can always implement them as header files, and then the user can #include them as they see fit. All you have to do is include it as part of the currently official distribution. No compiler bloat, faster error reporting and correcting, easily patchable and modifiable... I have an ide...