<html> <head> <meta http-equiv="Content-Type" content="text/html; charset=Windows-1252"> <style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style> </head> <body dir="ltr"> <div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);" class="elementToProof"> Mark,</div> <div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);" class="elementToProof"> <br> </div> <div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);" class="elementToProof"> This is very helpful - thank you.</div> <div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);" class="elementToProof"> <br> </div> <div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);" class="elementToProof"> I have now: dropped the grounded implementation; renamed the unique implementation to "mt"; added an array.copy() to ensure that the old state array is not updated; included a derived shared random type, including runif() and set_seed() predicates; and revised the test.</div> <div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);" class="elementToProof"> <br> </div> <div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);" class="elementToProof"> Kindly, Mark.<br> </div> <div id="appendonsend"></div> <hr style="display:inline-block;width:98%" tabindex="-1"> <div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" style="font-size:11pt" color="#000000"><b>From:</b> Mark Brown <mark@mercurylang.org><br> <b>Sent:</b> 20 February 2023 14:32<br> <b>To:</b> Mark Clements <mark.clements@ki.se><br> <b>Cc:</b> users <users@lists.mercurylang.org><br> <b>Subject:</b> Re: [m-users.] Announcement: mercury-mt Mercury only Mersenne Twister random numbers</font> <div> </div> </div> <div class="BodyFragment"><font size="2"><span style="font-size:11pt;"> <div class="PlainText">[You don't often get email from mark@mercurylang.org. Learn why this is important at <a href="https://aka.ms/LearnAboutSenderIdentification">https://aka.ms/LearnAboutSenderIdentification</a> ]<br> <br> Hi Mark,<br> <br> As Volker said, thanks for sharing your work. In regards to your question,<br> <br> > Question for Mercury users: is it better to do this as a ground generator or as a unique generator?<br> <br> Short answer: it's generally best for a library to supply as many<br> instances as it can correctly implement. For this application it<br> would be better to implement it as a unique generator, and use<br> adapters to make other instances as needed.<br> <br> Long answer: the choice for the user is similar to the choice of<br> whether or not to use destructive update in other circumstances. If<br> there is a significant amount of state that is frequently updated,<br> then destructive update is probably worthwhile even though the user<br> will need to manage uniqueness as a result. Since MT requires about<br> 2.5k of state and is likely to be used when lots of calls are made,<br> the unique generator will be the one that most users would want.<br> <br> If a shared generator is required by the user and they don't mind<br> copying the state for every number generated, this can be done with<br> one of the adapters down near the bottom of the random.m interface<br> rather than re-implementing the algorithm. For example, you can import<br> mt_unique and define a type:<br> <br> :- type random == shared_random(params, ustate).<br> <br> Then create a generator as follows:<br> <br> seed(12345u32, P, R),<br> S0 = make_shared_random(P, R)<br> <br> See the attached file test_mt_1.m.<br> <br> It's also possible to attach a generator to the I/O state using<br> make_io_urandom, so that the generator can be used more conveniently<br> without copying, and without the generator itself needing to be<br> unique.<br> <br> The "ugly truth" answer: the shared implementation you've given still<br> uses arrays underneath, and these are destructively updated and aren't<br> supposed to be shared. As a consequence this typeclass instance<br> doesn't behave correctly if the generator is not actually used in a<br> unique way. For example, if the runif test was called twice in<br> sequence, passing the final state from the first call as the initial<br> state of the second, we'd expect to observe different numbers. If it<br> was run a third time, using the *first* state, it should reproduce the<br> first numbers again, but the attached file test_mt_2.m shows that it<br> doesn't.<br> <br> The underlying problem is that if arrays are put in data structures,<br> and you refer to an array that has been elsewhere updated, you can get<br> the new value when logically it should return the old value. The<br> correct way to handle this, aside from not referring to old arrays in<br> the first place, is to copy arrays when required. That's what the<br> shared adapter does via urandom_dup Ideally, the compiler would<br> check that this is done correctly, but unfortunately the current<br> implementation doesn't always do that - see the caveat at the start of<br> array.m.<br> <br> I've included in the above test some code to demonstrate the problem<br> just with arrays on their own.<br> <br> Shortest answer: I'll add some docs to the random.m module to clarify<br> for instance implementors. :-)<br> <br> Hope this helps!<br> <br> Mark<br> <br> On Sun, Feb 12, 2023 at 12:38 AM Mark Clements <mark.clements@ki.se> wrote:<br> ><br> > I have written a Mercury-only implementation of the Mersenne Twister random number generator. This is a very common and reasonably good pseudo random number generator. For details:<br> ><br> > <a href="https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fmclements%2Fmercury-mt&data=05%7C01%7Cmark.clements%40ki.se%7Cb48b433f28744e59a55d08db1346e47e%7Cbff7eef1cf4b4f32be3da1dda043c05d%7C0%7C0%7C638124967789194891%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000%7C%7C%7C&sdata=V4x8%2FyBaB5Eh11m73B2k4x9RhVEqTJVGyaAoT6Tc2DM%3D&reserved=0"> https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fmclements%2Fmercury-mt&data=05%7C01%7Cmark.clements%40ki.se%7Cb48b433f28744e59a55d08db1346e47e%7Cbff7eef1cf4b4f32be3da1dda043c05d%7C0%7C0%7C638124967789194891%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000%7C%7C%7C&sdata=V4x8%2FyBaB5Eh11m73B2k4x9RhVEqTJVGyaAoT6Tc2DM%3D&reserved=0</a><br> ><br> > Sincerely, Mark.<br> ><br> ><br> > När du skickar e-post till Karolinska Institutet (KI) innebär detta att KI kommer att behandla dina personuppgifter. Här finns information om hur KI behandlar personuppgifter<<a href="https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fki.se%2Fmedarbetare%2Fintegritetsskyddspolicy&data=05%7C01%7Cmark.clements%40ki.se%7Cb48b433f28744e59a55d08db1346e47e%7Cbff7eef1cf4b4f32be3da1dda043c05d%7C0%7C0%7C638124967789194891%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000%7C%7C%7C&sdata=B7iJrTi5U7KcEbmXLh%2ByzeNZUQo9nq67xEsspy48Qu0%3D&reserved=0">https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fki.se%2Fmedarbetare%2Fintegritetsskyddspolicy&data=05%7C01%7Cmark.clements%40ki.se%7Cb48b433f28744e59a55d08db1346e47e%7Cbff7eef1cf4b4f32be3da1dda043c05d%7C0%7C0%7C638124967789194891%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000%7C%7C%7C&sdata=B7iJrTi5U7KcEbmXLh%2ByzeNZUQo9nq67xEsspy48Qu0%3D&reserved=0</a>>.<br> ><br> ><br> > Sending email to Karolinska Institutet (KI) will result in KI processing your personal data. You can read more about KI’s processing of personal data here<<a href="https://ki.se/en/staff/data-protection-policy">https://ki.se/en/staff/data-protection-policy</a>>.<br> > _______________________________________________<br> > users mailing list<br> > users@lists.mercurylang.org<br> > <a href="https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.mercurylang.org%2Flistinfo%2Fusers&data=05%7C01%7Cmark.clements%40ki.se%7Cb48b433f28744e59a55d08db1346e47e%7Cbff7eef1cf4b4f32be3da1dda043c05d%7C0%7C0%7C638124967789194891%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000%7C%7C%7C&sdata=Ztb33y1WBBmyQA%2B3aasFSbbzuHOMATU0a10NxMYN0Xw%3D&reserved=0"> https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.mercurylang.org%2Flistinfo%2Fusers&data=05%7C01%7Cmark.clements%40ki.se%7Cb48b433f28744e59a55d08db1346e47e%7Cbff7eef1cf4b4f32be3da1dda043c05d%7C0%7C0%7C638124967789194891%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000%7C%7C%7C&sdata=Ztb33y1WBBmyQA%2B3aasFSbbzuHOMATU0a10NxMYN0Xw%3D&reserved=0</a><br> </div> </span></font></div> </body> </html>