Re: [Mono-dev] Mono's JIT and SSE instructions
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Paolo Molaro wrote: We don't auto vectorize from IL code, but we're developing an extension that allows people to take advantage of sse instructions with intrinsics (people watching the changes list can see it already). That said, what versions of SSE will be supported? Original SSE only, since it's probably available on every mono-compatible machine, everything up to SSSE3/SSE4..? -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkjv7O4ACgkQU1h3wAxL79ithwCfWawHcmn3qd4XB/WfcDZVWXLL qtoAn0ayqtxqMofMsHB9QXFzsk02xD8Q =DtZP -END PGP SIGNATURE- ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] svn trunk broken?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Jacob Gladish wrote: I just sync'd up to the trunk and my build is failing. Any ideas? You might want to run a full clean (svn clean, make clean, reconfigure) and retry - that fixed that for me a few days ago. If not, deleting your local copy (saving all diffs, if applicable) and checking out again should do. bedah -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkjwCPMACgkQU1h3wAxL79gucQCgxlEimKF0V/WuyKfupBb4BILd 0QEAoKTCAZUc+AG9rE/U0uUfYWNEWyvb =yzOi -END PGP SIGNATURE- ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Add a libgmp-wrapper to mono
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Aefvadh, I think it would give it a lot of visibility there. True - and that might even attract others to work on the library. It certainly sounds convincing, but on the other hand I don't think it's a good idea to aid the ongoing process of information centralisation around google... That's a completely different topic, though. We could host it on the Mono SVN server, but it does not have all the services that Google offers for your project. I must admit that I am not particularly familiar with the services of cgc, and that said, I do believe I would not even use or even need many of them. But I'll give it a look :) And additionally, we do not think that we would include it in the standard Mono distribution, not until the API was deemed frozen, so that would take some time. I already thought so, yes :) But no problem with it, in fact I support that. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkjU0b0ACgkQU1h3wAxL79i2OgCgg8UG+sRfks+0dmClLnb39m2/ fjEAoKXyHnanc6MgPbccDszV7mq5lJya =WrkY -END PGP SIGNATURE- ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Add a libgmp-wrapper to mono
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Launchpad und Bitbucket seem to be nice, thank you both very much :) Olexandr Melnyk wrote: Or Launchpad: https://launchpad.net/ On 9/20/08, Seo Sanghyeon [EMAIL PROTECTED] wrote: 2008/9/20 Dax [EMAIL PROTECTED]: True - and that might even attract others to work on the library. It certainly sounds convincing, but on the other hand I don't think it's a good idea to aid the ongoing process of information centralisation around google... That's a completely different topic, though. If you don't like Google (I don't) you may want to try other project hosting sites like Bitbucket or Kenai. http://www.bitbucket.org/ http://kenai.com/ -- Seo Sanghyeon ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkjVposACgkQU1h3wAxL79gA6ACfQ2Ut66JHY2hMRkIy/fzx9bww rhMAniu2SDL+y8PT6kGdDhAfKZbN0vv2 =d8M6 -END PGP SIGNATURE- ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Add a libgmp-wrapper to mono
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Nice idea, but I would rather not rely on Google (yes, Google is big, and Google is always available, and many other things..). However, if it's the best solution for everyone involved, I will of course upload to cgc. Miguel de Icaza wrote: This sounds really good; My suggestion would be to host this on code.google.com and then building packages through the OpenSUSE Build Service which can create packages for multiple distributions at once. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkjTnQQACgkQU1h3wAxL79jJBgCgnBLF8RZsp91X/fIm7e0D7PQz H2AAnAg/7a4ZmgtI7FYCy3zKFCREsvAw =ucsL -END PGP SIGNATURE- ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
[Mono-dev] Add a libgmp-wrapper to mono
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Aefvadh, I have here a fresh wrapper for libgmp, which I would be happy to add to mono. Currently, it provides an interface exposing almost all directly usable usable functions to .NET (excluding file-transport functions and some minor others, the latter mostly obsolete or depracated), featuring an intuitive interface (overloaded operators und instance methods that act on the current object itself, making efficient computations of long formulae with multiplrecision variables possible, e.g. x.Add(1).Multiply(3).Remainder(5) would act on only one object, whereas (x + 1) * 3 % 5 would create two intermediate objects, not including the result, wich is another, distinct object in its own right). As of now, it is not CLS-compatible, but I am working on it. Also, the GMP APIs are preferredly exported using native long int, realized through platform switches (PLATFORM_64 define). The 32bit version will work on 64bit machines, the reverse is not true, however. It is planned to reduce the exposure of user code to this internal to a minimum, maybe by making all such input (u)int and all output (u)long, but that as well is not yet clear. Some extensions to the GMP API are also planned, such as .Raise for rationals (for which GMP only exposes + - / *), exp/log for all data types, trigonometric functions, more complex structures, ... - but priority lies on CLS-compliance for now. Please let me know what you think about this :) bedah -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkjSscsACgkQU1h3wAxL79gvxACfTIuq9E5Tmxq3WV8/8J9zDQBO BcQAnRDT4rwq80+GALGQcSsgzdi2jkjI =GzXC -END PGP SIGNATURE- ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
Re: [Mono-dev] Performance of calls
Aefvadh, The latter is by design, as methods of MBR classes are never inlined. Something like that I've assumed it to be.. But the questions on differences between csc and gmcs and why static/virtual calls to MBR classes are faster are still there.. bedah ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list
[Mono-dev] AppDomain.RelativeSearchPath is faulty?
Aefvadh, very recently I discovered something kind of odd. Please save this as Default.aspx and run it - click the button. script runat=server protected void bar(object sender, EventArgs e) { btn.Text = AppDomain.CurrentDomain.RelativeSearchPath; } /script form runat=server asp:Button ID=btn runat=server OnClick=bar / /form You will see, that the button's Text is ~/Bin;~/bin, ~ replaced by the directory your Default.aspx is saved in. I have already tried to trace it, but I got lost somewhere in the runtime... jolan'tru ___ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list