[PD] freeverb crashes on new computer
Hi, im trying to run old patches on my new computer but freeverb~ always crashes , from time to time it stop doing sound and i get nan if i conect an env~ object after it. ive tried putting a clip~ -1 1 before freeverb~ but its still crashing Do anybody have an idea why is this behaivor and how to solve it? Im on macosx 10.6.7, 2.2 ghz inetl core i7 thanks R. ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
[PD] [OT] SSE/MMX tips?
I am trying to code a simple operation using SSE2 instructions where possible. I have a feeling that what I want to do is just a matter of a couple of shufps and haddps instructions but I can't get it. Lazyweb please help! The operation is integration. I have a vector of 4 single floats (v4sf) and a carry-in float to start. For example CI F0 F1 F2 F3 5 1 0 10 -5 Yields F0 F1 F2 F3 6 6 16 11 So far iteration on plain floats seems to be the best I can come up with, but HADDPS is tantalizingly close to what I want to do. Any hints? Thanks, Bill Gribble ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
[PD] family affairs
Hello list, im trying to use the patch Frank posted some years ago in the list about parent- child relation with sliders. The idea is that with the parent slider you control the child values, but rather than just setting the child to the parents value the patch add the parents value in the childer sliders. but the patch has a limitation( bug? ): it just work when you have the parent value set to 0 , if you dont have the parent slider set to 0 and you change the childs sliders value and then the parent slider it wont work you are going to get jumps in the childs sliders. Which would be the best way to avoiding these jumps and make it work not only then the parent slider is set to 0? thanks here is the patch family_affair.pd Description: Binary data ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [OT] SSE/MMX tips?
On Wed, 7 Sep 2011, Bill Gribble wrote: So far iteration on plain floats seems to be the best I can come up with, but HADDPS is tantalizingly close to what I want to do. Any hints? Once I thought that with some commutativity you could speed things up like this : (f0+f1+f2+f3)+(f4+f5+f6+f7)+... can be rearranged as : (f0+f4+...)+(f1+f5+...)+(f2+f6+...)+(f3+f7+...) But I don't remember whether I tried it or not. That's a speedup without even using MMX/SSE... theoretically, it can double the speed of a summation like this, and you can apply this boost to sum-of-products to get a certain amount of speedup too. ___ | Mathieu Bouchard tél: +1.514.383.3801 Villeray, Montréal, QC___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] freeverb crashes on new computer
On Tue, 6 Sep 2011, ronni montoya wrote: Hi, im trying to run old patches on my new computer but freeverb~ always crashes , from time to time it stop doing sound and i get nan if i conect an env~ object after it. AFAIK, if you get nan out of [env~], it's because there's at least one nan per window in the output of [freeverb~]... you can use [print~] for greater detail, so that you can figure out the difference between « there is one nan every n sample » or « every sample is nan » for a more official bug report. Do anybody have an idea why is this behaivor and how to solve it? nan is often due to trying to do 0/0 or log(-1) or other out-of-range things. (some other out-of-range operations yield +inf or -inf instead, it depends) Actually... is [clip~ -1 1] really able to get rid of nan ? can you check that your input of freeverb never has any nan ? I think that in theory, [clip~] knows how to get rid of +inf and -inf, but doesn't know what to do with nan. If not, then someone has to be reading the freeverb source and look for such possible mistakes. ___ | Mathieu Bouchard tél: +1.514.383.3801 Villeray, Montréal, QC ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
[PD] exporting text file and formatting question
Greetings all I would like to export data to a text file every time the metro is banged with the format MMDDhhmmss, float, float, float I can create the data but the questions I have 1) I keep getting a write failed. 2) how can I get the commas between the values? 3) is there a way to do leading zeros? Here's a link to my pd http://dl.dropbox.com/u/6576402/questions/pd/export_text.pd Ps: I'm using Linux ubuntu 10.04 64bit thanks #N canvas 644 98 688 570 10; #X obj 102 162 date; #X obj 167 158 time; #X obj 168 81 metro 1000; #X obj 160 27 tgl 15 0 empty empty empty 17 7 0 10 -262144 -1 -1 0 1; #X obj 183 134 bng 15 250 50 0 empty empty empty 17 7 0 10 -262144 -1 -1; #X floatatom 22 209 5 0 0 0 - - -; #X floatatom 347 214 5 0 0 0 - - -; #X obj 345 101 i; #X msg 408 57 0; #X obj 403 99 + 2; #X obj 401 6 loadbang; #X floatatom 65 208 5 0 0 0 - - -; #X floatatom 111 210 5 0 0 0 - - -; #X floatatom 169 212 5 0 0 0 - - -; #X floatatom 218 212 5 0 0 0 - - -; #X floatatom 261 212 5 0 0 0 - - -; #X obj 225 469 textfile; #X obj 167 386 bng 15 250 50 0 empty empty empty 17 7 0 10 -262144 -1 -1; #X obj 44 290 pack f f f; #X obj 207 290 pack f f f; #X obj 56 248 bondo 3; #X msg 221 405 print; #X msg 220 434 write /tmp/testfile.txt; #X obj 398 168 + 3; #X obj 450 169 - 1; #X floatatom 399 214 5 0 0 0 - - -; #X floatatom 458 214 5 0 0 0 - - -; #X obj 214 250 bondo 3; #X obj 391 247 bondo 3; #X obj 227 328 bondo 3; #X obj 383 284 pack f f f; #X obj 218 367 pack f f f; #X obj 222 505 print list; #X text 114 249 MMDDhhmmss; #X connect 0 0 5 0; #X connect 0 1 11 0; #X connect 0 2 12 0; #X connect 1 0 13 0; #X connect 1 1 14 0; #X connect 1 2 15 0; #X connect 2 0 4 0; #X connect 2 0 0 0; #X connect 2 0 7 0; #X connect 2 0 1 0; #X connect 3 0 2 0; #X connect 5 0 20 0; #X connect 6 0 28 0; #X connect 7 0 6 0; #X connect 7 0 9 0; #X connect 7 0 23 0; #X connect 7 0 24 0; #X connect 8 0 7 1; #X connect 9 0 7 1; #X connect 10 0 8 0; #X connect 11 0 20 1; #X connect 12 0 20 2; #X connect 13 0 27 0; #X connect 14 0 27 1; #X connect 15 0 27 2; #X connect 16 0 32 0; #X connect 18 0 29 0; #X connect 19 0 29 1; #X connect 20 0 18 0; #X connect 20 1 18 1; #X connect 20 2 18 2; #X connect 21 0 17 0; #X connect 21 0 22 0; #X connect 22 0 16 0; #X connect 23 0 25 0; #X connect 24 0 26 0; #X connect 25 0 28 1; #X connect 26 0 28 2; #X connect 27 0 19 0; #X connect 27 1 19 1; #X connect 27 2 19 2; #X connect 28 0 30 0; #X connect 28 1 30 1; #X connect 28 2 30 2; #X connect 29 0 31 0; #X connect 29 1 31 1; #X connect 29 2 31 2; #X connect 30 0 29 2; #X connect 31 0 21 0; ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] freeverb crashes on new computer
On 07/09/11 15:23, Mathieu Bouchard wrote: On Tue, 6 Sep 2011, ronni montoya wrote: Do anybody have an idea why is this behaivor and how to solve it? nan is often due to trying to do 0/0 or log(-1) or other out-of-range things. Probably inf - inf or 0 * inf are NaN, I haven't checked the specs. (some other out-of-range operations yield +inf or -inf instead, it depends) Actually... is [clip~ -1 1] really able to get rid of nan ? No, because NaN compares always result in false: http://pure-data.git.sourceforge.net/git/gitweb.cgi?p=pure-data/pure-data;a=blob;f=src/d_math.c;h=f76bffef18607334a67b147fc498b7513dd6372b;hb=HEAD#l42 Workaround/fix might be: if (!(f x-lo)) f = x-lo; if (!(f x-hi)) f = x-hi; But this might all be screwed up depending on how Pd and/or freeverb is compiled, I know some inf/nan things can behave even more strangely with -ffast-math (for example, Pd .git has no mention of it as far as I can tell). Claude ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [OT] SSE/MMX tips?
On 07/09/11 12:17, Bill Gribble wrote: The operation is integration. Try calling it 'scan' and you might end up with more productive searches, at least to my mind integration is more about summing over continuous regions than something discrete like this. Searching for data parallel algorithm scan leads to pages like: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.85.1876rep=rep1type=pdf http://http.developer.nvidia.com/GPUGems3/gpugems3_ch39.html Claude ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [OT] SSE/MMX tips?
Ah! Thanks. Apparently also called prefix sum. Thanks, Bill Gribble On Wed, 2011-09-07 at 16:21 +0100, Claude Heiland-Allen wrote: On 07/09/11 12:17, Bill Gribble wrote: The operation is integration. Try calling it 'scan' and you might end up with more productive searches, at least to my mind integration is more about summing over continuous regions than something discrete like this. Searching for data parallel algorithm scan leads to pages like: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.85.1876rep=rep1type=pdf http://http.developer.nvidia.com/GPUGems3/gpugems3_ch39.html Claude ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
[PD] Problem with msgfile in zexy
Hello, I have a problem with [msgfile]. When i send [add2 1 2 3 4( to [msgfile], Pd crash. This on Ubuntu 11.04, Pd 0.42.6 and Zexy 2.2.3. ++ Jack ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] exporting text file and formatting question
Hi Rick, your patch won't work. There are several wrong idioms, please take a look at the help file of [textfile]. The idea is you write each single line with the [add ( message and eventually write the whole text to a .txt file. The [print ( message in your patch is not needed. Also be careful while using [pack]. Remember that whenever it receives something on the first (hot) inlet, it outputs all the values it stored. In your patch [pack] behaviour is controlled properly. See also [pack] help file. A proper use of [pack] wouldn't require the [bondo] obj. cheers, M I can create the data but the questions I have 1) I keep getting a write failed. 2) how can I get the commas between the values? 3) is there a way to do leading zeros? Here's a link to my pd http://dl.dropbox.com/u/6576402/questions/pd/export_text.pd Ps: I'm using Linux ubuntu 10.04 64bit -- Marco Donnarumma Independent New Media and Sonic Arts Practitioner, Performer, Teacher ACE, Sound Design MSc by Research (ongoing) The University of Edinburgh, UK ~ Portfolio: http://marcodonnarumma.com Research: http://res.marcodonnarumma.com | http://www.thesaddj.com | http://www.flxer.net Director: http://www.liveperformersmeeting.net ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] (breaking symbols) was Re: find a list of numbers in a text file
On Sep 6, 2011, at 4:21 PM, martin.pe...@sympatico.ca martin.pe...@sympatico.ca wrote: From: h...@at.or.at To: martin.pe...@sympatico.ca Date: Tue, 6 Sep 2011 15:57:32 -0400 CC: pd-list@iem.at Subject: Re: [PD] (breaking symbols) was Re: find a list of numbers in a text file On Sep 6, 2011, at 1:16 PM, martin.pe...@sympatico.ca martin.pe...@sympatico.ca wrote: Date: Tue, 6 Sep 2011 08:12:15 +0200 From: f...@footils.org To: h...@at.or.at CC: pd-list@iem.at Subject: Re: [PD] (breaking symbols) was Re: find a list of numbers in a text file Hi, On Mon, Sep 05, 2011 at 02:34:30PM -0400, Hans-Christoph Steiner wrote: Then having the patch rely on the error: float: no method for 'symbol' error that is normally generated in that case. Well, it prints an error because there *is* an error in the patch. The author should never have converted a float to a symbol and expect to still do float calculations with it. Yes, I think it ought to be up to the external to implement its own symbol-to-float converter since there is no universally valid way of doing it. The same pattern will have different meanings in different context and Pd can't be expected to know what is expected in a specific case. Martin That sounds like a recipe for disaster, then we end up with multiple, incompatible ways of doing this. I don't think anyone is proposing that symbols like four get converted to 4, but only symbols like 43, 5.22, -234, etc. Things that can be directly converted from a char* to a float with a standard function. I think someone might well want to make an external that can take 'four' as 4. Or a symbol '4+2' that would become a float 6.0. And obviously Pd should not be doing that all the time because most of the time that is not the right thing to do. The disaster would be if Pd tried to guess what the user intends based on some general idea of what Pd is 'for'. For instance I've been working on an xbee external where I need to specify numbers like 0x0a060123 for the remote address. I do that by interpreting the symbol as a 64-bit integer internally but I would not expect any other external to do this unless it needs to. Otherwise the user would have to manually convert the number (which is conventionally given in hexadecimal form) into a bunch of decimal numbers small enough to be represented as float and then the external would combine them internally. Martin Yeah, I agree, I have no problem if someone wants to make an externals that interprets 4+2 or four as 4, but it doesn't seem like something that should be in the core. About hex 0x form, that could conceivably fall under the anything that looks like a number is a float, as least for programmers. In most programming languages, 0x10 is just a number like 16. .hc If you are not part of the solution, you are part of the problem. ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] pd-extended 0.43 windows 7 installation
Hi Hans, Thanks for the help with this but I now get the same error message except it now says 'libstd++-6.dll is missing...' Im downloading from here BTW http://autobuild.puredata.info/auto-build/latest Thanks again On Tue, Sep 6, 2011 at 8:20 PM, Hans-Christoph Steiner h...@at.or.atwrote: By the way, I added libpthread-2.dll and libgcc_s_dw2-1.dll to Pd-extended 0.43, please try it again and see if it works now. http://pure-data.svn.**sourceforge.net/viewvc/pure-** data?view=revisionrevision=**15285http://pure-data.svn.sourceforge.net/viewvc/pure-data?view=revisionrevision=15285 .hc On Sep 6, 2011, at 10:07 AM, Patrice Colet wrote: - Mathieu Bouchard ma...@artengine.ca a écrit : -static-libstdc++ is not in gcc in Ubuntu 10.10 (« gcc version 4.4.5 (Ubuntu/Linaro 4.4.4-14ubuntu5) ») Is it a MinGW-only feature ? I don't think so $ gcc -v Using built-in specs. COLLECT_GCC=C:\MinGW\bin\gcc.**exe COLLECT_LTO_WRAPPER=c:/mingw/**bin/../libexec/gcc/mingw32/4.** 5.2/lto-wrapper.exe Target: mingw32 Configured with: ../gcc-4.5.2/configure --enable-languages=c,c++,ada,** fortran,obj c,obj-c++ --disable-sjlj-exceptions --with-dwarf2 --enable-shared --enable-libgo mp --disable-win32-registry --enable-libstdcxx-debug --enable-version-specific-r untime-libs --disable-werror --build=mingw32 --prefix=/mingw Thread model: win32 gcc version 4.5.2 (GCC) too, if I'm not mistaken. E.g. could it solve the libstdc++ problem that we were having with Gem, GridFlow, ATI drivers, and perhaps some other drivers or libs ? what problem? __**__** ___ | Mathieu Bouchard tél: +1.514.383.3801 Villeray, Montréal, QC -- Patrice Colet __**_ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/** listinfo/pd-list http://lists.puredata.info/listinfo/pd-list --**--** http://at.or.at/hans/ __**_ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/** listinfo/pd-list http://lists.puredata.info/listinfo/pd-list ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] pd-extended 0.43 windows 7 installation
Ok, I think that's the last missing DLL. At least it works for me now on XP. Try tomorrow's build. http://pure-data.svn.sourceforge.net/viewvc/pure-data?view=revisionrevision=15290 .hc On Sep 7, 2011, at 2:57 PM, ALAN BROOKER wrote: Hi Hans, Thanks for the help with this but I now get the same error message except it now says 'libstd++-6.dll is missing...' Im downloading from here BTW http://autobuild.puredata.info/auto-build/latest Thanks again On Tue, Sep 6, 2011 at 8:20 PM, Hans-Christoph Steiner h...@at.or.at wrote: By the way, I added libpthread-2.dll and libgcc_s_dw2-1.dll to Pd- extended 0.43, please try it again and see if it works now. http://pure-data.svn.sourceforge.net/viewvc/pure-data?view=revisionrevision=15285 .hc On Sep 6, 2011, at 10:07 AM, Patrice Colet wrote: - Mathieu Bouchard ma...@artengine.ca a écrit : -static-libstdc++ is not in gcc in Ubuntu 10.10 (« gcc version 4.4.5 (Ubuntu/Linaro 4.4.4-14ubuntu5) ») Is it a MinGW-only feature ? I don't think so $ gcc -v Using built-in specs. COLLECT_GCC=C:\MinGW\bin\gcc.exe COLLECT_LTO_WRAPPER=c:/mingw/bin/../libexec/gcc/mingw32/4.5.2/lto- wrapper.exe Target: mingw32 Configured with: ../gcc-4.5.2/configure --enable-languages=c,c+ +,ada,fortran,obj c,obj-c++ --disable-sjlj-exceptions --with-dwarf2 --enable-shared -- enable-libgo mp --disable-win32-registry --enable-libstdcxx-debug --enable- version-specific-r untime-libs --disable-werror --build=mingw32 --prefix=/mingw Thread model: win32 gcc version 4.5.2 (GCC) too, if I'm not mistaken. E.g. could it solve the libstdc++ problem that we were having with Gem, GridFlow, ATI drivers, and perhaps some other drivers or libs ? what problem? ___ | Mathieu Bouchard tél: +1.514.383.3801 Villeray, Montréal, QC -- Patrice Colet ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list http://at.or.at/hans/ ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list I have the audacity to believe that peoples everywhere can have three meals a day for their bodies, education and culture for their minds, and dignity, equality and freedom for their spirits. - Martin Luther King, Jr. ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] (breaking symbols) was Re: find a list of numbers in a text file
- Original Message - From: Hans-Christoph Steiner h...@at.or.at To: Jonathan Wilkes jancs...@yahoo.com Cc: fbar f...@footils.org; pd-list@iem.at pd-list@iem.at Sent: Tuesday, September 6, 2011 4:04 PM Subject: Re: [PD] (breaking symbols) was Re: find a list of numbers in a text file On Sep 6, 2011, at 12:30 PM, Jonathan Wilkes wrote: - Original Message - From: fbar f...@footils.org To: pd-list@iem.at pd-list@iem.at Cc: Sent: Tuesday, September 6, 2011 3:53 AM Subject: Re: [PD] (breaking symbols) was Re: find a list of numbers in a text file On Tue, Sep 06, 2011 at 09:44:33AM +0200, Frank Barknecht wrote: I'm not sure what appears in the patch should mean. It definitly means that numercial-symbol selectors don't get shown and cannot be written into a patch, so you cannot use them in the editor where real selectors should be written, like in [route]. Forgot to add: Of course it is possible and legal to use numerical or non-printable symbols as selectors, but they have to be constructed dynamically and cannot be typed, in accordance with the restrictions mentioned in the manual. Instead something like this can be used: [makefilename %d] | | [makefilename %d] | | [select symbol-dummy] I used [makefilename %d] a lot in the rj library's [m_chorddict] dictionary for chords, where some chord names are proper symbols, like m7, while others are floats like 7. The float-names get converted to symbols internally to look up chord notes in a data structure array keyed by symbols only (using [m_symbolarray]). At what point are you using numerical-symbol selectors? Everything you've described has the selector 'symbol'. If you mean you let the user send symbols or floats as the key and convert internally, that's _exactly_ what I'm proposing. I guess I'm not clear on your proposal. Is it that a symbol selector automatically converts things to a symbol? That makes a lot of sense, and would help with other issues. Then you could also make symbols with spaces, like: [symbol 43( [symbol /home/hans/My Documents( Well, that's something I've wanted for a long time. But what I am proposing has to do with selectors, not symbol messages. Problem: convert from symbol-atom to float-atom Proposal: if a selector happens to be in a form that can be interpreted by the naked eye as a valid Pd float, and the object receiving the message has a float method (and no anything method), then send a float to the object. [r infinite-expressivity] | [1( - float | [makefilename %d] -- converted to symbol message (and the message arg is convert to a symbol-atom) | [list trim] -- now we have a message with the selector 1 and no arguments | [route float] -- seriously, it's a symbol-atom, not a float | + | [float] -- my proposal: give [float] a float-atom instead of a symbol-atom in this case | [route float] | [set $1, bang( | [s infinite-expressivity] But if there were a really nice quoting mechanism, that would probably be much clearer. etc. A quoting mechanism would also help. We could probably get away with only \. For example, \ for spaces, like: [symbol /home/hans/My\ Documents( [symbol I\ like\ lots\ of\ \ \ \ \ spaces( [symbol commas\,\ in\ symbols( [symbol semi-colon\;\ in\ symbols( That looks really ugly to me. What's wrong with quotes? And last but least, and its already in there: [symbol \43( [symbol \-21343( [symbol \-0.2e59( Anything that just \ couldn't cover? [openpanel] - outputs /home/hans/My documents | [set symbol $1( | [ ( -- What's printed here? ...My documents or ...My\ documents? .hc If nature has made any one thing less susceptible than all others of exclusive property, it is the action of the thinking power called an idea, which an individual may exclusively possess as long as he keeps it to himself; but the moment it is divulged, it forces itself into the possession of everyone, and the receiver cannot dispossess himself of it. - Thomas Jefferson ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [OT] SSE/MMX tips?
On Wed, 7 Sep 2011, Mathieu Bouchard wrote: On Wed, 7 Sep 2011, Bill Gribble wrote: So far iteration on plain floats seems to be the best I can come up with, but HADDPS is tantalizingly close to what I want to do. Any hints? Once I thought that with some commutativity you could speed things up like this : (f0+f1+f2+f3)+(f4+f5+f6+f7)+... can be rearranged as : (f0+f4+...)+(f1+f5+...)+(f2+f6+...)+(f3+f7+...) But what I said does not apply to your case, because you want a scan, whether I didn't really read and assumed a fold. I don't know how to optimise a scan. ___ | Mathieu Bouchard tél: +1.514.383.3801 Villeray, Montréal, QC ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [OT] SSE/MMX tips?
On Wed, Sep 7, 2011 at 7:59 PM, Mathieu Bouchard ma...@artengine.ca wrote: On Wed, 7 Sep 2011, Mathieu Bouchard wrote: On Wed, 7 Sep 2011, Bill Gribble wrote: So far iteration on plain floats seems to be the best I can come up with, but HADDPS is tantalizingly close to what I want to do. Any hints? Sorry, what's HADDPS? Once I thought that with some commutativity you could speed things up like this : (f0+f1+f2+f3)+(f4+f5+f6+f7)+... can be rearranged as : (f0+f4+...)+(f1+f5+...)+(f2+f6+...)+(f3+f7+...) But what I said does not apply to your case, because you want a scan, whether I didn't really read and assumed a fold. I don't know how to optimise a scan. This is really interesting. Your compiler probably knows how to optimize this kind of information. SSE3 is really about memory allocation. The instructions pack floats into a bigger section of memory. In SSE3, this means 4 floats in a 128-bit single operation. SSE 4.2 has 256-bit wide (8 flops per clock)--the latest increase in single-threaded computing power is in favor of single-precision float (lucky for Pd-ers) ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] [OT] SSE/MMX tips?
I noticed that your suggestion did not apply, but assumed it was a subtle riddle taunting me for an offtopic post! I think the best I can do is 2 vector adds and 2 shifts in place of 4 float adds per 4 floats. Not much of a savings, but with the loop and fetch overhead it may be worth it. I'll benchmark and see! It's really just for fun anyway. Thanks, Bill Gribble On Sep 7, 2011, at 20:59, Mathieu Bouchard ma...@artengine.ca wrote: On Wed, 7 Sep 2011, Mathieu Bouchard wrote: On Wed, 7 Sep 2011, Bill Gribble wrote: So far iteration on plain floats seems to be the best I can come up with, but HADDPS is tantalizingly close to what I want to do. Any hints? Once I thought that with some commutativity you could speed things up like this : (f0+f1+f2+f3)+(f4+f5+f6+f7)+... can be rearranged as : (f0+f4+...)+(f1+f5+...)+(f2+f6+...)+(f3+f7+...) But what I said does not apply to your case, because you want a scan, whether I didn't really read and assumed a fold. I don't know how to optimise a scan. ___ | Mathieu Bouchard tél: +1.514.383.3801 Villeray, Montréal, QC ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] OSC Confusion
Have you tried [O] | [import mrpeach] then [udprecieve ] (the is whatever port your are using), | [routeOSC] etc? On Thu, Sep 8, 2011 at 2:00 PM, Jim Aikin midigur...@sbcglobal.net wrote: I've been using Pd 0.41.4-extended in Windows 7 64-bit. Attempting to open up OSCroute-help.pd consistently crashes Pd. So I do the sensible thing and upgrade to 0.42.5. However, I now see messages telling me that sendOSC, dumpOSC, and OSCroute, which 0.41.4 seemed to have no trouble with, are now deprecated. I'm advised (by the messages in the Pd window) to Consider switching to mrpeach's [routeOSC]. I'd be happy to do so. But for some reason these objects don't seem to be included in 0.42.5-extended. So I go to a page on puredata.info that has a link to Get OSC for all platforms. Sweet -- except, not. To start with, it appears that I may need to compile the stuff I downloaded, as it includes .c files. Not being a Linux person, I don't even have a C compiler, nor do I feel inclined to work out how to use one. The README does not explain whether I need to compile the code, much less how I would do it. The readme does, however, tell me that I will also need a separate set of objects that implement the transport (OSI-Layer 4), for instance [udpsend]/[udpreceive]. And of course, the README doesn't say where I'll find those, or how to install them when I've found them. They're not on the puredata.info OSC page, as far as I can see. This leads me to a concatenation of questions. (1) Does the fact that OSCroute, sendOSC, and dumpOSC are deprecated mean that they actually don't work, or do they still work? (2) Where can I get a version of Pd-extended that includes the new, preferred OSC objects? (3) If there is no such version, how exactly would I go about incorporating them in 0.42.5? (4) What sort of patching will I need to do in order to connect the mrpeach objects with the [udp] objects after I've installed them, and where is this patching documented? Sorry to be a pest about this. Right at the moment I'm trying to document the usage of OSC to let Pd communicate with Csound. Andres Cabrera's very nice tutorial video (on YouTube) uses the old Pd objects. I'd like to use the new objects in the tutorial I'm writing, but I can hardly do so when I don't understand where the objects are or how to install them. Can anyone clear up the mysteries for me? Thanks! --Jim Aikin __**_ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/** listinfo/pd-list http://lists.puredata.info/listinfo/pd-list -- Richie ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list
Re: [PD] OSC Confusion
hello, mrPeach OSC objects, and many other ones are stuffed into pd-extended's nightly builds, you can find them at this url: http://autobuild.puredata.info/auto-build/latest/ - Jim Aikin midigur...@sbcglobal.net a écrit : I've been using Pd 0.41.4-extended in Windows 7 64-bit. Attempting to open up OSCroute-help.pd consistently crashes Pd. So I do the sensible thing and upgrade to 0.42.5. However, I now see messages telling me that sendOSC, dumpOSC, and OSCroute, which 0.41.4 seemed to have no trouble with, are now deprecated. I'm advised (by the messages in the Pd window) to Consider switching to mrpeach's [routeOSC]. I'd be happy to do so. But for some reason these objects don't seem to be included in 0.42.5-extended. So I go to a page on puredata.info that has a link to Get OSC for all platforms. Sweet -- except, not. To start with, it appears that I may need to compile the stuff I downloaded, as it includes .c files. Not being a Linux person, I don't even have a C compiler, nor do I feel inclined to work out how to use one. The README does not explain whether I need to compile the code, much less how I would do it. The readme does, however, tell me that I will also need a separate set of objects that implement the transport (OSI-Layer 4), for instance [udpsend]/[udpreceive]. And of course, the README doesn't say where I'll find those, or how to install them when I've found them. They're not on the puredata.info OSC page, as far as I can see. This leads me to a concatenation of questions. (1) Does the fact that OSCroute, sendOSC, and dumpOSC are deprecated mean that they actually don't work, or do they still work? (2) Where can I get a version of Pd-extended that includes the new, preferred OSC objects? (3) If there is no such version, how exactly would I go about incorporating them in 0.42.5? (4) What sort of patching will I need to do in order to connect the mrpeach objects with the [udp] objects after I've installed them, and where is this patching documented? Sorry to be a pest about this. Right at the moment I'm trying to document the usage of OSC to let Pd communicate with Csound. Andres Cabrera's very nice tutorial video (on YouTube) uses the old Pd objects. I'd like to use the new objects in the tutorial I'm writing, but I can hardly do so when I don't understand where the objects are or how to install them. Can anyone clear up the mysteries for me? Thanks! --Jim Aikin ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list -- Patrice Colet ___ Pd-list@iem.at mailing list UNSUBSCRIBE and account-management - http://lists.puredata.info/listinfo/pd-list