http://forums.leaflabs.com/topic.php?id=35 A place to share, learn, and grow... en-US Fri, 22 Jan 2016 00:18:28 +0000 http://bbpress.org/?v=1.0.2 q http://forums.leaflabs.com/search.php http://forums.leaflabs.com/topic.php?id=35#post-2754 Fri, 19 Nov 2010 16:21:24 +0000 gbulmer 2754@http://forums.leaflabs.com/ <p>Have you looked at the generated code to see that gcc doesn't use the MAC instructions?</p> <p>ARM say Cortex-M3 isn't aimed at being DSP alternatives. </p> <p>Cortex-M4 is a bit more performance orientated with hardware floating point and some single-instruction stream, multiple-data stream (SIMD) instructions.<br /> Cortex-M4 runs at 150MHz-ish. The NXP version has a Cortex-M0 co-processor on board!</p> <p>Feeescale and NXP are first to announce, but ST have licensed Cortex-M4 too, and will likely announce soon.</p> <p>If you want a lot of arithmetic performance, but want to use Cortex-M3, consider Actel SmartFusion, which is a Cortex-M3 + FPGA in a single chip. On some problems, it should give Intel a run for their money (and may win on some specialised problems). </p> http://forums.leaflabs.com/topic.php?id=35#post-2751 Fri, 19 Nov 2010 14:36:09 +0000 larryang 2751@http://forums.leaflabs.com/ <p>Out of curiousity, is there a pragma or key word or something so that the code you write takes advantage of the multiply and accumulate instruction?</p> <p>It doesn't looks like the chip have other DSP like features though, like circular buffers or loop buffers. </p> http://forums.leaflabs.com/topic.php?id=35#post-2670 Wed, 17 Nov 2010 02:22:59 +0000 metaphysician 2670@http://forums.leaflabs.com/ <p>hi there - just a quick bump on this from a new guy, as i want to stay on top of whatever's being done in terms of audio DSP development on the Maple - i really enjoyed Okie's processing demo and was wondering where that project (or any project involving audio DSP on the Maple) was at these days. </p> http://forums.leaflabs.com/topic.php?id=35#post-2013 Mon, 25 Oct 2010 20:17:39 +0000 gbulmer 2013@http://forums.leaflabs.com/ <p>as have I :-) </p> http://forums.leaflabs.com/topic.php?id=35#post-2007 Mon, 25 Oct 2010 19:59:28 +0000 poslathian 2007@http://forums.leaflabs.com/ <p>just answered that ;) </p> http://forums.leaflabs.com/topic.php?id=35#post-2004 Mon, 25 Oct 2010 19:51:16 +0000 gbulmer 2004@http://forums.leaflabs.com/ <p>poslathian - fixed point please!</p> <p>I posted a strawman list of requirements in the thread <a href="http://forums.leaflabs.com/topic.php?id=259&amp;page=2">http://forums.leaflabs.com/topic.php?id=259&amp;page=2</a> </p> http://forums.leaflabs.com/topic.php?id=35#post-2000 Mon, 25 Oct 2010 19:45:33 +0000 poslathian 2000@http://forums.leaflabs.com/ <p>We did achieve some very fast FFT results with a C based implementation, but it is not baked in to wirish/libmaple. Perhaps this would make a good library if anyone is interested in writing one. We have also been toying with the idea of adding a fixed point library, which I find essential for most audio and synthesizer work. </p> http://forums.leaflabs.com/topic.php?id=35#post-1951 Sat, 23 Oct 2010 19:46:30 +0000 anotherDJ 1951@http://forums.leaflabs.com/ <p>Hi,</p> <p>I'm new to all this, I'll post sometime soon to figure out if Maple is exactly right for my needs. But just wondering can these DSP algorithms shifty is talking about be used on maple? as in integrated with wirish. </p> <p>thanks </p> http://forums.leaflabs.com/topic.php?id=35#post-130 Fri, 28 May 2010 22:25:45 +0000 bnewbold 130@http://forums.leaflabs.com/ <p>Bootloader code is available at <a href="http://github.com/leaflabs/maple-bootloader" rel="nofollow">http://github.com/leaflabs/maple-bootloader</a>; the relevant file is main.c. It could probably use a refactoring/cleanup/size-optimization which would make maintaining a tracking branch difficult, but if it's a simple change that shouldn't be a problem. We already have button-detection code that enables either the hardware serial bootloader or the "endless bootloader" modes. Because of the wiring you won't be able to use the extra button on the board (connected straight to BOOT0), but you could definitely add an external button to any of the GPIO pins and check for that. </p> <p>See, eg, modifications to run the most recent bootloader on the rev1 Maple hardware (forums.leaflabs.com/topic.php?id=32#post-126 ); you should be able to maintain IDE compatibility. </p> http://forums.leaflabs.com/topic.php?id=35#post-129 Fri, 28 May 2010 22:07:05 +0000 shifty 129@http://forums.leaflabs.com/ <p>Hiya! </p> <p> The document is very sparse on the optimization details, but the good news is that it lists tables of times for all of the operations. e.g. 1024 point complex FFT takes 2.2ms @ 72 MHz. That's approx. 512K samples per second. That means, if Fs = 48KH, it's running about 10x sample rate, which is good, because a typical FFT/IFFT loop requires 4 such operations to complete, (e.g. with 50% overlap). Therefore, it's running at about 2.5X necessary speed for a mono FFT/IFFT Loop. So, stereo is likely even possible. Additionally, there are ways to further use this. for example, if the code could be modified to only work on real input and output, it could perhaps be 50% faster, but so far that's not necessary, energy can be spent on more fun things :)</p> <p>p.s. the PID loop executes in 750ns! And the doc says "Analysis of the PID timing shows that assembly code is not as fast as C code. The compiler is more efficient in accessing variables than manual optimization (offset computation and data placement in literal pool)."</p> <p>As far as the bootloader, are you releasing the source to it? If so, I think the easiest way is to make my own subtle variation on it, which requires a button to be held down during boot up to enter image upgrade mode. Will that be straightforward? </p> http://forums.leaflabs.com/topic.php?id=35#post-128 Fri, 28 May 2010 18:33:07 +0000 bnewbold 128@http://forums.leaflabs.com/ <p>Awesome, I have that downloaded but haven't read through it all yet. I was using the arduino PID library for something else but it would be better to have it implemented more efficiently.</p> <p>I know you're also interested in very fast startup bootloaders; I think the best way to go about this would be to do away with the bootloader entirely and have a make target in our command line tools which would use modified linker scripts to generate standalone .bin programs which must be uploaded over JTAG or the serial bootloader. This wouldn't work with the unmodified IDE though... what do you think? </p> http://forums.leaflabs.com/topic.php?id=35#post-127 Fri, 28 May 2010 18:05:51 +0000 shifty 127@http://forums.leaflabs.com/ <p>Hey all.</p> <p> I'm pretty excited about this new project, Maple. Here is a pdf from STM on DSP specifically for this chip. It's got sections on FIRs, IIRs, and FFTs, and PID controllers.</p> <p>DSP = Digital Signal Processing<br /> FIRs are simple, reliable filters for equalizers, synthesizers, etc.<br /> IIRs are harder to work with (less stable), but more efficient<br /> FFTs are for constructing/deconstructing the signal into its raw frequency components<br /> PIDs are algorithms to keep things balanced like a Segway scooter, or otherwise control the speed of something very carefully, such as a milling machine.</p> <p><a href="http://www.st.com/stonline/books/pdf/docs/14988.pdf" rel="nofollow">http://www.st.com/stonline/books/pdf/docs/14988.pdf</a> </p>