search for: beamforming

Hey folks, Some progress has been made since the first thread: http://lists.llvm.org/pipermail/llvm-dev/2017-March/111129.html And also I think the consensus is to enable "-ffp-contract=on" by default (instead of "fast"), which seems to be working on some preliminary tests I made. I just ran the test-suite on x86_64 and AArch64. The former is ok, the latter still has some

Hi, Below you can see the updated benchmark results for the new SLP-vectorizer. As you can see, there is a small number of compile time regressions, a single major runtime *regression, and many performance gains. There is a tiny increase in code size: 30k for the whole test-suite. Based on the numbers below I would like to enable the SLP-vectorizer by default for -O3. Please let me know if you

Hi, LLVM’s SLP-vectorizer is a new pass that combines similar independent instructions in a straight-line code. It is currently not enabled by default, and people who want to experiment with it can use the clang command line flag “-fslp-vectorize”. I ran LLVM’s test suite with and without the SLP vectorizer on a Sandybridge mac (using SSE4, w/o AVX). Based on my performance measurements

Dear all, As part of working on making test-suite less demanding of exact FP results so my FP-contraction patch can go back into trunk and stay there, today I analyzed "MultiSource/Benchmarks/VersaBench/beamformer". I found that the raw output from that program is 2789780 bytes [i.e. ~2.7 _megabytes_] of floating-point text, which IMO is too much to put into a patch -- or at least

...e the values of the minimum and maximum cutoff frequencies for each coding version of the 44.1 kHz sampled data? For instance, are the values fmin=100 Hz and fmax=12 kHz valid? Thank you very much in advance. Kind regards, ? Fernando A. Marengo Rodriguez, PhD Post-doctoral fellow on Acoustics and Beamforming -- Laboratory of Noise and Vibration (LVA) Federal University of Santa Catarina (Brazil) -- Acoustics and Electroacoustics Laboratory School of Electronic Engineering National University of Rosario (Argentina) -------------- next part -------------- An HTML attachment was scrubbed... URL: http://li...

On Jul 13, 2013, at 11:30 PM, Nadav Rotem <nrotem at apple.com> wrote: > Hi, > > LLVM’s SLP-vectorizer is a new pass that combines similar independent instructions in a straight-line code. It is currently not enabled by default, and people who want to experiment with it can use the clang command line flag “-fslp-vectorize”. I ran LLVM’s test suite with and without the SLP

Sorry for not posting sooner. I forgot to send an update with the results. I also have some benchmark data. It confirms much of what you posted -- binary size increase is essentially 0, performance increases across the board. It looks really good to me. However, there was one crash that I'd like to check if it still fires. Will update later today (feel free to ping me if you don't hear

On Jul 14, 2013, at 9:52 PM, Chris Lattner <clattner at apple.com> wrote: > > On Jul 13, 2013, at 11:30 PM, Nadav Rotem <nrotem at apple.com> wrote: > >> Hi, >> >> LLVM’s SLP-vectorizer is a new pass that combines similar independent instructions in a straight-line code. It is currently not enabled by default, and people who want to experiment with it

Renato, I see that you were not on CC to the following email thread. I would appreciate if you could provide some guidance on how to fix the test-suite. On Mon, Oct 24, 2016 at 3:08 PM, Sebastian Pop <sebpop at gmail.com> wrote: > On Thu, Oct 20, 2016 at 5:59 PM, Renato Golin <renato.golin at linaro.org> wrote: >> So, let's separate the FP_CONTRACT from the FAST_MATH

Hi, Sorry for the delay in response. I measured the code size change and noticed small changes in both directions for individual programs. I found a 30k binary size growth for the entire testsuite + SPEC. I attached an updated performance report that includes both compile time and performance measurements. Thanks, Nadav On Jul 14, 2013, at 10:55 PM, Nadav Rotem <nrotem at apple.com>

...t; >If 32 bits is not enough, the next logical amount is 64 bits. Fact is that FLAC is highly economical in items such as reserving 20 bits for sampling rate or 3 bits in the middle of the middle of a byte for number of channels (which are, in fact, currently too few for applications such as beamforming that use arrays of several dozen microphones), so the next logical amount could easily be 40 bits, which would suffice for 50 h recording at 192 kHz / 32 bits / 8 channels; or even 48 bit, which allows for 1.5 year recordings (not unlikely, however, in long-term soundscape continuous recordings...

Cool! What changes have you seen to generated code size? I'll take it for a spin on our benchmarks. On Sat, Jul 13, 2013 at 11:30 PM, Nadav Rotem <nrotem at apple.com> wrote: > Hi, > > LLVM’s SLP-vectorizer is a new pass that combines similar independent > instructions in a straight-line code. It is currently not enabled by > default, and people who want to

...quencies for each coding version of the 44.1 kHz sampled data? For >> instance, are the values fmin=100 Hz and fmax=12 kHz valid? >> Thank you very much in advance. >> Kind regards, >> >> Fernando A. Marengo Rodriguez, PhD >> Post-doctoral fellow on Acoustics and Beamforming >> -- >> Laboratory of Noise and Vibration (LVA) >> Federal University of Santa Catarina (Brazil) >> -- >> Acoustics and Electroacoustics Laboratory >> School of Electronic Engineering >> National University of Rosario (Argentina) >> >> ________...

...and increases load. The Alsa-plugins of Speex echo just don’t seem to work whilst repos using the speex libs do. https://github.com/voice-engine/ec I actually use the speex AGC Alsa-plugin as its a great tool and then run Echo via the above. Coupled to WebRTC_VAD and a python based MFCC routine. Beamforming and DOA I have given up hope for as I am running so many extremely similar FFT routines concurrently in different process space that my load usage makes me believe there isn’t a chance. All of them have extremely similar frame based spectra analysis collection that collate and process over differe...

...maximum cutoff > frequencies for each coding version of the 44.1 kHz sampled data? For > instance, are the values fmin=100 Hz and fmax=12 kHz valid? > Thank you very much in advance. > Kind regards, > > Fernando A. Marengo Rodriguez, PhD > Post-doctoral fellow on Acoustics and Beamforming > -- > Laboratory of Noise and Vibration (LVA) > Federal University of Santa Catarina (Brazil) > -- > Acoustics and Electroacoustics Laboratory > School of Electronic Engineering > National University of Rosario (Argentina) > > ____________________________________________...

...es for each coding version of the 44.1 kHz sampled data? For instance, are the values fmin=100 Hz and fmax=12 kHz valid? >> >>Thank you very much in advance. >>Kind regards, >> >>? >>Fernando A. Marengo Rodriguez, PhD >>Post-doctoral fellow on Acoustics and Beamforming >>-- >>Laboratory of Noise and Vibration (LVA) >>Federal University of Santa Catarina (Brazil) >>-- >>Acoustics and Electroacoustics Laboratory >>School of Electronic Engineering >>National University of Rosario (Argentina) >> >>________________...

...uff that people in this list might found interesting. Frequency and Music An overview of frequency, harmonic (Fourier) series, and their relationship to music. http://cnx.org/content/col10338/latest/ Audio Localization This course has been created as an introduction to audio localization, and how beamforming can be applied in a real-time environment. http://cnx.org/content/col10250/latest/ (is this useful for Ghost?) Fundamentals of Digital Signal Processing Lab The purpose of this lab is to familiarize students with the DSP development workstation in the signal processing lab by examining sampling, a...

Dear Ulrich, Thanks for your answer. >Well, today 4 GiB is about half an hour of 8-channel, 96 kHz, 24-bit >uncompressed audio, or about 0.9 % of the capacity of a modest 2 TB >HDD. Not much, in other words, and who hasn't cursed yet at artificial >4 GiB (or even 2 GiB) limitations? So I wouldn't be too sure about the >"ever", even though it does seem very far

Hello, I was interested in how much Type-Based Alias Analysis helps to optimize code. For that purpose, I've compared three sets of benchmarks: compiled without TBAA, compiled with a default TBAA metadata format, and compiled with new TBAA metadata format. As a set of benchmarks, I've used the LLVM test suite (http://llvm.org/docs/TestingGuide.html#test-suite-overview) which has a lot of

Hi all, I've started looking at the GlobalMerge pass, enabled by default on ARM and AArch64. I think we should reconsider that, at least for AArch64. As is, the pass just merges all globals together, in groups of 4KB (AArch64, 128B on ARM). At the time it was enabled, the general thinking was "it's almost free, it doesn't affect performance much, we might as well use it".