Statistics
| Branch: | Tag: | Revision:

root @ 05c117f3

Name Size
  cmake
  docs
  dtools
  gnuradio-core
  gr-atsc
  gr-audio
  gr-comedi
  gr-digital
  gr-fcd
  gr-fft
  gr-filter
  gr-howto-write-a-block
  gr-noaa
  gr-pager
  gr-qtgui
  gr-shd
  gr-trellis
  gr-uhd
  gr-utils
  gr-video-sdl
  gr-vocoder
  gr-wavelet
  gr-wxgui
  grc
  gruel
  volk
.gitattributes 323 Bytes
.gitignore 15 Bytes
AUTHORS 1.1 kB
CMakeLists.txt 9 kB
COPYING 34.3 kB
README 4.7 kB
README-win32-mingw-short.txt 3.9 kB
README.building-boost 1.7 kB
README.hacking 10 kB

Latest revisions

# Date Author Comment
05c117f3 06/15/2012 07:59 pm Tom Rondeau

filter: updating adaptive FIR filters.

No need for our own adaptive_fir class; can do everything with fir_filter.

With QA code.

1e7aae36 06/15/2012 03:53 pm Tom Rondeau

filter: minor coding things.

408f44a7 06/15/2012 03:52 pm Tom Rondeau

filter: mods to fir filter bases to make consistent and simpler.

44b525e7 06/15/2012 01:59 pm Tom Rondeau

volk: adding new kernels to test and profile.

59ad06e0 06/15/2012 01:57 pm Tom Rondeau

filter: improving documentation and adding GRC blocks for fir_filter_xxx blocks.

5585c712 06/15/2012 12:43 pm Tom Rondeau

filter: adding ssc and fsf versions of filter with associated new Volk kernels.

These routines work and pass QA. They could use some performance work. the FSF is just slightly slower than before; the SCC version is more noticably slower.

Both could benefit, probably, by using SSE2 intrinsics to handle the shorts.

a74286a2 06/14/2012 08:51 pm Tom Rondeau

filter: fixed FIR filter QA code to make sure it excersizes SIMD call.

1a5a1d4e 06/14/2012 08:13 pm Tom Rondeau

filter: added a ccf Volk dot product to use with ccf filters and used it in fir_filter_ccf.

Produces improved results to previous version.

4e2f3852 06/14/2012 08:13 pm Tom Rondeau

filter: switched ccc FIR filter to use aligned VOLK calls.

dae3b390 06/14/2012 04:26 pm Tom Rondeau

filter: fff FIR filters use VOLK machine information to set alignment.

Works with SSE and AVX machines. SSE results comparable to float_dotprod still. AVX slightly faster.

View all revisions | View revisions

Also available in: Atom