I'm not really sure where to put this code for expanding WIB2 frame ADCs to 16-bit (see description below). It will potentially be useful in the future for DQM and data selection's HLF. But I'd like it to go somewhere, so it's findable by me and other people in the future when it may be needed. There is similar code for WIB1 in readout, so maybe readout is the right place?
This commit introduces test_fast_expand_wib2frame_app.cxx which holds
code that takes a WIB2 frame with packed 14-bit ADCs and expands the
ADCs into an array of 16-bit values for later processing. AVX2 vector
instructions are used for speed. On one core of epdtdifogkv01.cern.ch
(Xeon(R) Platinum 8280L CPU @ 2.70GHz), I measure 15 MHz of frames
unpacked, which is about 0.8 of the rate of frames produced by a
single APA.
I'm not really sure where to put this code for expanding WIB2 frame ADCs to 16-bit (see description below). It will potentially be useful in the future for DQM and data selection's HLF. But I'd like it to go somewhere, so it's findable by me and other people in the future when it may be needed. There is similar code for WIB1 in
readout, so maybereadoutis the right place?This commit introduces test_fast_expand_wib2frame_app.cxx which holds code that takes a WIB2 frame with packed 14-bit ADCs and expands the ADCs into an array of 16-bit values for later processing. AVX2 vector instructions are used for speed. On one core of epdtdifogkv01.cern.ch (Xeon(R) Platinum 8280L CPU @ 2.70GHz), I measure 15 MHz of frames unpacked, which is about 0.8 of the rate of frames produced by a single APA.