Prescale the muon shower lines

Test prescaling the muon shower lines, as they seem to take a lot of BW on data (900 MB/s).

A test on run 305589, shows:

nominal: 1.1G Sep  9 22:04 test.turbo.mdf
prescaled to 10%: 779M Sep  9 23:01 test.turbo.mdf

so a deducting of about 30% in the Turbo stream out of HLT2.

added PR/Moore_hlt2_and_spruce_bandwidth PR/Moore_hlt2_and_spruce_bandwidth_SMOG2 labels

/ci-test

added RTA label

added ci-test-triggered label

• [2024-09-09 20:19] Validation started with lhcb-2024-patches-mr#1417

changed the description

marked this merge request as ready

@cagapopo @msaur We need a patch release of Moore with this change only, to reduce the BW to disk (and tape).

Could this be done quickly? Thanks

assigned to @cagapopo

fyi @mvesteri, also @lugrazet given that this is a qee line and @lohenry as the author. Just to be clear, given the significant impact on bandwidth, we will prescale this as soon as possible, any further optimisations that you may want to perform can come in a separate MR.

approved this merge request

approved this merge request

approved this merge request

Another thing (more of a curiosity), looking a bit at the bandwidth test of 2024-patches (before this fix), the BW of this line looks significantly lower than what reported here, around 8 MB/s , see here . Any idea why the result is so different on data @decianm ?

resolved all threads

mentioned in commit ccd1016e

merged

added qee label

mentioned in merge request !3732 (merged)