Improve safeguarding against large Sudakov reweighting factors
Before, each Sudakov trial emission was checked individually, and rew factors below -9 or above 10 were ignored. Now, we check for a large accumulated weight (absolute value larger than 1000) after processing all trial emissions for a given parton.
This fixes a situation, in which we have found a single large accumulated reweighting factor of ~1e6 for an IS bottom quark, where many rejected Sudakov trial emissions featured individual reweighting factors of ~1-5. This observation was made when reweighting the input scale of the PDF ratios in the backwards evolution, and for trial emission scales around t = 100 GeV^2.
The above threshold (default: 1e3) can be adjusted by the CSS_MAX_REWEIGHT_FACTOR setting.
Edited by Enrico Bothmann
Merge request reports
Activity
Filter activity
Please register or sign in to reply