You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I was looking to use this to get around a simple issue that I have with numpy histograms: specifically, numpy histograms work by calculating a cumulative sum across the bin ranges, which means that when weights vary by a factor of 10e6 or so from one end of the spectrum to the other (i.e. in simulated LHC dijet spectra) the low-weight end is lost to roundoff error.
I took a look and it wasn't clear if numpy histograms are being used behind the scenes. If not, great, this is probably exactly what I need. If so, also great, because as part of this project we should probably fix numpy histograms.
The text was updated successfully, but these errors were encountered:
However, I'm in the process of deprecating Histogrammar in favor of histbook, which also doesn't use Numpy histograms. histbook uses Numpy arrays to store and manipulate aggregated data, but the actual bin updates are performed with numpy.add.at, nothing cumulative.
I was looking to use this to get around a simple issue that I have with numpy histograms: specifically, numpy histograms work by calculating a cumulative sum across the bin ranges, which means that when weights vary by a factor of 10e6 or so from one end of the spectrum to the other (i.e. in simulated LHC dijet spectra) the low-weight end is lost to roundoff error.
I took a look and it wasn't clear if numpy histograms are being used behind the scenes. If not, great, this is probably exactly what I need. If so, also great, because as part of this project we should probably fix numpy histograms.
The text was updated successfully, but these errors were encountered: