AlgoCore_tau_bdt.vhd File Reference

Core of the BDT tau Algorithm. More...

Go to the source code of this file.

Entities
AlgoCore_tau_bdt	entity
	Core of the electromagnetic algorithm. More...
Behavioral	architecture
	Core of the electromagnetic algorithm. More...

Detailed Description

Core of the BDT tau Algorithm.

The total latency of this block is 10 clock cycles. Inputs: IN_ParBDT: array of 3 8-bit BDT thresholds used to compute the BDT condition, allowing loose/medium/tight working points. These values are compared with the 8 MSB bits of the BDT score to produce a number between 0 and 3 which is written to OUT_TOB.Core.BDT IN_ParFrac: array of 3 8-bit parameters used to multiply the EM estimate, compare the resulting 3 values to the HAD estimate and produce a loose/medium/tight condition for triggering on LLPs which is written out to OUT_TOB.Core.Frac IN_Min_E_threshold: A 16-bit ET threshold used to compute OUT_TOB.Core.Energy. If the energy is below this threshold, The latter field will be all zeros. IN_Min_BDT_E_threshold: A 16-bit ET threshold for low ET cutoff of the BDT condition. Below it the BDT WP is set to "tight". IN_Max_cond_E_threshold: A 16-bit ET threshold for high ET cutoff of both BDT and hadronic fraction conditions. Beyond it the corresponding WP is set to "tight". IN_Data: an array of TriggerTowers with nine elements, representing a single TOB.

The input data is mapped like this: +----—+----—+----—+ | 6 | 7 | 8 | | |0 1 2 3| | | | | | +----—+----—+----—+ | 3 | 4 | 5 | | | |0 1 2 3| | | | | +----—+----—+----—+

	0 1 2 3
0	1	2

+----—+----—+----—+

The BDT's inputs are sums over groups of symmetrically positioned supercells around the TOB axis. For example, a sum over all the corner supercells of layer 2 is one variable. The BDT has 11 such variables. The design is constructed of several stages:

1. Produce all necessary sums inside the AdderTree entity:
   • 11 BDT variables
   • EM estimate
   • HAD estimate (used along with EM estimate to compute LLP discriminant, aka Frac condition)
   • 9 Towers (used to determine if the TOB is seeded)
   • ET estimate The AdderTree.vhd file is automatically generated and contains in it's docstring a table describing the latencies of when the outputs are ready.
2. Compute the BDT score (BDTModel), determine if the TOB is seeded (TauSeedFinder) and perform the multiplication (MultiMultiplier)
3. Evaluate the conditions (TauConditionsEnergyAndSeed, TauConditionsBDT, TauConditionsFrac) All signal delays are handled inside the DelayTree entity, which is automatically generated. The reason for having AdderTree and DelayTree automatically generated is to be able to quickly change the BDT model to another one and also be able to modify the way energy estimates are computed (that is, what supercells to sum) without having to manually adjust delays.

   Author

   David Reikher

Definition in file AlgoCore_tau_bdt.vhd.