ZPD LUT Formats

The LUT formats are defined most exactly in their VHDL code. A close second which is more readable by most people is the C++ code in L1DctConfig/L1DZpdLut.cc. A few of the more common and/or changeable LUTs are documented here.

Decision Module Cuts

Bit 0 A10 track cuts		Bit 0 A7 track cuts
Addr	Cut	Addr	Cut
0x100	min curvature	0x110	min curvature
0x101	max curvature	0x111	max curvature
0x102	min tanλ	0x112	min tanλ
0x103	max tanλ	0x113	max tanλ
0x104	min z0	0x114	min z0
0x105	max z0	0x115	max z0
0x106	max z0 error	0x116	max z0 error
0x107	max missing segments	0x117	max missing segments

The cuts for bit 1 start at 0x200, for bit 2 at 0x300, etc.

The cuts are implemented as min < value < max, i.e., less-than, not less-than-or-equal...).

Decision Module Max Possible Segments

Each memory address contains a 4-bit maximum number of segments possible for a certain curvature (rho) and tanλ (dip) combination. Ideally we would use all 8 bits of both quantities to form a 16 bit address (2¹⁶ = 65k) LUT. But we don't have enough memory (or address space) for that so we need to drop some bits of precision. Since the LUT would be symmetric in positive/negative curvature, we drop one bit by taking the absolute value. Then we drop the remaining 2 least significant bits. Since the number of possible segments depends more strongly on tanλ, we only drop one LSB there.

Here is a snippet of C++ code which converts rho and dip into the memory address for that combination:

// addr[15:12] = 0001
// addr[11: 7] = abs(rho[7:0]) [6:2]
// addr[ 6: 0] = dip[7:1]
unsigned getMaxSegsLutAddr(int rho, int dip)
{
  // rho and dip are 8 bit signed quantities range [-128,127]
  assert(-128 <= rho && rho < 128);
  assert(-128 <= dip && dip < 128);
  
  // take |rho|, use +127 for |-128| because +128 takes 9 bits, not 8...
  unsigned absrho = ( -128==rho ? 127 : abs(rho) );

  unsigned irho = (absrho>>2) & 0x1f;  // 5 bits
  unsigned idip = (dip>>1) & 0x7f;     // 7 bits

  unsigned addr = 0x1000 + (irho<<7) + idip;
  return addr;
}

Basic Parameters

Addr	Value	Units
0x11	Number of curvature bins
0x12	min pT	MeV
0x13	number of tanλ bins (A10)
0x14	min tanλ (A10)	10-3
0x15	max tanλ (A10)	10-3
0x16	number of tanλ bins (A7)
0x17	min tanλ (A7)	10-3
0x18	max tanλ (A7)	10-3
0x19	B field	milliTesla
0x1a	IPx	10-3 cmz
0x1b	IPy	10-3 cm
0x1c	IPz	10-3 cm
0x1d	ZPD octant

The Finders also store the basic parameters which were used to generate their LUT contents. They are stored in each Finder block (0x2, 0x8, 0x20, 0x80, 0x200, 0x800) at addresses 0x110 to 0x116:

Addr	Value	Units
0x110	Number of curvature bins
0x111	min pT	MeV
0x112	number of tanλ bins (A10)
0x113	min tanλ (A10)	10-3
0x114	max tanλ (A10)	10-3
0x115	phi merge
0x116	B field	milliTesla

The ZPD configuration code checks the consistency of the Finder and Fitter basic parameters to ensure that they match.

Other LUTs