forked from kostrzewa/tmLQCD
-
Notifications
You must be signed in to change notification settings - Fork 0
/
README.phmc
42 lines (35 loc) · 1.31 KB
/
README.phmc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
See doc/2+1+1_howto.text for more details
as a test the following parameter set can be used (see also
sample-input/sample-hmc2.input):
Kappa = 0.170
2KappaMu = 0.01
2Kappamubar = 0.1105
2Kappaepsbar = 0.0935
beta=3.3
tlsym gauge action
antiperiodic boundary conditions for quark fields
P = 0.53347(17)
P_rec = 0.30393(22)
How it roughly works:
- Decide for a fixed order n of the less accurate polynomial
P
- Decide also for an approximation interval
- compute the n roots of the polynomial using the programme
in util/laguer/
This will store the monomials in Square_root_BR_roots.dat
and the normalisation factor in normierungLocal.dat
- the normalisation factor in the latter file corresponds
to n-th root of C_total. In the monomial repr. the square
root of this is needed. This rooting is done in phmc_tm.c
- the polynomial approximates 1/sqrt(x)
- copy the files normierungLocal.dat and Square_root_BR_roots.dat
to your running directory
- the order used in the PHMC for P must be fixed to the same
order as used for the root computation (or vice versa)
todo:
- fixed order of P as input parameter
- maybe more meaningful names for input/output files?
- generic way to use n_f=2 PHMC degenerate and non-degenerate
- eigenvalue computation - in parallel?
- on BG/L?
- reduce memory usage?