|
HARM
harm and utilities
|
|
HARM
harm and utilities
|
U->P inversion definitions for utoprim_jon.c. More...
Go to the source code of this file.
Macros | |
| #define | JONGENNEWTSTUFF 1 |
| whether to turn on jon's modificatoins to Scott's general Newton's method More... | |
| #define | CRAZYNEWCHECK 0 |
| #define | WHICHHOTINVERTER 3 |
| 1: 1-D E'(W') (really has 3 roots, but seemingly always positive-most root found AND correct root) 2: 2-D E'(W',v^2) and P^2(W',v^2) (unsure how many roots) 3: 1-D E'(W') Jon's method without intermediate v^2 and using {u}^2 instead of v^2 – avoids catastrophic cancellation (still has 3 roots) More... | |
| #define | PRIMFROMVSQ 0 |
| whether to do final calculation of primitives from v^2 or {u}^2 0: from {u}^2 – avoids catastrophic cancellation at high Lorentz factors 1: from v^2 – requires computing from 1/sqrt(1-v^2) which has problems when v^2 1 More... | |
| #define | WHICHCOLDINVERTER 2 |
| 0: original 1-D but with E' (W and v^2, but can give nan's when Wp<0 slightly – maybe fixed with new normalization?) (apparently also not giving unique solution) 1: Jon's simple 1-D E' (no use of v^2, but obtains an inconsistent v^i[P^2,E']) (apparently also not giving unique solution) Uses same as WHICHHOTINVERTER==3 2: simple 1-D P^2 (no use of v^2, obtains a consistent v^i[P^2]) (but 4 roots and never sure which one is right – not always positive root. More... | |
| #define | WHICHENTROPYINVERTER 0 |
| 0: 1-D method using equation Sc = Ss = D Ss with Ss specific entropy to find W' More... | |
| #define | METHODTYPE 3 |
| #define | NEWT_DIM 2 |
| #define | USE_LINE_SEARCH -2 |
| #define | MAX_NEWT_RETRIES 0 /*Max. # of retries of N-R procedure, while increasing guess for W by *10 after each time*/ |
U->P inversion definitions for utoprim_jon.c.
Definition in file utoprim_jon.h.
| #define CRAZYNEWCHECK 0 |
Definition at line 12 of file utoprim_jon.h.
| #define JONGENNEWTSTUFF 1 |
whether to turn on jon's modificatoins to Scott's general Newton's method
Definition at line 10 of file utoprim_jon.h.
| #define MAX_NEWT_RETRIES 0 /*Max. # of retries of N-R procedure, while increasing guess for W by *10 after each time*/ |
Definition at line 43 of file utoprim_jon.h.
| #define METHODTYPE 3 |
Definition at line 39 of file utoprim_jon.h.
| #define NEWT_DIM 2 |
Definition at line 40 of file utoprim_jon.h.
| #define PRIMFROMVSQ 0 |
whether to do final calculation of primitives from v^2 or {u}^2 0: from {u}^2 – avoids catastrophic cancellation at high Lorentz factors 1: from v^2 – requires computing from 1/sqrt(1-v^2) which has problems when v^2 1
Definition at line 22 of file utoprim_jon.h.
| #define USE_LINE_SEARCH -2 |
Definition at line 41 of file utoprim_jon.h.
| #define WHICHCOLDINVERTER 2 |
0: original 1-D but with E' (W and v^2, but can give nan's when Wp<0 slightly – maybe fixed with new normalization?) (apparently also not giving unique solution) 1: Jon's simple 1-D E' (no use of v^2, but obtains an inconsistent v^i[P^2,E']) (apparently also not giving unique solution) Uses same as WHICHHOTINVERTER==3 2: simple 1-D P^2 (no use of v^2, obtains a consistent v^i[P^2]) (but 4 roots and never sure which one is right – not always positive root.
If D<0, then can be negative root) 3: 1-D P^ error on all 3 components of P^i so there is a unique answer (NOT DONE YET)
Definition at line 31 of file utoprim_jon.h.
| #define WHICHENTROPYINVERTER 0 |
0: 1-D method using equation Sc = Ss = D Ss with Ss specific entropy to find W'
Definition at line 35 of file utoprim_jon.h.
| #define WHICHHOTINVERTER 3 |
1: 1-D E'(W') (really has 3 roots, but seemingly always positive-most root found AND correct root) 2: 2-D E'(W',v^2) and P^2(W',v^2) (unsure how many roots) 3: 1-D E'(W') Jon's method without intermediate v^2 and using {u}^2 instead of v^2 – avoids catastrophic cancellation (still has 3 roots)
Definition at line 17 of file utoprim_jon.h.
1.8.3.1
