Local variables. Each process pi manages the following local variables: parti is used to locally store a copy of the snapshot object PART ; counti is a local counter; and groupi a binary variable whose value belongs to {1, 2}. − Behavior of a process pi. Algorithm 1 describes the behavior of a process pi. When it invokes propose(ini) (where ini is the value it proposes), pi first indicates it is participating (line 1). Then it invokes the snapshot object until at least n t processes are participating (lines 2-4). When this oc- curs, pi enters group 1 or group 2 according to the value of its counter counti (line 5), and launches in parallel two threads T 1 and T 2 (line 6). In the thread T 1, pi loop forever until DEC contains a proposed value. When this happens pi decides it (line 7). The execution of return() at line 7 or 12 terminates the invocation of propose(). The thread T 2 is the core of the algorithm. Process pi tries to enter the critical section controlled by either the f -mutex or the m-mutex object MUTEX [groupi] (line 9). If it succeeds and DEC has still its initial default value, pi assigns it the value ini it proposed (line 10). Finally, pi releases the critical section (line 11), and decides (line 12). Let us remind that, as far as MUTEX [1] (respectively, MUTEX [2]) is concerned, up to f (respectively, m) processes can simultaneously execute line 10.
Local variables. Each process pi manages the following local variables: parti is used to locally store a copy of the snapshot object PART ; counti is a local counter; idi and new_namei are used to store the original and new names, respectively. − Behavior of a process pi. Algorithm 2 describes the behavior of a process pi. Every process pi keeps on taking snapshots until it notices that n t processes (including itself) are participating. Then, the process invokes a rename operation of a RENAMINGf object, stores the value of its new name in new_namei, and returns this value.
Local variables. For each iteration j, party Pi maintains several variables: i • L(j) contains the set of values to which Pi is “locked,” or ⊥ if it is not locked. Pi will not commit to any other set in iteration j if it is locked. i • HardLocked(j) ∈ {0, 1} is a boolean variable indicating whether Pi is “hard locked” to the set L(j). When Xx is hard locked to a set, it will commit to the set regardless of any proposal in iteration j. i • T (j) records the sets from “valid proposals” received in iteration j. As part of the security proof, we show that the output must always be such a set. Since, apart from the proposal and preround, the protocol only requires parties to exactly compare sets, it is ok to replace the encodings of sets in these rounds with the hash of the set.
Local variables all omvar integer real i da , , j db , , k e1a , e1b , e2a 1 , e2b , v1a , v1b , v2a , v2b 2 , v3a , b1a , b1b , b2a , v3b 3 *if def,MHD 4 , b2b , q11 , , b3a q12 , , b3b q2 , q3 *endif MHD The following arrays are never used, and are pla ed here only to show how arrays an be de lared and then equivalen ed to "global worker arrays" so that the size of the exe utable is not in reased. real array1d (ijkx) real array2d (idim,jdim) real array3d ( in, jn, kn) equivalen e ( array1d , wa1d ) equivalen e ( array2d , wa2d ) equivalen e ( array3d , wa3d ) ----------------------------------------------------------------------- Input parameters: da , db array and boundary values for density e1a, e1b array and boundary values for first internal energy e2a, e2b array and boundary values for se ond internal energy v1a, v1b array and boundary values for 1-velo ity v2a, v2b array and boundary values for 2-velo ity v3a, v3b array and boundary values for 3-velo ity b1a, b1b array and boundary values for 1-magneti field b2a, b2b array and boundary values for 2-magneti field b3a, b3b array and boundary values for 3-magneti field namelist / pgen / 1 da , db , e1a , e1b , e2a 2 , e2b , v1a , v1b , v2a , v2b 3 , v3a , b1a , b1b , b2a , b2b 4 , b3a , b3b Set default values Read namelist pgen. read (ioin , pgen) write (iolog, pgen) Set field arrays. Metri fa tors in the magneti field settings are ne xxxxxx to preserve the solenoidal ondition. Note that the first internal energy is initialised even if the total energy equation is being solved. If needed, routine TOTNRG is alled by SETUP to initialise the total energy "et". do 30 k=ksmnm2,kemxp3 do 20 j=jsmnm2,jemxp3 do 10 i=ismnm2,iemxp3 d (i,j,k) = da v1(i,j,k) = v1a v2(i,j,k) = v2a *if -def,ISO *endif -ISO v3(i,j,k) = v3a e1(i,j,k) = e1a *if def,TWOFLUID e2(i,j,k) = e2a *endif TWOFLUID *if def,MHD b1(i,j,k) = b1a b2(i,j,k) = b2a * g2bi (i) b3(i,j,k) = b3a * g31bi(i) * g32bi(j) *endif MHD 10 ontinue 20 ontinue 30 ontinue *if -def,ISYM Set inflow boundary arrays. *if def,MHD q11 = ( v2b * b3b - v3b * b2b ) * dx1a(ism1) q12 = ( v2b * b3b - v3b * b2b ) * dx1a(ism2) q2 = ( v3b * b1b - v1b * b3b ) * g2a (is ) q3 = ( v1b * b2b - v2b * b1b ) * g31a(is ) *endif MHD do 50 k=ksmnm2,kemxp3 diib1 (j,k) = db diib2 (j,k) = db v1iib1 (j,k) = v1b v1iib2 (j,k) = v1b v1iib3 (j,k) = v1b v2iib1 (j,k) = v2b v2iib2 (j,k) = v2b v3iib1 (j,k) = v2b v3iib2 (j,k) = v2b *if -def,ISO e1iib1 (j,k) = e1b e1iib2 (j,k) = e1b *endif -ISO *if def,TWOFLUID e2iib1 (j,...
