Changeset 9fa538c


Ignore:
Timestamp:
May 3, 2021, 9:15:34 PM (6 months ago)
Author:
Michael Brooks <mlbrooks@…>
Branches:
arm-eh, jacob/cs343-translation, master, new-ast-unique-expr
Children:
9e2341b4, c0c940a
Parents:
d3ba775
Message:

Fixed new-array subscripting to work on 32-bit builds, partly with a workaround for #247.

Also changed its declarations to static inline, getting them to show equal performance to a C array on a simple test.

Files:
1 edited
2 moved

Legend:

Unmodified
Added
Removed
  • libcfa/src/containers/array.hfa

    rd3ba775 r9fa538c  
    2121    };
    2222
    23     Timmed & ?[?]( arpk(N, S, Timmed, Tbase) & a, ptrdiff_t i ) {
     23    // About the choice of integral types offered as subscript overloads:
     24    // Intent is to cover these use cases:
     25    //    float foo( ptrdiff_t i ) { return a[i]; }           // i : ptrdiff_t
     26    //    forall( [N] ) ... for( i; N ) { total += a[i]; }    // i : typeof( sizeof(42) )
     27    //    for( i; 5 ) { total += a[i]; }                      // i : int
     28    // It gets complicated by:
     29    // -  CFA does overloading on concrete types, like int and unsigned int, not on typedefed
     30    //    types like size_t.  So trying to overload on ptrdiff_t vs int works in 64-bit mode
     31    //    but not in 32-bit mode.
     32    // -  Given bug of Trac #247, CFA gives sizeof expressions type unsigned long int, when it
     33    //    should give them type size_t.
     34    //   
     35    //                          gcc -m32         cfa -m32 given bug         gcc -m64
     36    // ptrdiff_t                int              int                        long int
     37    // size_t                   unsigned int     unsigned int               unsigned long int
     38    // typeof( sizeof(42) )     unsigned int     unsigned long int          unsigned long int
     39    // int                      int              int                        int
     40
     41    static inline Timmed & ?[?]( arpk(N, S, Timmed, Tbase) & a, int i ) {
    2442        return (Timmed &) a.strides[i];
    2543    }
    2644
    27     Timmed & ?[?]( arpk(N, S, Timmed, Tbase) & a, int i ) {
     45    static inline Timmed & ?[?]( arpk(N, S, Timmed, Tbase) & a, unsigned int i ) {
    2846        return (Timmed &) a.strides[i];
    2947    }
    3048
    31     Timmed & ?[?]( arpk(N, S, Timmed, Tbase) & a, size_t i ) {
     49    static inline Timmed & ?[?]( arpk(N, S, Timmed, Tbase) & a, long int i ) {
    3250        return (Timmed &) a.strides[i];
    3351    }
    3452
    35     size_t ?`len( arpk(N, S, Timmed, Tbase) & a ) {
     53    static inline Timmed & ?[?]( arpk(N, S, Timmed, Tbase) & a, unsigned long int i ) {
     54        return (Timmed &) a.strides[i];
     55    }
     56
     57    static inline size_t ?`len( arpk(N, S, Timmed, Tbase) & a ) {
    3658        return z(N);
    3759    }
    3860
    3961    // workaround #226 (and array relevance thereof demonstrated in mike102/otype-slow-ndims.cfa)
    40     void ?{}( arpk(N, S, Timmed, Tbase) & this ) {
     62    static inline void ?{}( arpk(N, S, Timmed, Tbase) & this ) {
    4163        void ?{}( S (&inner)[z(N)] ) {}
    4264        ?{}(this.strides);
    4365    }
    44     void ^?{}( arpk(N, S, Timmed, Tbase) & this ) {
     66    static inline void ^?{}( arpk(N, S, Timmed, Tbase) & this ) {
    4567        void ^?{}( S (&inner)[z(N)] ) {}
    4668        ^?{}(this.strides);
     
    5375
    5476forall( Te )
    55 Te mkar_( tag(Te) ) {}
     77static inline Te mkar_( tag(Te) ) {}
    5678
    5779forall( [N], ZTags ... , Trslt &, Tatom & | { Trslt mkar_( tag(Tatom), ZTags ); } )
    58 arpk(N, Trslt, Trslt, Tatom) mkar_( tag(Tatom), tag(N), ZTags ) {}
     80static inline arpk(N, Trslt, Trslt, Tatom) mkar_( tag(Tatom), tag(N), ZTags ) {}
    5981
    6082// based on https://stackoverflow.com/questions/1872220/is-it-possible-to-iterate-over-arguments-in-variadic-macros
     
    90112
    91113forall( TA &, TB &, TC &, IxAB, IxBC ... | { TB & ?[?]( TA &, IxAB ); TC & ?[?]( TB &, IxBC ); } )
    92 TC & ?[?]( TA & this, IxAB ab, IxBC bc ) {
     114static inline TC & ?[?]( TA & this, IxAB ab, IxBC bc ) {
    93115    return this[ab][bc];
    94116}
     
    99121
    100122forall( TA &, TB &, TC &, IxAB_0, IxBC | { TB & ?[?]( TA &, IxAB_0 ); TC & ?[?]( TB &, IxBC ); } )
    101 TC & ?[?]( TA & this, IxAB_0 ab, IxBC bc ) {
     123static inline TC & ?[?]( TA & this, IxAB_0 ab, IxBC bc ) {
    102124    return this[ab][bc];
    103125}
    104126
    105127forall( TA &, TB &, TC &, IxAB_0, IxAB_1, IxBC | { TB & ?[?]( TA &, IxAB_0, IxAB_1 ); TC & ?[?]( TB &, IxBC ); } )
    106 TC & ?[?]( TA & this, IxAB_0 ab0, IxAB_1 ab1, IxBC bc ) {
     128static inline TC & ?[?]( TA & this, IxAB_0 ab0, IxAB_1 ab1, IxBC bc ) {
    107129    return this[[ab0,ab1]][bc];
    108130}
    109131
    110132forall( TA &, TB &, TC &, IxAB_0, IxAB_1, IxAB_2, IxBC | { TB & ?[?]( TA &, IxAB_0, IxAB_1, IxAB_2 ); TC & ?[?]( TB &, IxBC ); } )
    111 TC & ?[?]( TA & this, IxAB_0 ab0, IxAB_1 ab1, IxAB_2 ab2, IxBC bc ) {
     133static inline TC & ?[?]( TA & this, IxAB_0 ab0, IxAB_1 ab1, IxAB_2 ab2, IxBC bc ) {
    112134    return this[[ab0,ab1,ab2]][bc];
    113135}
     
    121143// Base
    122144forall( [Nq], [Sq], Tbase & )
    123 tag(arpk(Nq, Sq, Tbase, Tbase)) enq_( tag(Tbase), tag(Nq), tag(Sq), tag(Tbase) ) {}
     145static inline tag(arpk(Nq, Sq, Tbase, Tbase)) enq_( tag(Tbase), tag(Nq), tag(Sq), tag(Tbase) ) {}
    124146
    125147// Rec
    126148forall( [Nq], [Sq], [N], [S], recq &, recr &, Tbase & | { tag(recr) enq_( tag(Tbase), tag(Nq), tag(Sq), tag(recq) ); } )
    127 tag(arpk(N, S, recr, Tbase)) enq_( tag(Tbase), tag(Nq), tag(Sq), tag(arpk(N, S, recq, Tbase)) ) {}
     149static inline tag(arpk(N, S, recr, Tbase)) enq_( tag(Tbase), tag(Nq), tag(Sq), tag(arpk(N, S, recq, Tbase)) ) {}
    128150
    129151// Wrapper
    130152struct all_t {} all;
    131153forall( [N], [S], Te &, result &, Tbase & | { tag(result) enq_( tag(Tbase), tag(N), tag(S), tag(Te) ); } )
    132 result & ?[?]( arpk(N, S, Te, Tbase) & this, all_t ) {
     154static inline result & ?[?]( arpk(N, S, Te, Tbase) & this, all_t ) {
    133155    return (result&) this;
    134156}
Note: See TracChangeset for help on using the changeset viewer.