IRIX Base Documentation 2002 November

home *** CD-ROM | disk | FTP | other *** search

/ IRIX Base Documentation 2002 November / SGI IRIX Base Documentation 2002 November.iso / usr / share / catman / u_man / cat1 / pca.z / pca

Wrap

Text File | 2002-10-03 | 46.7 KB | 886 lines

PCA(1) Last changed: 3-5-99 NNAAMMEE ppccaa - Power C Analyzer SSYYNNOOPPSSIISS //uussrr//lliibb//ppccaa [ _o_p_t_i_o_n_s ] ... //uussrr//lliibb6644//ccmmppllrrss//ppccaa [ _o_p_t_i_o_n_s ] ... IIMMPPLLEEMMEENNTTAATTIIOONN IRIX systems (--oo3322 ABI only) DDEESSCCRRIIPPTTIIOONN For --nn3322 and --6644 ABIs, ppccaa has been replaced by the --aappoo option. ppccaa is a source-to-source optimizing C preprocessor that discovers parallelism in C code. ppccaa concurrentizes C by restructuring certain portions of code and then adding parallel programming directives where possible. The amount of code that is concurrentized varies depending on the value of certain ppccaa command line options, directives, and assertions. For more information, refer to the _I_R_I_S _P_o_w_e_r _C _U_s_e_r'_s _G_u_i_d_e. ppccaa is usually invoked as an option to the cccc(1) command, although it can be run separately. When ppccaa is used as part of a cccc compilation, the ppccaa _o_p_t_i_o_ns must be passed via the --WWKK mechanism. See the cccc(1) man page for details of the --WW option. ppccaa requires a single input source file, which can be specified using the --ii option. Specifying a filename as a command-line argument is also sufficient, as any argument not recognized as an option is treated as an input filename. ppccaa can produce two types of output. The first type is the concurrentized C source code, and the second type is the annotated listing, whose contents vary depending on the value of the --lloo command line option. Filenames for these two types of output can be set using the --ccmmpp and --ll options, respectively. If the --ccmmpp option is not specified, the output code will be sent to standard output. If no --ll option is specified, and some type of listing information is requested using the --lloo option, then this listing information will also be sent to standard output. ppccaa accepts the following command line options: --aarrll==<_i_n_t_e_g_e_r> Long name: --aaddddrreessss__rreessoolluuttiioonn__lleevveell==<_i_n_t_e_g_e_r> Default value: --aaddddrreessss__rreessoolluuttiioonn__lleevveell==11 The --aarrll option controls the assumptions ppccaa makes about memory aliases. The integer value, which ranges from 0 (no assumptions) to 4 (assume everything), specifies the level of assumptions that ppccaa is to make about the behavior of the code. Each level is cumulative (i.e., level 4 also makes all the assumptions of the lower levels). The following are the assumption level definitions: 00 No assumptions are made. 11 Assumes that no pointer self references. 22 Assumes that function arguments are distinct from each other. 33 Assumes that local pointers and arrays are distinct from global pointers and arrays. 44 Assumes that all pointers and arrays are distinct from each other. --aarrccllmm==<_i_n_t_e_g_e_r> Long name: --aarrcclliimmiitt==<_i_n_t_e_g_e_r> Default value: --aarrcclliimmiitt==55000000 The --aarrcclliimmiitt option sets the size of the dependence arc data structure that ppccaa uses to perform data dependence analysis. This data structure is dynamically allocated on a loop nest by loop nest basis. By default, this data structure is allocated with a size = max (# of statements * 4, --aarrcclliimmiitt value). If a loop contains too many dependence relationships and cannot be represented in the dependence data structure, ppccaa will surrender optimization of the loop. You may use the --aarrcclliimmiitt option to increase the size of the data structure to enable ppccaa to perform more optimizations. (Most users do NOT need to change this value.) --cchhll==<_i_n_t_e_g_e_r> Long name: --ccaacchheelliinnee==<_i_n_t_e_g_e_r> Default value: --ccaacchheelliinnee==6644 The --ccaacchheelliinnee option informs ppccaa of the width of the memory channel (in bytes) between cache and main memory. --cchhss==<_i_n_t_e_g_e_r> Long name: --ccaacchheessiizzee==<_i_n_t_e_g_e_r> Default value: --ccaacchheessiizzee==6644 The --ccaacchheessiizzee option informs ppccaa of the size (in kilobytes) of the cache memory. --ccppllcc==<_i_n_t_e_g_e_r>[,<_i_n_t_e_g_e_r>] Long name: --ccaacchhee__pprreeffeettcchh__lliinnee__ccoouunntt==<_i_n_t_e_g_e_r>[,<_i_n_t_e_g_e_r>] Default value: --ccaacchhee__pprreeffeettcchh__lliinnee__ccoouunntt==00 The --ccppllcc option informs ppccaa of the number of additional cache lines that are fetched during a cache miss. The first number corresponds to the primary cache, and the second number, if specified, corresponds to the secondary cache. --[[nn]]ccmmpp[[==<_f_i_l_e>] Long name: --[[nnoo]]ccmmpp[[==<_f_i_l_e>] Default value: ssttaannddaarrdd oouuttppuutt The --ccmmpp (compilable) option instructs ppccaa to place the optimized C program in a specified transformed program file. If --ccmmpp==<_f_i_l_e> is specified, the transformed C is written to that file. If --ccmmpp is specified, the transformed code is written to _f_i_l_e_n_a_m_e..mm, where _f_i_l_e_n_a_m_e is the input file name with any trailing ..cc removed. If no --ccmmpp option is specified, the transformed code is written to standard output. To disable generation of the C output file, specify --nnooccmmpp on the command line. --[[nn]]ccoonncc Long name: --[[nnoo]]ccoonnccuurrrreennttiizzee Default value: --ccoonnccuurrrreennttiizzee This command line option directs ppccaa to perform concurrentization, while --nnooccoonncc directs ppccaa to perform only scalar optimizations. --[[nn]]ccpp==<_l_i_s_t> Long name: --[[nnoo]]ccmmppooppttiioonnss==<_l_i_s_t> Default value: --nnooccmmppooppttiioonnss The --ccmmppooppttiioonnss option specifies additional information for inclusion in the transformed code file. The following option can be specified: ii Insert line numbers that reference the original source. --ddoollllaarr No short name. Default value: ooffff The --ddoollllaarr option allows dollar signs to be used in identifiers under both Kernighan and Ritchie and ANSI C. --ddpprr==<_i_n_t_e_g_e_r> Long name: --ddpprreeggiisstteerrss==<_i_n_t_e_g_e_r> Default value: --ddpprreeggiisstteerrss==1122 The --ddpprreeggiisstteerrss option specifies the number of double precision floating point registers each processor has available for expression evaluation. --eeiiiiffgg==<_i_n_t_e_g_e_r> Long name: --eeaacchh__iinnvvaarriiaanntt__iiff__ggrroowwtthh==<_i_n_t_e_g_e_r> Default value: --eeaacchh__iinnvvaarriiaanntt__iiff__ggrroowwtthh==2200 The --eeaacchh__iinnvvaarriiaanntt__iiff__ggrroowwtthh option controls the rewriting of iiffs nested within loops. The value of the option (which must be in the range 0 to 100) specifies a limit on the number of executable statements in the nested iiff. If the number of statements in the iiff exceeds this limit, then ppccaa will not rewrite the code. If there are fewer statements, then ppccaa will interchange the loop and if statements for improved execution speed. --ffllooaatt No short name. Default value: ooffff In Kernighan and Ritchie C, all variables declared as type ffllooaatt are promoted to type ddoouubbllee by default. The --ffllooaatt option prevents this promotion to double. This option is ignored under --ssyynnttaaxx==aa, because ANSI C does not promote variables of type ffllooaatt to ddoouubbllee. --ffpprr==<_i_n_t_e_g_e_r> Long name: --ffpprreeggiisstteerrss==<_i_n_t_e_g_e_r> Default value: --ffpprreeggiisstteerrss==1122 The --ffpprreeggiisstteerrss option specifies the number of single precision floating point registers each processor has available for expression evaluation. --[[nn]]ffuussee Long name: --[[nnoo]]ffuussee Default value: --ffuussee This command line option enables lloooopp ffuussiioonn, a conventional compiler optimization that transforms two adjacent loops into a single loop. Setting --ssccaallaarroopptt to at least 2, --ooppttiimmiizzee==55 is also required to enable loop fusion. --hheeaapp==<_l_i_s_t> Long name: --hheeaapplliimmiitt==<_i_n_t_e_g_e_r> Default value: --hheeaapplliimmiitt==110000 The --hheeaapplliimmiitt option specifies the maximum size in megabytes to which the memory heap is allowed to grow to when ppccaa processes a source file. If this limit is reached, ppccaa will stop processing the source code. --hhllii==<_l_i_s_t> Long name: --hhooiisstt__lloooopp__iinnvvaarriiaannttss==<_i_n_t_e_g_e_r> Default value: --hhooiisstt__lloooopp__iinnvvaarriiaannttss==11 This option controls code hoisting of loop-invariant expressions from loops. Note that this switch is independent of the switches that control the floating of invariant-IFs out of loops, --eeaacchh__iinnvvaarriiaanntt__iiff__ggrroowwtthh and --mmaaxx__iinnvvaarriiaanntt__iiff__ggrroowwtthh. This option accepts the following settings: 00 Turns off the hoisting of invariant code from loops. 11 Floats all loop invariant expressions that are not under the control of an IF-structure within the given loop nest. This is the default setting. 22 The same behavior as level 1. 33 Floats all loop-invariant expressions from loops. If there is invariant code that is protected by an IF- structure and the hoisting value is less than 3, then ppccaa will generate a message to that effect in your output listing. --ii==<_f_i_l_e> Long name: --iinnppuutt==<_f_i_l_e> Default value: nnoonnee The --ii option instructs ppccaa to read its input code from the specified file. --iinnll==<_l_i_s_t> Long name: --iinnlliinnee==<_l_i_s_t> Default value: ooffff --iinnllcc==<_l_i_s_t> Long name: --iinnlliinnee__aanndd__ccooppyy==<_l_i_s_t> Default value: ooffff --iippaa==<_l_i_s_t> Long name: --iippaa==<_l_i_s_t> Default value: ooffff The --iinnlliinnee, --iinnlliinnee__aanndd__ccooppyy, and --iippaa options provide ppccaa with a list of routines to analyze. If the option is specified without an argument list, ppccaa will try to inline/analyze all the called functions in the _i_n_l_i_n_i_n_g _u_n_i_v_e_r_s_e (specified by the --iinnlliinnee__ffrroomm__ffiilleess, --iinnlliinnee__ffrroomm__lliibbrraarriieess, --iippaa__ffrroomm__ffiilleess, or --iippaa__ffrroomm__lliibbrraarriieess options). If a list of names is included, for example, --iinnlliinnee==mmkkccooeeff,,yyvvaall, then just the routines named will be inlined/analyzed. The --iinnlliinnee and --iippaa command line options are overridden by the ##pprraaggmmaa iinnlliinnee and ##pprraaggmmaa iippaa directives. The --iinnlliinnee__aanndd__ccooppyy option functions like the --iinnlliinnee option, except that if all references to a function are inlined, the text of the routine is not optimized, but is copied unchanged to the transformed code file. This is intended for use when inlining routines from the same file as the call, and has no special effect when the routines being inlined are being taken from a library or another source file. When a subprogram has been inlined everywhere it is used, leaving it unoptimized saves compilation time. When a program involves multiple source files, the unoptimized routine will still be available in case one of the other source files contains a reference to it, so no errors will result. NOTE: the iinnlliinnee__aanndd__ccooppyy algorithm assumes that all calls and references to the routine precede it in the source file. If the routine is referenced after the text of the routine, and that particular call site cannot be inlined, the unoptimized version of the routine will be invoked. --iinnccrr==<_f_i_l_e> Long name: --iinnlliinnee__ccrreeaattee==<_f_i_l_e> Default value: ooffff --iippaaccrr==<_f_i_l_e> Long name: --iippaa__ccrreeaattee==<_f_i_l_e> Default value: ooffff These options instruct ppccaa to build a library file containing partially-analyzed routines for later inlining/analyzing. The library created is used with the --iinnlliinnee__ffrroomm__lliibbrraarriieess or --iippaa__ffrroomm__lliibbrraarriieess option. Libraries created with --iinnlliinnee__ccrreeaattee can be used with either inlining or interprocedural analysis, since they contain essentially complete descriptions of the functions included. Libraries created with --iippaa__ccrreeaattee can be used only with interprocedural analysis, since they do not have the complete text of the functions, just the data relationships information. Any filename can be used for the library name. An extension ..kklliibb is preferred, for maximum compatibility with the --iinnlliinnee__ffrroomm__lliibbrraarriieess and --iippaa__ffrroomm__lliibbrraarriieess options. --iinndd==<_i_n_t_e_g_e_r> Long name: --iinnlliinnee__ddeepptthh==<_i_n_t_e_g_e_r> Default value: --iinnlliinnee__ddeepptthh==22 This option sets the maximum level of subprogram nesting that ppccaa will attempt to inline. Higher values instruct ppccaa to trace function calls further and thus produce more inlined code. The ##pprraaggmmaa [[nnoo]]iinnlliinnee directive, when enabled, is not affected by the --iinnlliinnee__ddeepptthh restrictions. The argument values and their meanings are: --11 Inline only routines that do not contain procedure or function calls. 00 Use the default value (2). 11--1100 Inline routines to this depth. --iinnffff==<_l_i_s_t> Long name: --iinnlliinnee__ffrroomm__ffiilleess==<_l_i_s_t> Default value: ccuurrrreenntt ssoouurrccee ffiillee --iinnffll==<_l_i_s_t> Long name: --iinnlliinnee__ffrroomm__lliibbrraarriieess==<_l_i_s_t> Default value: ooffff --iippaaffff==<_l_i_s_t> Long name: --iippaa__ffrroomm__ffiilleess==<_l_i_s_t> Default value: ccuurrrreenntt ssoouurrccee ffiillee --iippaaffll==<_l_i_s_t> Long name: --iippaa__ffrroomm__lliibbrraarriieess==<_l_i_s_t> Default value: ooffff These options provide ppccaa with the locations of functions available for inlining/interprocedural analysis. (The total set of available functions is called the inlining (or IPA) universe.) The --iinnlliinnee__ffrroomm__ffiilleess and --iippaa__ffrroomm ffiilleess options take the names of source files and directories containing source files. Including a directory, for example, --iippaaffff==//uussrr//iippaalliibb, is equivalent to the UNIX notation //uussrr//iippaalliibb//**..cc. (Do not use shell wild card characters in the list of files and directories.) Note that the specified files must have already been pre-processed by ccpppp or aaccpppp (if necessary). The --iinnlliinnee__ffrroomm__lliibbrraarriieess and --iippaa__ffrroomm__lliibbrraarriieess options take the names of libraries created with the --iinnlliinnee__ccrreeaattee and --iippaa__ccrreeaattee options and directories containing such libraries. In directories, the ppccaa libraries are identified by the extension ..kklliibb. Multiple files/libraries or directories may be specified in one option, separated by commas. Multiple occurrences of these options may be specified on the command line. --iinnllll==<_i_n_t_e_g_e_r> Long name: --iinnlliinnee__lloooopplleevveell==<_i_n_t_e_g_e_r> Default value: --iinnllll==22 --iippaallll==<_i_n_t_e_g_e_r> Long name: --iippaa__lloooopplleevveell==<_i_n_t_e_g_e_r> Default value: --iippaallll==22 The --iinnlliinnee__lloooopplleevveell and --iippaa__lloooopplleevveell options enable the user to limit inlining to just functions which are referenced in nested loops, where the effects of reduced function call overhead or enhanced optimizations will be multiplied. The parameter is defined from the most deeply nested function reference. --iinnllll==11 restricts inlining to functions referenced in the deepest loop nest. --iinnllll==33 restricts inlining to those routines referenced at the three deepest levels. The ffoorr loop nest level of each function reference is included in the optional calling tree section of the listing files. The ##pprraaggmmaa [[nnoo]]iinnlliinnee and ##pprraaggmmaa [[nnoo]]iippaa directives, when enabled, are not affected by the --iinnlliinnee__lloooopplleevveell and --iippaa__lloooopplleevveell restrictions. --iinnmm Long name: --iinnlliinnee__mmaannuuaall Default value: ooffff --iippaamm Long name: --iippaa__mmaannuuaall Default value: ooffff These options instruct ppccaa to recognize the ##pprraaggmmaa [[nnoo]]iinnlliinnee and ##pprraaggmmaa [[nnoo]]iippaa directives. This allows manual control over which functions are inlined/analyzed at which call sites. The default is to ignore these ##pprraaggmmaa directives. When --iinnlliinnee__mmaann or --iippaa__mmaann is included on the command line, the ##pprraaggmmaa directives are enabled. Because ##pprraaggmmaa [[nnoo]]iinnlliinnee and ##pprraaggmmaa [[nnoo]]iippaa are not affected by the --iinnlliinnee__lloooopplleevveell and --iippaa__lloooopplleevveell command line options, they can be used with command line control to select functions or call sites that the regular selection algorithm would reject. --llmm==<_i_n_t_e_g_e_r> Long name: --lliimmiitt==<_i_n_t_e_g_e_r> Default value: --lliimmiitt==55000000 To reduce the compile time, ppccaa estimates how long it takes to analyze each loop nest construct. If a loop is too deeply nested, ppccaa ignores the outer loop and recursively visits the inner loops. The loop nest limit can be used to control what ppccaa considers too deeply nested. For more information, refer to the _I_R_I_S _P_o_w_e_r _C _U_s_e_r'_s _G_u_i_d_e. --llnn==<_i_n_t_e_g_e_r> Long name: --lliinneess==<_i_n_t_e_g_e_r> Default value: --lliinneess==5555 This option enables the ppccaa listing to be paginated for printing. The number of lines per page on the listing may be changed by using the --lliinneess option. The --lliinneess==00 option directs ppccaa to paginate only at subroutine boundaries. --[[nn]]ll[[==<_f_i_l_e>]] Long name: --[[nnoo]]lliisstt[[==<_f_i_l_e>]] Default value: ssttaannddaarrdd oouuttppuutt If some type of listing information is requested using the --lloo option, the --lliisstt option specifies the destination of that listing. If --lliisstt==<_f_i_l_e> is specified, the listing is written to the specified file. If --lliisstt is specified without a filename, the listing file is written to _f_i_l_e_n_a_m_e..llsstt, where _f_i_l_e_n_a_m_e is the input file name with any trailing ..cc removed. If no --lliisstt option is specified, any listing information requested is written to the standard output. To disable generation of any listing information, enter --nnoolliisstt on the command line. --lloo==<_l_i_s_t> Long name: --lliissttooppttiioonnss==<_l_i_s_t> Default value: nnoo lliissttiinngg The --lliissttooppttiioonnss option tells ppccaa what information to include in the listing file. --lliissttooppttiioonnss can enable the following options to be included in the listing: cc Calling tree. kk Print ppccaa options active within the program unit. ll Loop-by-loop optimization table. nn Program unit names as processed (written to error file). pp Compilation performance statistics. ss Summary of the optimizations performed. --llww==<_i_n_t_e_g_e_r> Long name: --lliissttiinnggwwiiddtthh==<_i_n_t_e_g_e_r> Default value: --lliissttiinnggwwiiddtthh==8800 This option sets the maximum line length for the listing file. This setting affects only the format of the loop summary table (--lloo==ll) and the options table (--lloo==kk). The only permissible values for --lliissttiinnggwwiiddtthh are 80 and 132. --[[nn]]mmaa==<_l_i_s_t> Long name: --[[nnoo]]mmaacchhiinnee==<_l_i_s_t> Default value: --mmaacchhiinnee==ss The --mmaacchhiinnee option is list-valued with three choices which may be set. The possible choices are: nn Tells ppccaa to concurrentize loops which generate non- stride-1 memory references. oo Tells ppccaa to only concurrentize outer loops. This capability is available to prevent concurrentization on applications that have small inner loop bounds, thereby reducing overhead costs. ppccaa determines concurrentization based on the overhead to benefit ratio. When the loop bounds are unknown at compile time, ppccaa may generate concurrent code for innermost loops, a practice that may be inefficient for the actual loop bounds. ss Tells ppccaa to concurrentize ffoorr loops which generate stride-1 memory references. The ss and nn choices may not be set simultaneously. --mmiiiiffgg==<_i_n_t_e_g_e_r> Long name: --mmaaxx__iinnvvaarriiaanntt__iiff__ggrroowwtthh==<_i_n_t_e_g_e_r> Default value: --mmaaxx__iinnvvaarriiaanntt__iiff__ggrroowwtthh==550000 The --mmaaxx__iinnvvaarriiaanntt__iiff__ggrroowwtthh option controls the rewriting of iiff statements nested within loops. The value of the option (which must be in the range 0 to 1000) specifies a limit on the total number of additional executable statements which can be generated by this optimization. If the total number of extra statements generated exceeds this limit, then ppccaa will stop rewriting the code. If there are fewer statements, then ppccaa will continue to interchange the loop and iiff statements for improved execution speed. --mmcc==<_i_n_t_e_g_e_r> Long name: --mmiinnccoonnccuurrrreenntt==<_i_n_t_e_g_e_r> Default value: --mmiinnccoonnccuurrrreenntt==11000000 The --mmiinnccoonnccuurrrreenntt option sets the minimum level of work which must be done by a loop in order for it to be a candidate for concurrentization. Loops that contain less work than --mmiinnccoonnccuurrrreenntt will not be concurrentized. Setting --mmiinnccoonnccuurrrreenntt==00 effectively allows all loops (if all other necessary conditions are met) to be concurrentized. --[[nn]]nnaammeeppaarrtt[[==<_i_n_t_e_g_e_r>,<_i_n_t_e_g_e_r>]] Long name: --[[nnoo]]nnaammeeppaarrttiittiioonniinngg[[==<_i_n_t_e_g_e_r>,<_i_n_t_e_g_e_r>]] Default value: --nnoonnaammeeppaarrttiittiioonniinngg This option looks at distinct array names and limits the number of arrays that that are referenced within a loop in order to avoid cache conflicts among them. For example, --nnaammeeppaarrtt can be used to break a loop that contains references to arrays aa and bb into two loops, one referencing array aa and the other referencing array bb. The two integer arguments specify the preferred minimum and maximum numbers of distinct arrays referenced in each distributed loop. If they are not specified, they default to 2 and 6. The --nnnnaammeeppaarrtt option turns off name partitioning. --ssccaallaarroopptt must be set to at least 3 for nnaammeeppaarrttiittiioonniinngg to work. --oo==<_i_n_t_e_g_e_r> Long name: --ooppttiimmiizzee==<_i_n_t_e_g_e_r> Default value: --ooppttiimmiizzee==55 The --ooppttiimmiizzee option sets the optimization level. Each optimization level is cumulative (i.e., level 5 performs everything up to and including this level). The meaning of the optimization levels are as follows: 00 No optimization is done. 11 Only simple optimizations are done. Induction variable recognition is enabled. 22 Recognizes reductions in concurrent loops. Performs lifetime analysis to determine when last-value assignment of scalars is necessary. Performs more powerful data dependence tests to find parallelism. 33 Special techniques are used to break data dependence cycles that otherwise prevent concurrentization. Triangular loops are recognized and loop interchanging will be attempted to improve memory referencing. Special case data dependence tests are used. Special index sets, called wrap-around variables, are also recognized. 44 Two versions of a loop are generated, if necessary, to break a data dependence arc. Exact data dependence tests are used to allow more parallelism to be discovered. 55 Array expansion and loop fusion are enabled. --rr==<_i_n_t_e_g_e_r> Long name: --rroouunnddooffff==<_i_n_t_e_g_e_r> Default value: --rroouunnddooffff==00 The --rroouunnddooffff option allows the user to specify the change from serial roundoff error that is tolerable. If an arithmetic reduction is accumulated in a different order than in the scalar program, the roundoff error is accumulated differently and the final result may differ from that of the original program's output. Although the difference is usually insignificant, certain restructuring transformations performed by ppccaa must be disabled to obtain exactly the same answers as the scalar program. ppccaa classifies its transformations by the amount of difference in roundoff error that can accumulate so the user can decide what level of roundoff error differences is allowable. The meanings of the switch values are as follows: 00 Allow no roundoff-changing transformations. Arithmetic recurrences and reductions are not concurrentized. Non-arithmetic reductions may still be concurrentized. Thus, the answers do not depend on the number of processors used. 11 Enable expression simplification which might generate overflow or underflow errors differently. Enable simplification of expressions with operands between binary and unary operators. Perform expression simplification which is exposed due to forward substitution. Enable code floating if the --ssccaallaarrooppttiimmiizzee option is greater than or equal to 1. 22 Allow loop interchanging around arithmetic reductions to be recognized. Perform concurrent reductions with pre-scheduled concurrent loops and local accumulation of reduction results. 33 Recognize real (float) induction variables. Enable sum reductions. Enable memory management optimizations if ssccaallaarrooppttiimmiizzee=3. --rroouuttiinnee==<_r_o_u_t_i_n_e>[[,,<_r_o_u_t_i_n_e>......]] --sskkiipp This option pair allows the user to request that ppccaa perform no optimization on the specified routines. The code in these routines will be written unchanged to the output file. --ssiiggnneedd No short name. Default value: ooffff By default, a variable declared as cchhaarr is handled as an uunnssiiggnneedd cchhaarr. The --ssiiggnneedd option causes a variable declared as cchhaarr to be handled as ssiiggnneedd cchhaarr. This option is necessary when porting code from platforms that have a C compiler that defaults cchhaarr to ssiiggnneedd cchhaarr. --ssoo==<_i_n_t_e_g_e_r> Long name: --ssccaallaarrooppttiimmiizzee==<_i_n_t_e_g_e_r> Default value: --ssccaallaarrooppttiimmiizzee==33 The --ssccaallaarrooppttiimmiizzee option sets the level of dusty-deck and other serial transformations performed. The following are the possible values for <_i_n_t_e_g_e_r>: 00 No scalar optimizations are performed. 11 Only simple scalar optimizations are performed - dead code elimination, global forward substitution and dusty-deck IF transformations. 22 Full range of scalar optimizations are performed - invariant IF's are floated out of loops, loop rerolling/unrolling, array expansion, loop fusion, loop peeling and induction variable recognition. 33 Memory management optimizations are enabled if rroouunnddooffff=3. 44 Reruns dead code elimination optimization after other optimizations. Sometimes the other optimizations expose more dead code which can be removed. --ssaasscc==<_i_n_t_e_g_e_r> Long name: --sseettaassssoocciiaattiivviittyy==<_i_n_t_e_g_e_r> Default value: --sseettaassssoocciiaattiivviittyy==11 The --sseettaassssoocciiaattiivviittyy option provides information on the mapping of physical addresses in main memory to cache pages. The default value (1) allows a datum in main memory to be placed in any of four places in cache. --sspprr==<_i_n_t_e_g_e_r> Long name: --sspprreeggiisstteerrss==<_i_n_t_e_g_e_r> Default value: --sspprreeggiisstteerrss==1122 The --sspprreeggiisstteerrss option specifies the number of single precision floating point registers each processor has available for expression evaluation. --ssttddiioo No short name. Default value: ooffff The --ssttddiioo option instructs ppccaa to perform strength reduction on calls to certain functions in the standard I/O library. Programs that use functions such as pprriinnttff or ssccaannff heavily will generally have improved I/O performance when this switch is used. The --ssccaallaarrooppttiimmiizzee=3 option is required to enable these transformations. Use of this option may produce code containing calls to functions defined in lliibbkkaappiioo..aa, a library of optimized low-level I/O routines. As a result, when linking your program using cccc(1) or lldd(1), you may have to include this library, using the command-line option --llkkaappiioo. If you use cccc(1) with a --ppccaa option to link your program, this is done automatically. If you explicitly specify that I/O using ssccaannff or ffssccaannff be done in parallel (for example, by using the ##pprraaggmmaa ccoonnccuurrrreenntt ccaallll directive), using --ssttddiioo might cause your program to yield incorrect results. To determine if your program might have this problem, search the ppccaa output code (if you are using cccc(1), use the --ppccaa kkeeeepp option to put the output code in a ..MM file) for the functions ____KKIIOOrrdd, ____KKIIOOrrff, and ____KKIIOOrrii. If you find these functions (from optimized calls to ssccaannff or ffssccaannff on doubles, floats, and integers, respectively), the I/O read operations involved will not be semaphored properly. This is only a problem if they occur in a parallel region, and operate on files that other threads may also be operating on concurrently. If you do no explicit specification of parallel regions, use of the --ssttddiioo option is safe. --ssuu==<_l_i_s_t> Long name: --ssuupppprreessss==<_l_i_s_t> Default value: nnoo mmeessssaaggee ssuupppprreessssiioonn The --ssuupppprreessss option tells ppccaa to disable the printing of individual classes of messages. These message classes range from syntax warning and error messages to messages about the optimizations performed. The following codes can be specified with the --ssuupppprreessss option: dd Data dependence messages. ee Syntax error messages. ii Informational messages. nn "Not optimized" messages. qq Questions. ss Standardized messages. ww Syntax warning messages. --ssyy==<_l_i_s_t> Long name: --ssyynnttaaxx==<_l_i_s_t> Default value: --ssyynnttaaxx==aa The --ssyynnttaaxx option tells ppccaa to check for compliance with certain syntactic rules. The --ssyynnttaaxx options are: aa Check for compliance with the SGI-extended Ansi C standard. kk Check for compliance with the SGI-enhanced Kernighan & Richie C specification. --uurr==<_i_n_t_e_g_e_r> Long name: --uunnrroollll==<_i_n_t_e_g_e_r> Default value: --uunnrroollll==44 --uurr22==<_i_n_t_e_g_e_r> Long name: --uunnrroollll22==<_i_n_t_e_g_e_r> Default value: --uunnrroollll22==110000 --uurr33==<_i_n_t_e_g_e_r> Long name: --uunnrroollll33==<_i_n_t_e_g_e_r> Default value: --uunnrroollll33==11 The --uunnrroollll, --uunnrroollll22, and --uunnrroollll33 options control how ppccaa unrolls inner loops. The value of the --ssccaallaarrooppttiimmiizzee option must be at least 2 in order for unrolling to be performed. The unrolling that ppccaa performs is also subject to the following constraints imposed by these three options: * The number of times a loop is unrolled will not be larger than the value given by the --uunnrroollll option. If the value given is zero, the default value is used instead. * The work done by the loop body is estimated by counting operands and operators. This estimate of the amount of work done by the unrolled loop body must be less than the value of the --uunnrroollll22 option and greater than the value of the --uunnrroollll33 option. The reason for the --uunnrroollll and --uunnrroollll22 options is that beyond a certain point, further unrolling of a loop, while continuing to increase code size, brings diminishing returns in performance since the work being done in the loop body is already large compared to the overhead associated with a loop iteration, which unrolling is attempting to amortize. The reason for the --uunnrroollll33 option is to prevent unrolling of very small loops which might be better unrolled or software-pipelined by the C compiler. Since these compiler optimizations are not on by default, the default value for this option is 1. If you would like to take advantage of these C compiler capabilities, you should set this option to a higher value. If the constraints imposed by the option values cannot be met for a given loop, then it will not be unrolled. If they can be met, then the largest degree of unrolling that satisfies the constraints will be used. --vvoollaattiillee No short name. Default value: ooffff The --vvoollaattiillee option indicates that all variables are implicitly volatile. Use of this option severely limits the optimization that can be done. --[[nn]]6644 Long name: --[[nnoo]]6644 Default value: ooffff This option tells ppccaa that the code is being compiled for a 64-bit machine. This affects the size of certain C data types. FFIILLEESS //uussrr//lliibb//ppccaa The pca program //uussrr//lliibb6644//ccmmppllrrss//ppccaa (32-bit and 64-bit versions) ffiillee..cc C source file ffiillee..mm C transformed file ffiillee..llsstt listing file //uussrr//lliibb//lliibbkkaappiioo..aa Optimized I/O library (for --ssttddiioo option) //uussrr//lliibb6644//mmiippss33//lliibbkkaappiioo..aa (mmiippss11, mmiippss33, and mmiippss44 versions) //uussrr//lliibb6644//mmiippss44//lliibbkkaappiioo..aa SSEEEE AALLSSOO cccc(1), ccpppp(1), mmppcc(1) _I_R_I_S _P_o_w_e_r _C _U_s_e_r'_s _G_u_i_d_e _I_R_I_S _P_o_w_e_r _C _Q_u_i_c_k _R_e_f_e_r_e_n_c_e _P_a_r_a_l_l_e_l _P_r_o_g_r_a_m_m_i_n_g _o_n _S_i_l_i_c_o_n _G_r_a_p_h_i_c_s _C_o_m_p_u_t_e_r _S_y_s_t_e_m_s _P_r_a_c_t_i_c_a_l _P_a_r_a_l_l_e_l _P_r_o_g_r_a_m_m_i_n_g by Dr. Barr Bauer, Academic Press, 1991. _I_n_t_r_o_d_u_c_t_i_o_n _t_o _P_a_r_a_l_l_e_l _P_r_o_g_r_a_m_m_i_n_g by Steven Brawer, Academic Press, 1989. This man page is available only online.