I am compiling my spec2017. In my config directory, there is an Example-gcc-linux-x86.cfg file.
I copied it to a new file. I modified the %define label and gcc path.
I also saw a config file used by others. I compared it. I found that its config file is somewhat different from mine.
The main differences are:
There is one more sentence in his config file than mine:
strict_rundir_verify = 0
Its compilation options are different from mine.
his is:
default=base: # flags for all base
OPTIMIZE = -g -O2 -static -march=native -fno-unsafe-math-optimizations -fno-tree-loop-vectorize
mine is:
default=base: # flags for all base
OPTIMIZE = -g -O3 -static -march=native -fno-unsafe-math-optimizations -fno-tree-loop-vectorize
I mainly use spec2017 for prefetching experiments, may I ask
What is the code that he added?
Which compile option is correct here? If different compilation options are used between the papers, the results measured by spec2017 are not comparable? Then why these papers did not disclose the compilation rules of spec2017 they used?
Is it OK to compile spec2017 with the following command? Is it unnecessary to set up copies and threads in addition?
runcpu --config=Example-gcc-linux-x86-test.cfg --reportable --action=run al
Thanks!
Related
This question already has answers here:
How to remove unused C/C++ symbols with GCC and ld?
(11 answers)
Closed 6 years ago.
I need some help for compiling with GCC under MinGW.
Say I have two files:
File a.c contains two functions, a1 and a2
File b.c contains two functions, b1 and b2.
Then I link the two objects into a shared library. The command used are like:
gcc -c a.c
gcc -c b.c
gcc -shared -Wl, --version-script v.ver -Wl, -Map=out.map -Wl, --strip-all -o mydll.dll a.o b.o
File v.ver looks like:
mylib {
global: a1;
a2;
local: *;
}
which is used to control which functions to be exported.
By checking the mapfile I can see that the two functions in b.c are also included into the .text section of the DLL file.
Because this DLL file only exports a1 and a2, and b1 and b2 are only defined in b.c, but never used anywhere. Is there an option I could add in GCC or ld so that b1 and b2 are not built into the DLL file so that I can save some space in the DLL file?
Yes, this is possible. To do this, add the following two flags when compiling your C source code to objects:
-ffunction-sections -fdata-sections
This will generate bigger object files, but will add a lot of information for the linker.
When calling the linker add the following flag:
--gc-sections
The linker will now throw away all functions and sections that are not used. Note that this might incur a performance penalty:
Only use these options when there are significant benefits from doing
so. When you specify these options, the assembler and linker create
larger object and executable files and are also slower. These options
affect code generation. They prevent
optimizations by the compiler and assembler using relative locations inside a translation unit since the locations are unknown
until link time. An example of such an optimization is relaxing calls
to short call instructions.
(man gcc)
See also this question: Query on -ffunction-section & -fdata-sections options of gcc for more information.
What is the right flag or order of flags to disable treating particular warning as error in GCC? I want to do that for -Wimplicit-interface.
>cat test.f90
call s
end
> gfortran -c -Werror -Wimplicit-interface -Wno-error=implicit-interface test.f90 -o test.o
test.f90:1.7:
call s
1
Warning: Procedure 's' called with an implicit interface at (1)
>ls test*
test.f90
No test.o was generated.
Without -Werror it works
> gfortran -c -Wimplicit-interface -Wno-error=implicit-interface test.f90 -o test.o
test.f90:1.7:
call s
1
Warning: Procedure 's' called with an implicit interface at (1)
> ls test*
test.f90 test.o
GCC version is gcc version 4.9.2 20141030 (Cray Inc.) (GCC).
This is not an explicit answer to the question. I found it educative enough and too long to be put as comment.
As you just found, you might not be able to achieve what you want if you combine -Werror and -Wno-error=implicit-interface. Let me explain: as opposed to what we have in the doc, especially the following sentence,
The combined effect of positive and negative forms is that more specific options have priority over less specific ones, independently of their position in the command-line.
It seems that it is not the case in the actual implementation. I had a similar problem recently, and by googling, I found this which contains this sentence:
'-w' permanently sets all warnings off no matter what specific warning is set on
It actually suggests that by using some non specific options, the actual implementation does not allow you to change specific option included in the non specific one.
As #innoSPG points out, the actual behaviour does not conform to what is claimed in the manual.
The comment by #MarkGlisse revealed that this have changed with GCC 5. Therefore it was probably a bug.
The solution is therefore to use the recent version, or to not use one of the -Werror and -Wimplicit-interface.
Or to really provide explicit interfaces everywhere, but that can be problematic, as MPI libraries differ in the amount of explicit interfaces provided in the mpi modules.
Let's say that you have prog.c that includes lib.h, whose functions are defined in lib.c, and you build the program with ̀gcc -O3 lib.c prog.c.
Does GCC merge both source files before compiling them?
Is GCC able to inline short functions of lib.c into the resulting binary?
Summary of answers
This does the trick: gcc -flto -O3 lib.c prog.c.
Both source files are still compiled individually, but the linker is able to inline functions from one file into the other one.
Does GCC merge both source files before compiling them?
No, it doesn't
Is GCC able to inline short functions of lib.c into the resulting binary?
Yes, at advanced optimization level. Look at Whole Program Optimization, Link Time Optimization and similar options
I wrote a Makefile and I can't get it to work. I have an option which is supposed to select which processor to compile to. However, when I run make from the commandline it says:
tandex#tandex-P-6860FX:~/emulators/nintendo sdks/3DS SDK [HomeBrew]$ make
gcc -march=arm7tdmi -static -fexceptions -fnon-call-exceptions -fstack-check test.c -c
test.c:1:0: error: bad value (arm7tdmi) for -march= switch
make: *** [ALL] Error 1
But in the man pages for gcc, it states that arm7tdmi is a permissible value. Am I missing something?
Makefile:
#3DS Compilation Makefile (c) TanDex (TEQ)RunawayFreelancers
#
#Version 0.99 (Alpha) For *nix Devices
#
#Please Check Back Soon for 3rd SDK
#SELECT THE COMPILER TO USE! GCC RECOMMENDED!
#FOR SANITY SAKE, USE C FILES WITH GCC AND CPP FILES WITH G++
CC=gcc
#CC=g++
#OBJECTCOPY REFERENCE, DO NOT REMOVE
OBJC=objcopy
OBJREFS= -O Binary
#SELECT THE PROCESSOR TO TUNE IT TO. ARMV7 (Nintendo DS) or ARMV9(Nintendo DS
(Graphical Support))
#or ARM11 Core ARM1176JZ-S and ARM1176JZF-S (3DS Processor? Not Sure if Correct. Try
and see if they Work?)
#
#NOTE: DS GAMES REQUIRE BOTH A ARM7 AND ARM9 BINARY. RUN THIS TWICE (ONCE FOR EACH)
#
#UNCOMMENT FOR PROCESOR
PROCESSOR=arm7tdmi
#PROCESSOR=arm946e-s
#PROCESSOR=arm1176jz-s
#PROCESSOR=arm1176jzf-s
#FILES
#
#PLACE ALL OF THE FILES HERE, THAT ARE BEING COMPILED!
FILES=test.c
#SET BIN FILE NAME BASED ON PROCESSOR SELECTED
ifeq($(PROCESSOR),arm7tdmi)\
NAME=ARM7.BIN
ifeq($(PROCESSOR), arm946e-s)\
NAME=ARM9.BIN
ifeq($(PROCESSOR), arm1176jz-s)\
NAME=ARM11.BIN
ifeq($(PROCESSOR), arm1176jzf-s)\
NAME=ARM11.BIN
#CREATE OBJECTS
ifeq($(CC), gcc)\
OBJECTS=$(FILES:.c=.o)
ifeq($(CC), g++)\
OBJECTS=$(FILES:.cpp=.o)
#FLAGS! DO NOT CHANGE THESE!!!!!!!!!!! THAT MEANS YOU!!!!!
#
#FOR THOSE WHO WANT TO KNOW WHAT THESE DO, HERE THEY ARE:
#-mtune=$(PROCESSOR) FORE THE COMPILER TO TUNE OUTPUT TO THE SPECIFIED
PROCESSOR
#-static REQUIRED FOR CLEAN BINARY OUTPUT?? (NOT SURE WHAT THIS
DOES, BUT WAS SUGESTED ON A POST ON STACKOVERFLOW)
#-fexceptions FORCE EXCEPTIONS
#-fnon-call-exceptions FORCE EXCEPTIONS TO ONLY BE RETURNED BY THE SYSTEM
(MEMORY AND FPU INSTRUTIONS FOR EXAMPLE)
#-fstack-check FORCE STACK CHECKING (DS / 3DS USE AWKWARD STACK
IMPLEMENTATION)
CFLAGS=-march=$(PROCESSOR) -static -fexceptions -fnon-call-exceptions -fstack-check
ALL:
$(CC) $(CFLAGS) $(FILES) -c
.c.o:
$(OBJC) $(OBJREFS) $(OBJECTS) $(NAME)
.cpp.o:
$(OBJC) $(OBJREFS) $(OBJECTS) $(NAME)
You are probably not calling the right gcc. You seem to be calling the gcc installed in your system, rather than the one that comes with the 3DS SDK.
It appears the problem is with -march=arm7tdmi.
I think the workaround is to avoid using -march=arm7tdmi; and use -march=cpu-type, where cpu-type is one of the ones listed at 3.17.4 ARM Options of the GCC manual.
Here's part of the page:
-march=name
This specifies the name of the target ARM architecture. GCC uses this name to determine what kind of instructions it can emit when
generating assembly code. This option can be used in conjunction with
or instead of the -mcpu= option. Permissible names are: ‘armv2’,
‘armv2a’, ‘armv3’, ‘armv3m’, ‘armv4’, ‘armv4t’, ‘armv5’, ‘armv5t’,
‘armv5e’, ‘armv5te’, ‘armv6’, ‘armv6j’, ‘armv6t2’, ‘armv6z’,
‘armv6kz’, ‘armv6-m’, ‘armv7’, ‘armv7-a’, ‘armv7-r’, ‘armv7-m’,
‘armv7e-m’, ‘armv7ve’, ‘armv8-a’, ‘armv8-a+crc’, ‘iwmmxt’, ‘iwmmxt2’,
‘ep9312’.
I am building a large project with a makefile that was originally built with icpc, and now I need to get it running with g++.
When it compiles the file that uses openmp, it uses the -c flag, and doesn't use any libraries, so it ends up being serial instead of openmp. All of the examples I am seeing aren't using this -c flag.
Is there some way to compile without linking, but using openmp?
edit:
I've been using the -lgomp flag(and the library is on the library path):
g++ -lgomp -c -w -O4 mainS.cpp
g++: -lgomp: linker input file unused because linking not done
Edit: my boss made several mistakes in the code, the makefile, and the documentation. Sorry to have wasted your time, at least it was less than the 5 hours I spend on it =/
Are you passing the flag to enable OpenMP (IIRC it's something like -fopenmp? If you don't chances are the compiler will ignore the OpenMP-related primitives and just produce serial code.
I don't think that -c (ie, compile only, don't like) has anything to do with your problem.
Perhaps the documentation helps...