Re: gcc and xgetbv

2013-01-25 Thread Vincent Habchi
On 25 janv. 2013, at 18:13, Rainer Müller wrote: > This could still be okay if the actual CPU support is checked at runtime > and a fallback code is provided. Via a CPUID instruction? Could be, but I’m fairly dubious. If it were the case, why check the CPU type at configure time… Anyhow, AFAIK

Re: gcc and xgetbv

2013-01-25 Thread Rainer Müller
On 2013-01-25 17:20, Vincent Habchi wrote: > On 25 janv. 2013, at 16:30, Rainer Müller wrote: > >> This will still become a problem if we are going to distribute this >> software as a pre-compiled binary, as the machine running the buildbot >> might have support for AVX instructions while the end

Re: gcc and xgetbv

2013-01-25 Thread Vincent Habchi
On 25 janv. 2013, at 16:30, Rainer Müller wrote: > This will still become a problem if we are going to distribute this > software as a pre-compiled binary, as the machine running the buildbot > might have support for AVX instructions while the end-user might not > have that. This is already the

Fwd: gcc and xgetbv

2013-01-25 Thread Vincent Habchi
On 25 janv. 2013, at 14:50, Michael Dickens wrote: > Maybe the answer is still found in this thread from a couple years back? > > < > http://lists.macosforge.org/pipermail/macports-dev/2011-October/016335.html >> > > On Oct 3, 2011, at 3:50 PM, vincent habchi wrote: >> Ok, finally I figured o

Re: gcc and xgetbv

2013-01-25 Thread Vincent Habchi
Michael: > Hi Vincent - Thanks for your reply; it seems like that thread from a > while back said pretty much what you're now saying. So, what I'm trying > to do is create a configure-time test which first tries to compile a > "xgetbv" instruction, and if that succeeds then it tries to execute th

Re: gcc and xgetbv

2013-01-25 Thread Rainer Müller
On 2013-01-25 16:13, Michael Dickens wrote: > Hi Vincent - Thanks for your reply; it seems like that thread from a > while back said pretty much what you're now saying. So, what I'm trying > to do is create a configure-time test which first tries to compile a > "xgetbv" instruction, and if that su

Re: gcc and xgetbv

2013-01-25 Thread Michael Dickens
Hi Vincent - Thanks for your reply; it seems like that thread from a while back said pretty much what you're now saying. So, what I'm trying to do is create a configure-time test which first tries to compile a "xgetbv" instruction, and if that succeeds then it tries to execute the binary. If the

Re: gcc and xgetbv

2013-01-25 Thread Joshua Root
On 2013-1-26 01:54 , vincent habchi wrote: > On 25 janv. 2013, at 15:07, Rainer Müller wrote: > >> When compiling the code with gcc-mp-4.7, it calls the assembler >> /opt/local/bin/as (which is the same or even a newer version than >> /usr/bin/as), > > There is no such file as /opt/local/bin/as.

Re: gcc and xgetbv

2013-01-25 Thread vincent habchi
On 25 janv. 2013, at 15:07, Rainer Müller wrote: > To make this more clear, a C compiler does not touch the assembler > instructions in the __asm__() statement. Rather, these are passed down in raw > form to the assembler. Thanks Rainer for elaborating. I was in a suburb train and had to be som

Re: gcc and xgetbv

2013-01-25 Thread Rainer Müller
On 2013-01-25 14:15, vincent habchi wrote: > > On 25 janv. 2013, at 04:02, Michael Dickens wrote: > >> but, this code does not compile using MacPorts' gcc 4.7; it errors out >> with "no such instruction: `xgetbv'". Apple's clang compiles this code >> without the #if, but it errors out when runni

Re: gcc and xgetbv

2013-01-25 Thread Michael Dickens
On Jan 24, 2013, at 10:02 PM, Michael Dickens wrote: > I've been told, and a quick internet search ("xgetbv") agrees, that the > following code snippet should be appropriate to use for GCC >= 4.4: > > {{{ > #if __GNUC__ > 4 || __GNUC__ == 4 && __GNUC_MINOR__ >= 4 > static inline unsigned long lo

Re: gcc and xgetbv

2013-01-25 Thread vincent habchi
On 25 janv. 2013, at 04:02, Michael Dickens wrote: > but, this code does not compile using MacPorts' gcc 4.7; it errors out > with "no such instruction: `xgetbv'". Apple's clang compiles this code > without the #if, but it errors out when running with "Illegal > instruction: 4". Ditto for MacPo

gcc and xgetbv

2013-01-24 Thread Michael Dickens
I'm trying to compile gnuradio using configure.compiler=macports-gcc-4.7 to see if their SIMD detection is working (a component they call VOLK). VOLK was originally designed for GCC-based compilers, and they are working on augmenting for CLANG. I've been told, and a quick internet search ("xgetbv