<p dir="ltr">Can someone see if adding</p>
<p dir="ltr"> Cc-options: -march=i686</p>
<p dir="ltr">To ghc-prim.cabal resolves the issue?</p>
<div class="gmail_quote">On Jul 17, 2014 9:06 PM, "Simon Marlow" <<a href="mailto:marlowsd@gmail.com">marlowsd@gmail.com</a>> wrote:<br type="attribution"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
gcc has -march=native which uses the current CPU's architecture, but I think it would be a really bad idea to turn that on by default, because it would mean that we have to be really careful which machine we build the distributions on.<br>
<br>
On my Linux box, gcc -v says it was configured with --with-arch-32=i686, which means that -march=i686 is the default for 32-bit code. We shouldn't go any later than that IMO.<br>
<br>
Anyway, this is all beside the point, if we aren't able to run the code generated by gcc (in whatever mode) then there's a bug somewhere.<br>
<br>
Cheers,<br>
Simon<br>
<br>
On 17/07/2014 07:39, Johan Tibell wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Alright, then which Make file do we need to fix to make sure GCC is<br>
called correctly? Also, I remember reading that some time during the<br>
4.x GCC series GCC switched to auto-detecting the arch to be that of<br>
the machine being used. Could someone try to just switch GCC to a<br>
newer version and see if it automatically stops trying to use i386,<br>
leading to Simon's problem?<br>
<br>
On Thu, Jul 17, 2014 at 8:37 AM, Niklas Larsson <<a href="mailto:metaniklas@gmail.com" target="_blank">metaniklas@gmail.com</a>> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
It certainly shouldn't be built with i386, because that is generating code<br>
for a processor that lacks all these fancy atomic instructions. The first of<br>
them appears on the 486.<br>
<br>
i686 should be safe, it goes all the way back to Pentium Pro.<br>
<br>
<br>
2014-07-17 8:33 GMT+02:00 Johan Tibell <<a href="mailto:johan.tibell@gmail.com" target="_blank">johan.tibell@gmail.com</a>>:<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
A perhaps silly question, *should* ghc-prim be built with i386 or i686?<br>
<br>
On Thu, Jul 17, 2014 at 8:33 AM, Niklas Larsson <<a href="mailto:metaniklas@gmail.com" target="_blank">metaniklas@gmail.com</a>><br>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
I just found exactly the same thing! Well, I used i686 instead.<br>
<br>
Sounds like it's worthwhile to see if this is limited to ghc-prim or if<br>
there's more stuff that's built with i386.<br>
<br>
<br>
2014-07-17 8:21 GMT+02:00 Páli Gábor János <<a href="mailto:pali.gabor@gmail.com" target="_blank">pali.gabor@gmail.com</a>>:<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
2014-07-17 0:51 GMT+02:00 Páli Gábor János <<a href="mailto:pali.gabor@gmail.com" target="_blank">pali.gabor@gmail.com</a>>:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
2014-07-17 0:47 GMT+02:00 Niklas Larsson <<a href="mailto:metaniklas@gmail.com" target="_blank">metaniklas@gmail.com</a>>:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
I hope they can just be done away with at the source, that is to<br>
make<br>
gcc<br>
generate the assembly primitives. GHC should already be built with<br>
i686, but<br>
does that reach ghc-prim?<br>
</blockquote>
<br>
This depends on GCC -- if no -march=XXX is explicitly set, I guess it<br>
will take its default, which may vary platform by platform.<br>
</blockquote>
<br>
All right, I have finally got a Windows (x64) machine and installed<br>
the msys2 environment by the GHC wiki [1]. This has GCC 4.5.2 (as<br>
Niklas wrote earlier), where the default -march is i386. You should<br>
see this line when trying to compile Johan's test program with the -v<br>
flag set:<br>
<br>
COLLECT_GCC_OPTIONS= ... '-v' '-mtune=i386' '-march=i386'<br>
<br>
With the -march=i586 flag explicitly set in the command line, no<br>
__sync_fetch_and_add_n() calls are generated.<br>
<br>
[1]<br>
<br>
<a href="https://ghc.haskell.org/trac/ghc/wiki/Building/Preparation/Windows/MSYS2" target="_blank">https://ghc.haskell.org/trac/<u></u>ghc/wiki/Building/Preparation/<u></u>Windows/MSYS2</a><br>
</blockquote>
<br>
<br>
</blockquote></blockquote>
<br>
<br>
</blockquote></blockquote>
</blockquote></div>