This is the mail archive of the cygwin mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: I sent in a bug report and never heard anything back

jeff writes:
> You are correct. The gcc people replied to my bug report with:
> "So the problem is your GMP/MPFR are compiled for one type of CPU and cannot be
> copied to other type."
> They closed it as a resolved issue.

Correct, since it's not their bug.

> So the problem is that gcc now requires GMP, and GMP was compiled with options that
> are not compatible with my CPU. I can fix the issue for myself by compiling everything
> from scratch.
> The bigger issue is can cygwin distribute a version of gcc that will work with all modern
> processors, including mine, with it's weird lack of AVX and AVX2. Since I ran into the
> issue, I suspect anyone else with a mobile processor like mine will run into the issue.
> I don't know enough about how cygwin is built and installed to know how different versions
> of the GMP library might be distributed for different Intel processors.

GMP for Cygwin is built as a "fat binary" and choses among different
implementations of their assembly at runtime.  The problem here is that
currently the selection is done per CPU family and not per feature
enabled on that CPU, so when a new CPU comes out that Intel marketing
decides to fuse some features off, there needs to be a fallback
implemented to some older type of CPU that didn't have that feature.

> But the problem is not going to go away, and really should be addressed in order to make
> cygwin run on all reasonable platforms.

I've just released a newer GMP package that might solve that problem,
please report back whether it does.  The fallback in GMP is however
implemented based on the availability of the BMI2 instructions rather
than AVX.  Since in all your mails you still didn't manage to tell us
which exact CPU you're having the problem with I can't tell if that's
going to work in this case, but I suspect that AVX availability should
be coupled with BMI2.

+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Samples for the Waldorf Blofeld:

Problem reports:
Unsubscribe info:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]