It is possible to build a program that calls a non-GEMM OpenBLAS routine from
a static initializer. Since the order of initialization is undefined, and even
less defined when using __attribute__((constructor)) in one TU and a C++ static
initializer in another TU, it can happen (and does, unfortunately) that
gotoblas_init is not called before the first BLAS routine. This results in a
segfault when trying to index into the gotoblas table.
The solution I have here is indirection: rather than directly using the table
use an inlined function to first check if it's been initialized. Since it will
only not have been done once, hopefully the branch prediction still keeps things
fast.
- Compiler TLS is now used only used when the compiler supports it
- If compiler TLS is unsupported, we use platform-specific TLS
- Only one variable (an index) is now in TLS
- We only access TLS once per alloc, and never when freeing
- Allocation / release info is now stored within the allocation itself, by
over-allocating; this saves having external structures do the bookkeeping, and
reduces some of the redundant data that was being stored (such as addresses)
- We never hit the alloc lock when not using SMP or when using OpenMP (that was
my fault)
- Now that there are fewer tracking structures I think this is a bit easier to
read than before
* Allow choosing the OpenMP scheduler and add range hint for GEMM_MULTITHREAD_THRESHOLD
* Amended description of GEMM_MULTITHREAD_THRESHOLD
to reflect #742 making it track floating point operations rather than matrix size
This patch adds the basic infrastructure for adding the SkylakeX (Intel Skylake server)
target. The SkylakeX target will use the AVX512 (AVX512VL level) instruction set,
which brings 2 basic things:
1) 512 bit wide SIMD (2x width of AVX2)
2) 32 SIMD registers (2x the number on AVX2)
This initial patch only contains a trivial transofrmation of the Haswell SGEMM kernel
to AVX512VL; more will follow later but this patch aims to get the infrastructure
in place for this "later".
Full performance tuning has not been done yet; with more registers and wider SIMD
it's in theory possible to retune the kernels but even without that there's an
interesting enough performance increase (30-40% range) with just this change.
For #1416 and #1529, more information about the options OpenBLAS was built with is needed. Additionally we may want to add this data to the openblas.pc file (but not all projects use pkgconfig, and as far as I am aware the cmake module for accessing it does not make such "private" declarations available)
Even if we're directly using the win32 threading driver and not pthreads,
pthread_atfork still works fine to register a pre-fork handler, and is
necessary to restore the threading server to a pre-initialized state.
Some Linux distributions (notably SuSE) have raised CPU_SETSIZE to 4096, apparently disregarding API limitations.
From #1348, the highest value to survive array initialization (on a desktop system) is 3232, and 1024 - which is the
more usual CPU_SETSIZE limit, was demonstrated to work fine on an actual bignuma system.
Failures to obtain or attach shared memory segments would lead to an exit without explanation of the exact cause.
This change introduces a more verbose error message and tries to make the code continue without setting cpu affinity.
Fixes#1351
oon3m0oo
Martin Kroeker
Craig Donner
Arjan van de Ven
Matthew Brett
zhiyong.dang
Alex Arslan
Erik M. Bray
Andrew