Details:
- NOTE: This is a merge commit of 'master' of git://github.com/amd/blis
into 'amd-master' of flame/blis.
- Fixed a bug in the downstream value of BLIS_NUM_ARCHS, which was
inadvertantly not incremented when the Zen2 subconfiguration was
added.
- In bli_gemm_front(), added a missing conditional constraint around the
call to bli_gemm_small() that ensures that the computation precision
of C matches the storage precision of C.
- In bli_syrk_front(), reorganized and relocated the notrans/trans logic
that existed around the call to bli_syrk_small() into bli_syrk_small()
to minimize the calling code footprint and also to bring that code
into stylistic harmony with similar code in bli_gemm_front() and
bli_trsm_front(). Also, replaced direct accessing of obj_t fields with
proper accessor static functions (e.g. 'a->dim[0]' becomes
'bli_obj_length( a )').
- Added #ifdef BLIS_ENABLE_SMALL_MATRIX guard around prototypes for
bli_gemm_small(), bli_syrk_small(), and bli_trsm_small(). This is
strictly speaking unnecessary, but it serves as a useful visual cue to
those who may be reading the files.
- Removed cpp macro-protected small matrix debugging code from
bli_trsm_front.c.
- Added a GCC_OT_9_1_0 variable to build/config.mk.in to facilitate gcc
version check for availability of -march=znver2, and added appropriate
support to configure script.
- Cleanups to compiler flags common to recent AMD microarchitectures in
config/zen/amd_config.mk, including: removal of -march=znver1 et al.
from CKVECFLAGS (since the -march flag is added within make_defs.mk);
setting CRVECFLAGS similarly to CKVECFLAGS.
- Cleanups to config/zen/bli_cntx_init_zen.c.
- Cleanups, added comments to config/zen/make_defs.mk.
- Cleanups to config/zen2/make_defs.mk, including making use of newly-
added GCC_OT_9_1_0 and existing GCC_OT_6_1_0 to choose the correct
set of compiler flags based on the version of gcc being used.
- Reverted downstream changes to test/test_gemm.c.
- Various whitespace/comment changes.
Details:
- Rewrote bli_cpuid_query() for ARM architectures to use stdio-based
functions such as fopen() and fgets() instead of popen(). The new code
does more or less the same thing as before--searches /proc/cpuinfo for
various strings, which are then parsed in order to determine the
model, part number, and features. Thanks to Dave Love for suggesting
this change in issue #335.
* Always use sqsumv to compute normfv on MacOS.
* Unconditionally disable the "dot trick" in normfv.
* Added explanatory comment to normfv definition.
Details:
- Added a comment above the unconditional disabling of the dotv-based
implementation to normfv. Thanks to Roman Yurchak, Devin Matthews,
and Isuru Fernando in helping with this improvement.
- CREDITS file update.
Details:
- Added section titled "Reproduction" to both Performance.md and
PerformanceSmall.md that briefly nudges the motivated reader in the
right direction if he/she wishes to run the same performance
benchmarks used to produce the graphs shown in those documents.
Thanks to Dave Love for making this suggestion.
Details:
- Updated test driver source in test, test/3, test/1m4m, and
test/mixeddt to iterate through the problem space backwards. This
can help avoid certain situations where the CPU frequency does not
immediately throttle up to its maximum. Thanks to Robert van de
Geijn for recommending this fix (originally made to test/sup drivers
in 57e422a).
- Applied off-by-one matlab output bugfix from b6017e5 to test drivers
in test, test/3, test/1m4m, and test/mixeddt directories.
Details:
- Fixed an off-by-one bug in the output of matlab row indices in
test/sup/test_gemm.c that only manifested when the problem size
increment was equal to 1.
- Disabled the building of rrc, rcr, rcc, crr, crc, and ccr storage
combinations for blissup drivers in test/sup. This helps make the
building of drivers complete sooner.
- Trivial changes to test/sup/runme.sh.
Details:
- Corrected links in contents section of docs/PerformanceSmall.md,
which were erroneously directing readers to the corresponding
sections of docs/Performance.md.
Details:
- Changed the value of alpha to 1.0 in test/sup/test_gemm.c. This is
needed because libxsmm currently only optimizes gemm operations where
alpha is unit (and beta is unit or zero).
- Adjusted the test/sup/Makefile to test libxsmm with netlib BLAS as its
fallback library. This is the library that will be called the
problem dimensions are deemed too large, or any other criteria for
optimization are not met. (This was done not because it is realistic,
but rather so that it would be very clear when libxsmm ceased handling
gemm calls internally when the data are graphed.)
Details:
- Switch the driver source in test/sup so that libxsmm_?gemm() is called
instead of ?gemm_() when compiling for / linking against libxsmm.
libxsmm's documentation isn't clear on whether it is even *trying* to
provide BLAS API compatibility, and I got tired of trying to figure it
out.
- Added missing -ldl in LDFLAGS when linking against libxsmm.
Details:
- Modified test/sup/Makefile to build drivers that test the performance
of skinny/small problems via libxsmm.
- Modified test/sup/runme.sh to run aforementioned drivers.
- Modified test/sup/test_gemm.c so that problem sizes are tested in
reverse order (from largest to smallest). This can help avoid certain
situations where the CPU frequency does not immediately throttle up
to its maximum. Thanks to Robert van de Geijn for recommending this
fix.
Details:
- Spell out the hack given to me by Sameer Agarwal in order to get Eigen
to build with -march=native (which is critically important for Eigen)
in docs/Performance.md and docs/PerformanceSmall.md.
Details:
- Previously, some versions of gcc would complain that the same
pointer, one_r, is being passed in for both alpha and beta in the
fourth call to the real gemm ukernel in bli_gemmtrsm4m1_ref.c. This
is understandable since the compiler knows that the real gemm ukernel
qualifies all of its floating-point arguments (including alpha and
beta) with restrict. A small hack has been inserted into the file
that defines a new variable to store the value 1.0, which is now used
in lieu of one_r for beta in the fourth call to the real gemm ukernel,
which should pacify the compiler now. Thanks to Dave Love for
reporting this issue (#328) and for Devin Matthews for offering his
'restrict' expertise.
Details:
- Added logic to configure that checks the version of the compiler
against known version ranges that could cause problems later in the
build process. For example, versions of gcc older than 4.9.0 use
different -march labels than version 4.9.0 or later
('-march=corei7-avx' vs '-march=sandybridge', respectively).
Similarly, before 6.1, compilation on Zen was possible, but you
need to start with -march=bdver4 and then disable instruction sets
that were discarded during the transition from Excavator to Zen. So
now, configure substitutes 'yes'/'no' values into anchors in
config.mk.in, which sets various make variables (e.g. GCC_OT_4_9_0),
which can be accessed and branched upon by the various
configurations' make_defs.mk files when setting their compiler flags.
- Updated config/haswell/make_defs.mk to branch on GCC_OT_4_9_0.
- Updated config/sandybridge/make_defs.mk to branch on GCC_OT_4_9_0.
- Updated config/zen/make_defs.mk to branch on GCC_OT_6_1_0.
Details:
- Added the page size, as returned via 'getconf -a | grep PAGE_SIZE',
and the location of the performance drivers to docs/Performance.md
(test/3) and docs/PerformanceSmall.md (test/sup). Thanks to Dave
Love for suggesting these additions in #325.
Details:
- Added a short section after the Intro of the README.md file titled
"Education and Learning" that directs interested readers to the
"LAFF-On Programming for High-Performance" massive open online course
(MOOC) hosted via edX.
Details:
- Added a new standalone test driver directory named '1m4m' that can
build and run performance experiments for BLIS 1m, 4m1a, assembly,
OpenBLAS, and the vendor library (MKL). This new driver directory
was used to regenerate performance results for the 1m paper.
- Added alternate (commented-out) cache blocksizes to
config/haswell/bli_cntx_init_haswell.c. These blocksizes tend to
work well on an a 12-core Intel Xeon E5-2650 v3.
Details:
- Updated the default MC cache blocksizes used by the haswell subconfig
for both row-preferential (the default) and column-preferential
microkernels.
Details:
- Updated the '-march=corei7-avx' flag in the sandybridge subconfig
to '-march=sandybridge' and the '-march=core-avx2' flag in the
haswell subconfig to '-march=haswell'. The older flags were used
by older versions of gcc and should have been updated to the newer
forms a long time ago. (The older flags were clearly working, even
though they are no longer documented in the gcc man page.)
Details:
- Updated most comments in testsuite modules that describe how the
correctness test is performed so that it is clear whether the vector
(normfv) or matrix (normfm) form of Frobenius norm is used.
Details:
- Defined two new static functions in bli_cntx.h:
bli_cntx_set_blksz_def_dt()
bli_cntx_set_blksz_max_dt()
which developers may find convenient when experimenting with different
values of cache blocksizes.
- Updated one- and two-socket multithreaded problem size range and
increment values in test/3/Makefile.
- Changed default to column storage in test/3/test_gemm.c.
- Fixed typo in comment in testsuite/src/test_subm.c.
Details:
- Fixed a bug in bli_l3_thrinfo_print_gemm_paths() and
bli_l3_thrinfo_print_trsm_paths(), defined in bli_l3_thrinfo.c,
whereby subnodes of the thrinfo_t tree are "dereferenced" near the
beginning of the functions, which may lead to segfaults in certain
situations where the thread tree was not fully formed because the
matrix problem was too small for the level of parallelism specified.
(That is, too small because some problems were assigned no work due
to the smallest units in the m and n dimensions being defined by the
register blocksizes mr and nr.) The fix requires several nested levels
of if statements, and this is one of those few instances where use of
goto statements results in (mostly) prettier code, especially in the
case of _gemm_paths(). And while it wasn't necessary, I ported this
goto usage to the loop body that prints the thrinfo_t work_id and
comm_id values for each thread. Thanks to Nicholai Tukanov for helping
to find this bug.
Details:
- Updated the BLASFEO performance graphs shown in PerformanceSmall.md
using a new commit of BLASFEO (2c9f312); updated PerformanceSmall.md
accordingly.
- Updated test/sup/octave/plot_l3sup_perf.m so that the .m files
containing the mpnpkp results do not need to be preprocessed in order
to plot half the problem size range (ie: up to 400 instead of the
800 range of the other shape cases).
- Trivial updates to runme.m.
Details:
- Use -qopenmp-simd instead of -fopenmp-simd when compiling with Intel
icc. Recall that this option is used for SIMD auto-vectorization in
reference kernels only. Support for the -f option has been completely
deprecated and removed in newer versions of icc in favor of -q. Thanks
to Victor Eijkhout for reporting this issue and suggesting the fix.
Details:
- Added a cpp-conditional directive block to bli_arch_config.h that
#includes "bli_family_thunderx2.h". The code has been missing since
adf5c17f. However, this never manifested as an error because the file
is virtually empty and not needed for thunderx2 (or most subconfigs).
Thanks to Jeff Diamond for helping to spot this.
Details:
- Updated the graphs linked in PerformanceSmall.md with BLASFEO results,
and added documenting language accordingly.
- Updated scripts in test/sup/octave to plot BLASFEO data.
- Minor tweak to language re: how OpenBLAS was configured for
docs/Performance.md.
Details:
- Changed starting problem and increment from 16 to 4.
- Added 'lll' (square problems) to list of problem size shapes to
compile and run with.
- Define BLASFEO location and added BLASFEO-related definitions.
