Details:
- Standard names for reference kernels (levels-1v, -1f and 3) are now
macro constants. Examples:
BLIS_SAXPYV_KERNEL_REF
BLIS_DDOTXF_KERNEL_REF
BLIS_ZGEMM_UKERNEL_REF
- Developers no longer have to name all datatype instances of a kernel
with a common base name; [sdcz] datatype flavors of each kernel or
micro-kernel (level-1v, -1f, or 3) may now be named independently.
This means you can now, if you wish, encode the datatype-specific
register blocksizes in the name of the micro-kernel functions.
- Any datatype instances of any kernel (1v, 1f, or 3) that is left
undefined in bli_kernel.h will default to the corresponding reference
implementation. For example, if BLIS_DGEMM_UKERNEL is left undefined,
it will be defined to be BLIS_DGEMM_UKERNEL_REF.
- Developers no longer need to name level-1v/-1f kernels with multiple
datatype chars to match the number of types the kernel WOULD take in
a mixed type environment, as in bli_dddaxpyv_opt(). Now, one char is
sufficient, as in bli_daxpyv_opt().
- There is no longer a need to define an obj_t wrapper to go along with
your level-1v/-1f kernels. The framework now prvides a _kernel()
function which serves as the obj_t wrapper for whatever kernels are
specified (or defaulted to) via bli_kernel.h
- Developers no longer need to prototype their kernels, and thus no
longer need to include any prototyping headers from within
bli_kernel.h. The framework now generates kernel prototypes, with the
proper type signature, based on the kernel names defined (or defaulted
to) via bli_kernel.h.
- If the complex datatype x (of [cz]) implementation of the gemm micro-
kernel is left undefined by bli_kernel.h, but its same-precision real
domain equivalent IS defined, BLIS will use a 4m-based implementation
for the datatype x implementations of all level-3 operations, using
only the real gemm micro-kernel.
Details:
- Added the ability to induce complex domain level-3 operations via new
virtual complex micro-kernels which are implemented via only real
domain micro-kernels. Two new implementations are provided: 4m and 3m.
4m implements complex matrix multiplication in terms of four real
matrix multiplications, where as 3m uses only three and thus is
capable of even higher (than peak) performance. However, the 3m method
has somewhat weaker numerical properties, making it less desirable
in general.
- Further refined packing routines, which were recently revamped, and
added packing functionality for 4m and 3m.
- Some modifications to trmm and trsm macro-kernels to facilitate indexing
into micro-panels which were packed for 4m/3m virtual kernels.
- Added 4m and 3m interfaces for each level-3 operation.
- Various other minor changes to facilitate 4m/3m methods.
Details:
- Removed a_next and b_next arguments to micro-kernels and replaced them
with a pointer to a new datatype, auxinfo_t, which is simply a struct
that holds a_next and b_next. The struct may hold other auxiliary
information that may be useful to a micro-kernel, such as micro-panel
stride. Micro-kernels may access struct fields via accessor macros
defined in bli_auxinfo_macro_defs.h.
- Updated all instances of micro-kernel definitions, micro-kernel calls,
as well as macro-kernels (for declaring and initializing the structs)
according to above change.
Details:
- Fixed copy-and-paste bug whereby [scz]gemmtrsm_u_opt_d4x4 kernels
for x86_64/core2 were calling the wrong reference code (l instead
of u).
- Fixed some unused variables in x86_64/core2 dotaxpyv and dotxaxpyf
kernels.
- Minor typecasting fix in testsuite/src/test_libblis.c.
- Makefile updates.
Details:
- Added infrastructure to support a new scalar representation, whereby
every object contains an internal scalar that defaults to 1.0. This
facilitates passing scalars around without having to house them in
separate objects. These "attached" scalars are stored in the internal
atom_t field of the obj_t struct, and are always stored to be the same
datatype as the object to which they are attached. Level-3 variants no
longer take scalar arguments, however, level-3 internal back-ends stll
do; this is so that the calling function can perform subproblems such
as C := C - alpha * A * B on-the-fly without needing to change either
of the scalars attached to A or B.
- Removed scalar argument from packm_int().
- Observe and apply attached scalars in scalm_int(), and removed scalar
from interface of scalm_unb_var1().
- Renamed the following functions (and corresponding invocations):
bli_obj_init_scalar_copy_of()
-> bli_obj_scalar_init_detached_copy_of()
bli_obj_init_scalar() -> bli_obj_scalar_init_detached()
bli_obj_create_scalar_with_attached_buffer()
-> bli_obj_create_1x1_with_attached_buffer()
bli_obj_scalar_equals() -> bli_obj_equals()
- Defined new functions:
bli_obj_scalar_detach()
bli_obj_scalar_attach()
bli_obj_scalar_apply_scalar()
bli_obj_scalar_reset()
bli_obj_scalar_has_nonzero_imag()
bli_obj_scalar_equals()
- Placed all bli_obj_scalar_* functions in a new file, bli_obj_scalar.c.
- Renamed the following macros:
bli_obj_scalar_buffer() -> bli_obj_buffer_for_1x1()
bli_obj_is_scalar() -> bli_obj_is_1x1()
- Defined new macros to set and copy internal scalars between objects:
bli_obj_set_internal_scalar()
bli_obj_copy_internal_scalar()
- In level-3 internal back-ends, added conditional blocks where alpha and
beta are checked for non-unit-ness. Those values for alpha and beta are
applied to the scalars attached to aliases of A/B/C, as appropriate,
before being passed into the variant specified by the control tree.
- In level-3 blocked variants, pass BLIS_ONE into subproblems instead of
alpha and/or beta.
- In level-3 macro-kernels, changed how scalars are obtained. Now, scalars
attached to A and B are multiplied together to obtain alpha, while beta
is obtained directly from C.
- In level-3 front-ends, removed old function calls meant to provide
future support for mixed domain/precision. These can be added back later
once that functionality is given proper treatment. Also, removed the
creating of copy-casts of alpha and beta since typecasting of scalars
is now implicitly handled in the internal back-ends when alpha and
beta are applied to the attached scalars.
Details:
- Updated micro-kernels for arm, bgq, loongson3a, and x86_64 so that
unimplemented kernel functions simply call the corresponding reference
implementation. (Previously, these unimplemented functions would
abort() with a "not yet implemented" message.)
Details:
- Added standalone trsm_l/trsm_u micro-kernels for x86_64 (core2).
These kernels are based on the gemmtrsm_l/gemmtrsm_u micro-kernels
that already existed in kernels/x86_64/core2-sse3/3.
Details:
- Removed restrict declaration from b_cast and c_cast from
bli_trsm_lu_ker_var2.c and bli_trsm_rl_ker_var2.c. Curiously, they
are causing problems for xlc only in those two files and no other
macro-kernels.
- Fixed (hopefully) kernel function parameter type declarations in
kernels/bgq/1f/bli_axpyf_opt_var1.c and kernels/bgq/3/bli_gemm_8x8.c.
Details:
- Added kernels for ARM, and configurations for Cortex-A9 and Cortex-A15.
Thanks to Francisco Igual for contributing these kernels and
configurations.
Details:
- Removed support for duplication from the gemmtrsm/trsm micro-kernels
and all framework code.
- Updated test suite modules according to above changes.
Details:
- Applied a patch from Tyler that fixes minor staleness in the piledriver
configuration and gemm micro-kernel.
- Very minor changes to test suite input files.
Details:
- Added a kernel directory for kernels developed by Francisco Igual for
the Sandy Bridge architecture, including a dgemm ukernel coded with
AVX intrinsics.
- Added a configuration for Sandy Bridge using values supplied by Fran.
Details:
- Updated level-1f kernels for x86_64 and bgq to use renamed fusing factor
macros. Meant to include this in 5e54f46c. Thanks to Fran for pointing
this out.
Details:
- Added a 'template' configuration, which contains stub implementations of the
level 1, 1f, and 3 kernels with one datatype implemented in C for each, with
lots of in-file comments and documentation.
- Modified some variable/parameter names for some 1/1f operations. (e.g.
renaming vector length parameter from m to n.)
- Moved level-1f fusing factors from axpyf, dotxf, and dotxaxpyf header files
to bli_kernel.h.
- Modifed test suite to print out fusing factors for axpyf, dotxf, and
dotxaxpyf, as well as the default fusing factor (which are all equal
in the reference and template implementations).
- Cleaned up some sloppiness in the level-1f unb_var1.c files whereby these
reference variants were implemented in terms of front-end routines rather
that directly in terms of the kernels. (For example, axpy2v was implemented
as two calls to axpyv rather than two calls to AXPYV_KERNEL.)
- Changed the interface to dotxf so that it matches that of axpyf, in that
A is assumed to be m x b_n in both cases, and for dotxf A is actually used
as A^T.
- Minor variable naming and comment changes to reference micro-kernels in
frame/3/gemm/ukernels and frame/3/trsm/ukernels.
Details:
- Added various micro-kernels for the following architectures:
Intel MIC
IBM BG/Q
IBM Power7
AMD Piledriver
Loogson 3A
and reorganized kernels directory. Thanks to Tyler Smith, Mike Kistler,
and Xianyi Zhang for contributing these kernels.
- Added configurations corresponding to above architectures, and renamed
"clarksville" configuration to "dunnington".
Details:
- Added a single-precision real gemm micro-kernel in
kernels/x86_64/3/bli_gemm_opt_d4x4.c.
- Adjusted the single-precision real register blocksizes in
config/clarksville/bli_kernel.h to be 8x4.
- Added a missing comment to bli_packm_blk_var2.c that was present in
bli_packm_blk_var3.c
Details:
- Fixed a bug in bli_acquire_mpart_t2b() and bli_acquire_mpart_l2r()
that cause incorrect partitioning when SUBPART0 was requested. This
bug was introduced in 46d3d09d49. Thanks to Bryan for isolating
this bug.
- Removed dupl kernels from kernels/x86_64/3 directory.
- Uncommented beta == 0 optimizaition code in
kernels/x86_64/3/bli_gemm_opt_d4x4.c.
Details:
- Modified x86_64 gemm microkernel so that when beta is zero, C is not read
from memory (nor scaled by beta).
- Fixed minor bug in test suite driver when "Test all combinations of storage
schemes?" switch is disabled, which would result in redundant tests being
executed for matrix-only (e.g. level-1m, level-3) operations if multiple
vector storage schemes were specified.
- Restored debug flags as default in clarksville configuration.
Details:
- Fixed over-cautious handling of unaligned operands in vector instrinsic
implementation of axpyf kernel.
- Fixed over- and under-cautious handling of unaligned operands in vector
intrinsic implementation of dotxaxpyf kernel.
Details:
- Updated level-1/-1f kernels so that non-unit and un-aligned cases are
handled by reference implementation (rather than aborted).
- Added -fomit-frame-pointer to default make_defs.mk for clarksville
configuration.
- Defined bli_offset_from_alignment() macro.
- Minor edits to old test drivers.
Details:
- Tweaked micro-kernels and configuration for clarksville.
- Updated/cleaned up old test drivers in test directory.
- Fixed syntax bug in trsv_unb_var1 and trsv_unf_var1 (introduced
recently).
Details:
- Fixed bugs in trmv_l and trsv_u due to backwards iteration resulting in
unaligned subpartitions. We were already going out of our way a bit to
handle edge cases in the first iteration for blocked variants, and this
was simply the unblocked-fused extension of that idea.
- Fixed control tree handling in her/her2/syr/syr2 that was not taking
into account how the choice of variant needed to be altered for
upper-stored matrices (given that only lower-stored algorithms are
explicitly implemented).
- Added bli_determine_blocksize_dim_f(), bli_determine_blocksize_dim_b()
macros to provide inlined versions of bli_determine_blocksize_[fb]() for
use by unblocked-fused variants.
- Integrated new blocksize_dim macros into gemv/hemv unf variants for
consistency with that of the bugfix for trmv/trsv (both of which now
use the same macros).
- Modified bli_obj_vector_inc() so that 1 is returned if the object is a
vector of length 1 (ie: 1 x 1). This fixes a bug whereby under certain
conditions (e.g. dotv_opt_var1), an invalid increment was returned, which
was invalid only because the code was expecting 1 (for purposes of
performing contiguous vector loads) but got a value greater than 1 because
the column stride of the object (e.g. rho) was inflated for alignment
purposes (albeit unnecessarily since there is only one element in the
object).
- Replaced some old invocations of set0 with set0s.
- Added alpha parameter to gemmtrsm ukernels for x86_64 and use accordingly.
- Fixed increment bug in cleanup loop of gemm ukernel for x86_64.
- Added safeguard to test modules so that testing a problem with a zero
dimension does not result in a failure.
- Tweaked handling of zero dimensions in level-2 and level-3 operations'
internal back-ends to correctly handle cases where output operand still
needs to be scaled (e.g. by beta, in the case of gemm with k = 0).
Details:
- Added two more arguments to the gemm and gemmtrsm microkernels: the
addresses of the next micro-panels of A and B. By passing these
pointers into the micro-kernel, we allow the micro-kernel author to
prefetch micro-panels of A and B as necessary (though this is
completely optional; these addresses may also be safely ignored).
- Updated all seven macro-kernels so that they compute and pass in
a_next and b_next. Note that ONLY the gemm macro-kernel computes
a_next and b_next with the precise semantics we want. I will go back
and fix the other macro-kernels in the near future.
- Added 'restrict' to various micro-kernels from which it was missing.
Details:
- Updated frame/util/dupl/bli_dupl_unb_var1.c to utilize PACKNR and NR
explicitly so navigate b1 so that situations where PACKNR > NR are
supported.
- Moved the 4x2 and 4x4 reference micro-kernels in frame/3/gemm/ukernels and
frame/3/trsm/ukernels to kernels/c99/.
- Updated clarksville and flame configurations.
Details:
- Fixed a bug in reference implementation of dupl (bli_dupl_unb_var1.c),
which resulted in incorrect duplication.
- Updated old test drivers according to recently updated packm control tree
creation interface.
- Added 'restrict' to x86 gemm microkernel interface.
Details:
- Changed all filename and function prefixes from 'bl2' to 'bli'.
- Changed the "blis2.h" header filename to "blis.h" and changed all
corresponding #include statements accordingly.
- Fixed incorrect association for Fran in CREDITS file.
Details:
- Removed #include of "blis2.h" from various lower-level, operation-specific
header files throughout the framework. Given that these low-level headers
are included within #blis2.h in a very specific order, #include'ing blis2.h
within them directly is unnecessary.
Details:
- Renamed x86 and x86_64 kernels to contain a 'd' before the fusing shape
to emphasize that the fusing shape is not for all datatype instances, but
rather just for one (that of double-precision real). Other fusing shapes
would be proportional to their precision and domain "byte footprints".
- Corresponding changes to config/clarksville/bl2_kernel.h.
Details:
- Added/renamed packm, unpackm kernels.
- Added machine value routines.
- Added param_map facility.
- Renamed AUTHORS to CREDITS.
- Added Makefile; continued to expand upon existing configure script.
- #define fuse_fac macros in operation headers if not defined already
(by the user in bl2_kernels.h).
