The code for SGEMM 16x4 and DGEMM 8x4 blocks on z14 and z15 uses
explicit unrolling and interleaving to improve performance. The code
employs an empty inline asm statement with operands that constrain the
compiler's instruction scheduling and thereby enforce proper overlapping
of load and compute phases. Fix an ifdef to apply that for clang builds,
as well.
Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
gcc's default setting for floating-point expression contraction is
"fast", which allows the compiler to emit fused multiply adds instead of
separate multiplies and adds (amongst others). Fused multiply-adds,
which assembly kernels typically apply, also bring a significant
performance advantage to the C implementation for matrix-matrix
multiplication on s390x. To enable that performance advantage for builds
with clang, add -ffp-contract=fast to the compiler options.
Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
... since it is not required and clang does not support that gcc
extension. Instead, use a variable-length array directly for these
operands.
Note that, while the actual inline assembly code does not directly use
these memory operands, they serve to inform the compiler that it cannot
reorder reads or writes to/from the input and output data across the
inline asm statements.
Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
... since clang does not support the instruction format for inline
assembly and also it is not required for current versions of clang.
Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
... as a bandaid for building with clang until LLVM's internal assembler
supports nops without operand.
Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
Some of the kernels written in assembly utilize a "load address"
instruction for loading an immediate value into a register. That is
both unnecessarily complex and LLVM's assembler does not understand that
specific syntax. Thus, replace with the appropriate "load immediate"
instruction, which is also clearer to read.
Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
* Fix misnaming of LAPACK_?ggsvp and ?ggsvd function prototypes as LAPACKE_
* Drop the LAPACKE matrix_layout parameter from the argument lists, change ints to pointers and add missing work arguments.
For the first iteration, it is better to use xvf*ger instead of xvf*gerpp
builtins which helps to avoid setting accumulators to zero. This helps
to reduce few instructions.
In 589c74a the cpuid detection was changed to use systemcfg, but a copy
and paste error was introduced during some refactoring that caused
POWER7 detection to reference CPUTYPE_POWER7 (which doesn't exist)
instead of CPUTYPE_POWER6.
In current code, no matter what number of threads specified, all
available CPU count is used when invoking OMP, which leads to very bad
performance if the workload is small while all available CPUs are big.
Lots of time are wasted on inter-thread sync. Fix this issue by really
using the number specified by the variable 'num' from calling API.
Signed-off-by: Chen, Guobing <guobing.chen@intel.com>
