amd/blis
1
0
Fork 0
mirror of https://github.com/amd/blis.git synced 2026-04-20 15:48:50 +00:00
Code Issues Packages Projects Releases Wiki Activity

Removed support for 3m, 4m induced methods.

Browse Source 
Details:
- Removed support for all induced methods except for 1m. This included
  removing code related to 3mh, 3m1, 4mh, 4m1a, and 4m1b as well as any
  code that existed only to support those implementations. These
  implementations were rarely used and posed code maintenance challenges
  for BLIS's maintainers going forward.
- Removed reference kernels for packm that pack 3m and 4m micropanels,
  and removed 3m/4m-related code from bli_cntx_ref.c.
- Removed support for 3m/4m from the code in frame/ind, then reorganized
  and streamlined the remaining code in that directory. The *ind(),
  *nat(), and *1m() APIs were all removed. (These additional API layers
  no longer made as much sense with only one induced method (1m) being
  supported.) The bli_ind.c file (and header) were moved to frame/base
  and bli_l3_ind.c (and header) and bli_l3_ind_tapi.h were moved to
  frame/3.
- Removed 3m/4m support from the code in frame/1m/packm.
- Removed 3m/4m support from trmm/trsm macrokernels and simplified some
  pointer arithmetic that was previously expressed in terms of the
  bli_ptr_inc_by_frac() static inline function (whose definition was
  also removed).
- Removed the following subdirectories of level-0 macro headers from
  frame/include/level0: ri3, rih, ri, ro, rpi. The level-0 scalar macros
  defined in these directories were used exclusively for 3m and 4m
  method codes.
- Simplified bli_cntx_set_blkszs() and bli_cntx_set_ind_blkszs() in
  light of 1m being the only induced method left within BLIS.
- Removed dt_on_output field within auxinfo_t and its associated
  accessor functions.
- Re-indexed the 1e/1r pack schemas after removing those associated with
  variants of the 3m and 4m methods. This leaves two bits unused within
  the pack format portion of the schema bitfield. (See bli_type_defs.h
  for more info.)
- Spun off the basic and expert interfaces to the object and typed APIs
  into separate files: bli_l3_oapi.c and bli_l3_oapi_ex.c; bli_l3_tapi.c
  and bli_l3_tapi_ex.c.
- Moved the level-3 operation-specific _check function calls from the
  operations' _front() functions to the corresponding _ex() function of
  the object API. (This change roughly maintains where the _check()
  functions are called in the call stack but lays the groundwork for
  future changes that may come to the level-3 object APIs.) Minor
  modifications to bli_l3_check.c to allow the check() functions to be
  called from the expert interface APIs.
- Removed support within the testsuite for testing the aforementioned
  induced methods, and updated the standalone test drivers in the 'test'
  directory so reflect the retirement of those induced methods.
- Modified the sandbox contract so that the user is obliged to define
  bli_gemm_ex() instead of bli_gemmnat(). (This change was made in light
  of the *nat() functions no longer existing.) Also updated the existing
  'power10' and 'gemmlike' sandboxes to come into compliance with the
  new sandbox rules.
- Updated BLISObjectAPI.md, BLISTypedAPI.md, Testsuite.md documentation
  to reflect the retirement of 3m/4m, and also modified Sandboxes.md to
  bring the document into alignment with new conventions.
- Updated various comments; removed segments of commented-out code.
This commit is contained in:
Field G. Van Zee
2021-10-28 16:05:43 -05:00
parent e8caf200a9
commit f065a8070f
163 changed files with 1455 additions and 17026 deletions
Download Patch File Download Diff File Expand all files Collapse all files

47
test/1m4m/Makefile
View File

@@ -186,20 +186,10 @@ BLA_DEF := -DBLAS
EIG_DEF := -DEIGEN
# Complex implementation type
D3MHW := -DIND=BLIS_3MH
D3M1 := -DIND=BLIS_3M1
D4MHW := -DIND=BLIS_4MH
D4M1B := -DIND=BLIS_4M1B
D4M1A := -DIND=BLIS_4M1A
D1M := -DIND=BLIS_1M
DNAT := -DIND=BLIS_NAT
# Implementation string
#STR_3MHW := -DSTR=\"3mhw\"
#STR_3M1 := -DSTR=\"3m1\"
#STR_4MHW := -DSTR=\"4mhw\"
#STR_4M1B := -DSTR=\"4m1b\"
STR_4M1A := -DSTR=\"4m1a_blis\"
STR_1M := -DSTR=\"1m_blis\"
STR_NAT := -DSTR=\"asm_blis\"
STR_OBL := -DSTR=\"openblas\"
@@ -234,19 +224,18 @@ all-st: blis-st openblas-st mkl-st
all-1s: blis-1s openblas-1s mkl-1s
all-2s: blis-2s openblas-2s mkl-2s
blis-st: blis-nat-st blis-1m-st blis-4m1a-st
blis-1s: blis-nat-1s blis-1m-1s blis-4m1a-1s
blis-2s: blis-nat-2s blis-1m-2s blis-4m1a-2s
blis-st: blis-nat-st blis-1m-st
blis-1s: blis-nat-1s blis-1m-1s
blis-2s: blis-nat-2s blis-1m-2s
#blis-ind: blis-ind-st blis-ind-mt
blis-nat: blis-nat-st blis-nat-1s blis-nat-2s
blis-1m: blis-1m-st blis-1m-1s blis-1m-2s
blis-4m1a: blis-4m1a-st blis-4m1a-1s blis-4m1a-2s
# Define the datatypes, operations, and implementations.
DTS := s d c z
OPS := gemm
BIMPLS := asm_blis 4m1a_blis 1m_blis openblas vendor
BIMPLS := asm_blis 1m_blis openblas vendor
EIMPLS := eigen
# Define functions to construct object filenames from the datatypes and
@@ -265,13 +254,6 @@ BLIS_1M_1S_BINS := $(patsubst %.o,%.x,$(BLIS_1M_1S_OBJS))
BLIS_1M_2S_OBJS := $(call get-2s-objs,1m_blis)
BLIS_1M_2S_BINS := $(patsubst %.o,%.x,$(BLIS_1M_2S_OBJS))
BLIS_4M1A_ST_OBJS := $(call get-st-objs,4m1a_blis)
BLIS_4M1A_ST_BINS := $(patsubst %.o,%.x,$(BLIS_4M1A_ST_OBJS))
BLIS_4M1A_1S_OBJS := $(call get-1s-objs,4m1a_blis)
BLIS_4M1A_1S_BINS := $(patsubst %.o,%.x,$(BLIS_4M1A_1S_OBJS))
BLIS_4M1A_2S_OBJS := $(call get-2s-objs,4m1a_blis)
BLIS_4M1A_2S_BINS := $(patsubst %.o,%.x,$(BLIS_4M1A_2S_OBJS))
BLIS_NAT_ST_OBJS := $(call get-st-objs,asm_blis)
BLIS_NAT_ST_BINS := $(patsubst %.o,%.x,$(BLIS_NAT_ST_OBJS))
BLIS_NAT_1S_OBJS := $(call get-1s-objs,asm_blis)
@@ -309,10 +291,6 @@ blis-1m-st: $(BLIS_1M_ST_BINS)
blis-1m-1s: $(BLIS_1M_1S_BINS)
blis-1m-2s: $(BLIS_1M_2S_BINS)
blis-4m1a-st: $(BLIS_4M1A_ST_BINS)
blis-4m1a-1s: $(BLIS_4M1A_1S_BINS)
blis-4m1a-2s: $(BLIS_4M1A_2S_BINS)
openblas-st: $(OPENBLAS_ST_BINS)
openblas-1s: $(OPENBLAS_1S_BINS)
openblas-2s: $(OPENBLAS_2S_BINS)
@@ -337,7 +315,6 @@ armpl-2s: vendor-2s
# automatically after building the binaries on which they depend.
.INTERMEDIATE: $(BLIS_NAT_ST_OBJS) $(BLIS_NAT_1S_OBJS) $(BLIS_NAT_2S_OBJS)
.INTERMEDIATE: $(BLIS_1M_ST_OBJS) $(BLIS_1M_1S_OBJS) $(BLIS_1M_2S_OBJS)
.INTERMEDIATE: $(BLIS_4M1A_ST_OBJS) $(BLIS_4M1A_1S_OBJS) $(BLIS_4M1A_2S_OBJS)
.INTERMEDIATE: $(OPENBLAS_ST_OBJS) $(OPENBLAS_1S_OBJS) $(OPENBLAS_2S_OBJS)
.INTERMEDIATE: $(EIGEN_ST_OBJS) $(EIGEN_1S_OBJS) $(EIGEN_2S_OBJS)
.INTERMEDIATE: $(VENDOR_ST_OBJS) $(VENDOR_1S_OBJS) $(VENDOR_2S_OBJS)
@@ -358,8 +335,7 @@ get-dt-cpp = $(strip \
get-in-cpp = $(strip \
$(if $(findstring 1m_blis,$(1)),-DIND=BLIS_1M,\
$(if $(findstring 4m1a_blis,$(1)),-DIND=BLIS_4M1A,\
-DIND=BLIS_NAT)))
-DIND=BLIS_NAT))
# A function to return other cpp macros that help the test driver
# identify the implementation.
@@ -371,7 +347,6 @@ get-in-cpp = $(strip \
get-bl-cpp = $(strip \
$(if $(findstring 1m_blis,$(1)),$(STR_1M) $(BLI_DEF),\
$(if $(findstring 4m1a_blis,$(1)),$(STR_4M1A) $(BLI_DEF),\
$(if $(findstring asm_blis,$(1)),$(STR_NAT) $(BLI_DEF),\
$(if $(findstring openblas,$(1)),$(STR_OBL) $(BLA_DEF),\
$(if $(and $(findstring eigen,$(1)),\
@@ -379,7 +354,7 @@ get-bl-cpp = $(strip \
$(STR_EIG) $(EIG_DEF),\
$(if $(findstring eigen,$(1)),\
$(STR_EIG) $(BLA_DEF),\
$(STR_VEN) $(BLA_DEF))))))))
$(STR_VEN) $(BLA_DEF)))))))
# Rules for BLIS and BLAS libraries.
@@ -456,16 +431,6 @@ test_%_$(P2_MAX)_1m_blis_2s.x: test_%_$(P2_MAX)_1m_blis_2s.o $(LIBBLIS_LINK)
$(CC) $(strip $< $(LIBBLIS_LINK) $(LDFLAGS) -o $@)
test_%_$(PS_MAX)_4m1a_blis_st.x: test_%_$(PS_MAX)_4m1a_blis_st.o $(LIBBLIS_LINK)
$(CC) $(strip $< $(LIBBLIS_LINK) $(LDFLAGS) -o $@)
test_%_$(P1_MAX)_4m1a_blis_1s.x: test_%_$(P1_MAX)_4m1a_blis_1s.o $(LIBBLIS_LINK)
$(CC) $(strip $< $(LIBBLIS_LINK) $(LDFLAGS) -o $@)
test_%_$(P2_MAX)_4m1a_blis_2s.x: test_%_$(P2_MAX)_4m1a_blis_2s.o $(LIBBLIS_LINK)
$(CC) $(strip $< $(LIBBLIS_LINK) $(LDFLAGS) -o $@)
test_%_$(PS_MAX)_asm_blis_st.x: test_%_$(PS_MAX)_asm_blis_st.o $(LIBBLIS_LINK)
$(CC) $(strip $< $(LIBBLIS_LINK) $(LDFLAGS) -o $@)

12
test/1m4m/runme.sh
View File

@@ -80,11 +80,10 @@ test_dts="s d c z"
test_ops="gemm"
# Implementations to test.
#test_impls="openblas vendor asm_blis 1m_blis 4m1a_blis"
#test_impls="asm_blis 1m_blis 4m1a_blis"
#test_impls="openblas vendor asm_blis 1m_blis"
#test_impls="asm_blis 1m_blis"
#test_impls="asm_blis"
#test_impls="4m1a_blis"
test_impls="asm_blis 4m1a_blis 1m_blis"
test_impls="asm_blis 1m_blis"
# Save a copy of GOMP_CPU_AFFINITY so that if we have to unset it, we can
# restore the value.
@@ -136,7 +135,7 @@ for th in ${threads}; do
for im in ${test_impls}; do
if [ "${dt}" = "s" -o "${dt}" = "d" ] && \
[ "${im}" = "1m_blis" -o "${im}" = "4m1a_blis" ]; then
[ "${im}" = "1m_blis" ]; then
continue
fi
@@ -164,8 +163,7 @@ for th in ${threads}; do
# Set the threading parameters based on the implementation
# that we are preparing to run.
if [ "${im}" = "asm_blis" ] || \
[ "${im}" = "1m_blis" ] || \
[ "${im}" = "4m1a_blis" ]; then
[ "${im}" = "1m_blis" ]; then
unset OMP_NUM_THREADS
export BLIS_JC_NT=${jc_nt}
export BLIS_PC_NT=${pc_nt}

6
test/1m4m/test_gemm.c
View File

@@ -108,9 +108,6 @@ int main( int argc, char** argv )
ind_t ind_mod = ind;
// A hack to use 3m1 as 1mpb (with 1m as 1mbp).
if ( ind == BLIS_3M1 ) ind_mod = BLIS_1M;
// Initialize a context for the current induced method and datatype.
cntx = bli_gks_query_ind_cntx( ind_mod, dt );
@@ -120,8 +117,7 @@ int main( int argc, char** argv )
#elif 0
#ifdef BLIS
if ( ind == BLIS_4M1A ) k_input = 128;
else if ( ind == BLIS_1M ) k_input = 128;
if ( ind == BLIS_1M ) k_input = 128;
else k_input = 256;
#else
k_input = 192;

10
test/3/Makefile
View File

@@ -187,20 +187,10 @@ BLA_DEF := -DBLAS
EIG_DEF := -DEIGEN
# Complex implementation type
D3MHW := -DIND=BLIS_3MH
D3M1 := -DIND=BLIS_3M1
D4MHW := -DIND=BLIS_4MH
D4M1B := -DIND=BLIS_4M1B
D4M1A := -DIND=BLIS_4M1A
D1M := -DIND=BLIS_1M
DNAT := -DIND=BLIS_NAT
# Implementation string
#STR_3MHW := -DSTR=\"3mhw\"
#STR_3M1 := -DSTR=\"3m1\"
#STR_4MHW := -DSTR=\"4mhw\"
#STR_4M1B := -DSTR=\"4m1b\"
#STR_4M1A := -DSTR=\"4m1a\"
#STR_1M := -DSTR=\"1m\"
STR_NAT := -DSTR=\"asm_blis\"
STR_OBL := -DSTR=\"openblas\"

3
test/3/test_gemm.c
View File

@@ -108,9 +108,6 @@ int main( int argc, char** argv )
ind_t ind_mod = ind;
// A hack to use 3m1 as 1mpb (with 1m as 1mbp).
if ( ind == BLIS_3M1 ) ind_mod = BLIS_1M;
// Initialize a context for the current induced method and datatype.
cntx = bli_gks_query_ind_cntx( ind_mod, dt );

3
test/3/test_hemm.c
View File

@@ -86,9 +86,6 @@ int main( int argc, char** argv )
ind_t ind_mod = ind;
// A hack to use 3m1 as 1mpb (with 1m as 1mbp).
if ( ind == BLIS_3M1 ) ind_mod = BLIS_1M;
// Initialize a context for the current induced method and datatype.
cntx = bli_gks_query_ind_cntx( ind_mod, dt );

3
test/3/test_herk.c
View File

@@ -88,9 +88,6 @@ int main( int argc, char** argv )
ind_t ind_mod = ind;
// A hack to use 3m1 as 1mpb (with 1m as 1mbp).
if ( ind == BLIS_3M1 ) ind_mod = BLIS_1M;
// Initialize a context for the current induced method and datatype.
cntx = bli_gks_query_ind_cntx( ind_mod, dt );

3
test/3/test_trmm.c
View File

@@ -91,9 +91,6 @@ int main( int argc, char** argv )
ind_t ind_mod = ind;
// A hack to use 3m1 as 1mpb (with 1m as 1mbp).
if ( ind == BLIS_3M1 ) ind_mod = BLIS_1M;
// Initialize a context for the current induced method and datatype.
cntx = bli_gks_query_ind_cntx( ind_mod, dt );

3
test/3/test_trsm.c
View File

@@ -91,9 +91,6 @@ int main( int argc, char** argv )
ind_t ind_mod = ind;
// A hack to use 3m1 as 1mpb (with 1m as 1mbp).
if ( ind == BLIS_3M1 ) ind_mod = BLIS_1M;
// Initialize a context for the current induced method and datatype.
cntx = bli_gks_query_ind_cntx( ind_mod, dt );