mirror of
https://github.com/amd/blis.git
synced 2026-05-11 09:39:59 +00:00
Merge pull request #50 from devinamatthews/fix_noopt_avx
Fix configuration issue where instruction set flags are not specified for debug builds.
This commit is contained in:
@@ -76,17 +76,32 @@ GIT_LOG := $(GIT) log --decorate
|
||||
#
|
||||
|
||||
# --- Determine the C compiler and related flags ---
|
||||
ifeq ($(CC),)
|
||||
CC := gcc
|
||||
CC_VENDOR := gcc
|
||||
endif
|
||||
ifneq ($(CC_VENDOR),gcc)
|
||||
$(error gcc is required for this configuration.)
|
||||
endif
|
||||
# Enable IEEE Standard 1003.1-2004 (POSIX.1d).
|
||||
# NOTE: This is needed to enable posix_memalign().
|
||||
CPPROCFLAGS := -D_POSIX_C_SOURCE=200112L
|
||||
CMISCFLAGS := -std=c99 -O3 -mfloat-abi=hard -mfpu=vfpv3 -marm -march=armv7-a #-g
|
||||
CMISCFLAGS := -std=c99 -mfloat-abi=hard
|
||||
CPICFLAGS := -fPIC
|
||||
CDBGFLAGS := #-g
|
||||
CWARNFLAGS := -Wall
|
||||
COPTFLAGS := -marm -march=armv7-a -mfpu=vfpv3 -O3 -mfloat-abi=hard #-g
|
||||
|
||||
ifneq ($(DEBUG_TYPE),off)
|
||||
CDBGFLAGS := -g
|
||||
endif
|
||||
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -O3
|
||||
endif
|
||||
|
||||
CVECFLAGS := -mfpu=vfpv3 -marm -march=armv7-a
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
CVECFLAGS := #-msse3 # -mfpmath=sse
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
# compilation, and one for each of the supported "special" compilation
|
||||
|
||||
@@ -76,18 +76,32 @@ GIT_LOG := $(GIT) log --decorate
|
||||
#
|
||||
|
||||
# --- Determine the C compiler and related flags ---
|
||||
ifeq ($(CC),)
|
||||
CC := gcc
|
||||
|
||||
CC_VENDOR := gcc
|
||||
endif
|
||||
ifneq ($(CC_VENDOR),gcc)
|
||||
$(error gcc is required for this configuration.)
|
||||
endif
|
||||
# Enable IEEE Standard 1003.1-2004 (POSIX.1d).
|
||||
# NOTE: This is needed to enable posix_memalign().
|
||||
CPPROCFLAGS := -D_GNU_SOURCE
|
||||
CMISCFLAGS := -std=c99 -march=armv8-a+fp+simd -ftree-vectorize -O3 -fopenmp -mcpu=cortex-a57.cortex-a53 -mtune=cortex-a57.cortex-a53
|
||||
CMISCFLAGS := -std=c99 -fopenmp
|
||||
CPICFLAGS := -fPIC
|
||||
CDBGFLAGS := -g #-g3 -gdwarf-2
|
||||
CWARNFLAGS := -Wall
|
||||
COPTFLAGS := -march=armv8-a+fp+simd -ftree-vectorize -O3 -mcpu=cortex-a57.cortex-a53 -mtune=cortex-a57.cortex-a53
|
||||
|
||||
ifneq ($(DEBUG_TYPE),off)
|
||||
CDBGFLAGS := -g
|
||||
endif
|
||||
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -O3 -ftree-vectorize -mtune=cortex-a57.cortex-a53
|
||||
endif
|
||||
|
||||
CVECFLAGS := -march=armv8-a+fp+simd -mcpu=cortex-a57.cortex-a53
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
CVECFLAGS := #-march=armv8-a+fp+simd -ftree-vectorize -O3 -mcpu=cortex-a57.cortex-a53 -mtune=cortex-a57.cortex-a53 #-march=armv8-a -O2 -mtune=cortex-a57 -mfpu=neon-fp-armv8
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
# compilation, and one for each of the supported "special" compilation
|
||||
|
||||
@@ -76,17 +76,32 @@ GIT_LOG := $(GIT) log --decorate
|
||||
#
|
||||
|
||||
# --- Determine the C compiler and related flags ---
|
||||
ifeq ($(CC),)
|
||||
CC := gcc
|
||||
CC_VENDOR := gcc
|
||||
endif
|
||||
ifneq ($(CC_VENDOR),gcc)
|
||||
$(error gcc is required for this configuration.)
|
||||
endif
|
||||
# Enable IEEE Standard 1003.1-2004 (POSIX.1d).
|
||||
# NOTE: This is needed to enable posix_memalign().
|
||||
CPPROCFLAGS := -D_POSIX_C_SOURCE=200112L
|
||||
CMISCFLAGS := -std=c99 -fopenmp
|
||||
CPICFLAGS := -fPIC
|
||||
CDBGFLAGS := -g
|
||||
CWARNFLAGS := -Wall
|
||||
COPTFLAGS := -O0 -malign-double -funroll-all-loops
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
ifneq ($(DEBUG_TYPE),off)
|
||||
CDBGFLAGS := -g
|
||||
endif
|
||||
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -O2 -malign-double -funroll-all-loops
|
||||
endif
|
||||
|
||||
CVECFLAGS := -mavx -mfma -march=bdver2 -mfpmath=sse
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
# compilation, and one for each of the supported "special" compilation
|
||||
|
||||
@@ -76,17 +76,32 @@ GIT_LOG := $(GIT) log --decorate
|
||||
#
|
||||
|
||||
# --- Determine the C compiler and related flags ---
|
||||
ifeq ($(CC),)
|
||||
CC := gcc
|
||||
CC_VENDOR := gcc
|
||||
endif
|
||||
ifneq ($(CC_VENDOR),gcc)
|
||||
$(error gcc is required for this configuration.)
|
||||
endif
|
||||
# Enable IEEE Standard 1003.1-2004 (POSIX.1d).
|
||||
# NOTE: This is needed to enable posix_memalign().
|
||||
CPPROCFLAGS := -D_POSIX_C_SOURCE=200112L
|
||||
CMISCFLAGS := -std=c99 -fopenmp
|
||||
CPICFLAGS := -fPIC
|
||||
CDBGFLAGS := #-g
|
||||
CWARNFLAGS := -Wall
|
||||
COPTFLAGS := -O2 -mfpmath=sse -fomit-frame-pointer
|
||||
|
||||
ifneq ($(DEBUG_TYPE),off)
|
||||
CDBGFLAGS := -g
|
||||
endif
|
||||
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -O2 -fomit-frame-pointer
|
||||
endif
|
||||
|
||||
CVECFLAGS := -mavx -mfma -march=native -mfpmath=sse
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
CVECFLAGS := -mavx -mfma -march=native
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
# compilation, and one for each of the supported "special" compilation
|
||||
|
||||
@@ -97,13 +97,10 @@ endif
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -march=armv7-a -mfpu=neon -O2
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS := #-msse3 -march=native # -mfpmath=sse
|
||||
COPTFLAGS := -O2
|
||||
endif
|
||||
|
||||
CVECFLAGS := -march=armv7-a #-msse3 -march=native # -mfpmath=sse
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
@@ -97,13 +97,10 @@ endif
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -march=armv7-a -mfpu=neon -O2 -mfloat-abi=hard
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS := #-msse3 -march=native # -mfpmath=sse
|
||||
COPTFLAGS := -O2
|
||||
endif
|
||||
|
||||
CVECFLAGS := -march=armv7-a #-msse3 -march=native # -mfpmath=sse
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
@@ -97,13 +97,10 @@ endif
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -O2 -mfpmath=sse -fomit-frame-pointer
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS := -msse3 -march=native
|
||||
COPTFLAGS := -O2 -fomit-frame-pointer
|
||||
endif
|
||||
|
||||
CVECFLAGS := -msse3 -march=native -mfpmath=sse
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
@@ -97,13 +97,10 @@ endif
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -O3 -march=native
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS := -mavx2 -mfma -mfpmath=sse #-msse3 -march=native # -mfpmath=sse
|
||||
COPTFLAGS := -O3
|
||||
endif
|
||||
|
||||
CVECFLAGS := -mavx2 -mfma -mfpmath=sse -march=native #-msse3 -march=native # -mfpmath=sse
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
@@ -97,13 +97,10 @@ endif
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -O3 -march=loongson3a -mtune=loongson3a
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS := #-msse3 -march=native # -mfpmath=sse
|
||||
COPTFLAGS := -O3 -mtune=loongson3a
|
||||
endif
|
||||
|
||||
CVECFLAGS := -march=loongson3a #-msse3 -march=native # -mfpmath=sse
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
@@ -100,10 +100,7 @@ else
|
||||
COPTFLAGS := -O3
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS :=
|
||||
endif
|
||||
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
@@ -97,13 +97,10 @@ endif
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -O2 -mfpmath=sse -fomit-frame-pointer
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS := -mavx -mfma -march=native
|
||||
COPTFLAGS := -O2 -fomit-frame-pointer
|
||||
endif
|
||||
|
||||
CVECFLAGS := -mavx -mfma -march=native -mfpmath=sse
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
@@ -100,10 +100,7 @@ else
|
||||
COPTFLAGS := -O3 -mtune=power7
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS := -mvsx
|
||||
endif
|
||||
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
@@ -100,10 +100,7 @@ else
|
||||
COPTFLAGS := -O2
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS := #-msse3 -march=native # -mfpmath=sse
|
||||
endif
|
||||
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
@@ -97,13 +97,10 @@ endif
|
||||
ifeq ($(DEBUG_TYPE),noopt)
|
||||
COPTFLAGS := -O0
|
||||
else
|
||||
COPTFLAGS := -O3 -march=native
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS := -mavx -mfpmath=sse #-msse3 -march=native # -mfpmath=sse
|
||||
COPTFLAGS := -O3
|
||||
endif
|
||||
|
||||
CVECFLAGS := -mavx -mfpmath=sse -march=native #-msse3 -march=native # -mfpmath=sse
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
@@ -100,10 +100,7 @@ else
|
||||
COPTFLAGS := -O2
|
||||
endif
|
||||
|
||||
ifneq ($(DEBUG_TYPE),noopt)
|
||||
CVECFLAGS := #-msse3 -march=native # -mfpmath=sse
|
||||
endif
|
||||
|
||||
CKOPTFLAGS := $(COPTFLAGS)
|
||||
|
||||
# Aggregate all of the flags into multiple groups: one for standard
|
||||
|
||||
3
configure
vendored
3
configure
vendored
@@ -206,6 +206,7 @@ main()
|
||||
;;
|
||||
enable-debug)
|
||||
debug_flag=1
|
||||
debug_type=noopt
|
||||
;;
|
||||
enable-debug=*)
|
||||
debug_flag=1
|
||||
@@ -327,7 +328,7 @@ main()
|
||||
|
||||
# Check if the debug flag was specified.
|
||||
if [ -n "${debug_flag}" ]; then
|
||||
if [ ${debug_type} = 'opt' ]; then
|
||||
if [ "x${debug_type}" = "xopt" ]; then
|
||||
echo "${script_name}: enabling debug symbols with optimizations."
|
||||
else
|
||||
debug_type='noopt'
|
||||
|
||||
Reference in New Issue
Block a user