Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

sse2neon -> SIMDe and added non-SIMD version #597

Merged
merged 6 commits into from
Apr 26, 2020
Merged

sse2neon -> SIMDe and added non-SIMD version #597

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
9e6fdd3
Changed sse2neon with SIMDe. Added building non-SIMD version.
mbrcic Apr 20, 2020
c172df7
fix for Neon
mbrcic Apr 20, 2020
f47e8a5
SIMDe made optional. Include paths changes for SIMDe.
mbrcic Apr 20, 2020
3e16e4e
Added documentation entry for added functionality, simde and no_simd.
mbrcic Apr 20, 2020
2b3403f
fix for Neon after test.
mbrcic Apr 21, 2020
66db9da
changed preprocessor conditionals for SIMDe
mbrcic Apr 22, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions .gitmodules
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,3 @@
[submodule "lib/simde"]
path = lib/simde
url = https://github.com/nemequ/simde.git
39 changes: 34 additions & 5 deletions Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,26 +1,40 @@
CFLAGS= -g -Wall -O2 -Wc++-compat #-Wextra
CPPFLAGS= -DHAVE_KALLOC
INCLUDES=
OBJS= kthread.o kalloc.o misc.o bseq.o sketch.o sdust.o options.o index.o chain.o align.o hit.o map.o format.o pe.o esterr.o splitidx.o ksw2_ll_sse.o
OBJS= kthread.o kalloc.o misc.o bseq.o sketch.o sdust.o options.o index.o chain.o align.o hit.o map.o format.o pe.o esterr.o splitidx.o
PROG= minimap2
PROG_EXTRA= sdust minimap2-lite
LIBS= -lm -lz -lpthread


ifeq ($(no_simd),) # if no_simd is not defined
ifeq ($(arm_neon),) # if arm_neon is not defined
OBJS+=ksw2_ll_sse.o
ifeq ($(sse2only),) # if sse2only is not defined
OBJS+=ksw2_extz2_sse41.o ksw2_extd2_sse41.o ksw2_exts2_sse41.o ksw2_extz2_sse2.o ksw2_extd2_sse2.o ksw2_exts2_sse2.o ksw2_dispatch.o
else # if sse2only is defined
OBJS+=ksw2_extz2_sse.o ksw2_extd2_sse.o ksw2_exts2_sse.o
endif
else # if arm_neon is defined
OBJS+=ksw2_extz2_neon.o ksw2_extd2_neon.o ksw2_exts2_neon.o
INCLUDES+=-Isse2neon
OBJS+=ksw2_ll_neon.o ksw2_extz2_neon.o ksw2_extd2_neon.o ksw2_exts2_neon.o
ifeq ($(simde),) # arm_neon without SIMDe -> use sse2neon
INCLUDES+=-Isse2neon
endif
ifeq ($(aarch64),) #if aarch64 is not defined
CFLAGS+=-D_FILE_OFFSET_BITS=64 -mfpu=neon -fsigned-char
else #if aarch64 is defined
CFLAGS+=-D_FILE_OFFSET_BITS=64 -fsigned-char
endif
endif
else
OBJS+=ksw2_ll_nosimd.o ksw2_extz2_nosimd.o ksw2_extd2_nosimd.o ksw2_exts2_nosimd.o
simde=1 # no_simd can be used only with SIMDe
endif

ifneq ($(simde),) # if simde is defined
CFLAGS+=-DSIMDE_ENABLE_NATIVE_ALIASES -DUSE_SIMDE
INCLUDES+=-Ilib/simde
endif

ifneq ($(asan),)
CFLAGS+=-fsanitize=address
Expand Down Expand Up @@ -56,10 +70,8 @@ sdust:sdust.c kalloc.o kalloc.h kdq.h kvec.h kseq.h ketopt.h sdust.h

# SSE-specific targets on x86/x86_64

ifeq ($(arm_neon),) # if arm_neon is defined, compile this target with the default setting (i.e. no -msse2)
ksw2_ll_sse.o:ksw2_ll_sse.c ksw2.h kalloc.h
$(CC) -c $(CFLAGS) -msse2 $(CPPFLAGS) $(INCLUDES) $< -o $@
endif

ksw2_extz2_sse41.o:ksw2_extz2_sse.c ksw2.h kalloc.h
$(CC) -c $(CFLAGS) -msse4.1 $(CPPFLAGS) -DKSW_CPU_DISPATCH $(INCLUDES) $< -o $@
Expand All @@ -84,6 +96,9 @@ ksw2_dispatch.o:ksw2_dispatch.c ksw2.h

# NEON-specific targets on ARM

ksw2_ll_neon.o:ksw2_ll_sse.c ksw2.h kalloc.h
$(CC) -c $(CFLAGS) $(CPPFLAGS) $(INCLUDES) $< -o $@

ksw2_extz2_neon.o:ksw2_extz2_sse.c ksw2.h kalloc.h
$(CC) -c $(CFLAGS) $(CPPFLAGS) -DKSW_SSE2_ONLY -D__SSE2__ $(INCLUDES) $< -o $@

Expand All @@ -93,6 +108,20 @@ ksw2_extd2_neon.o:ksw2_extd2_sse.c ksw2.h kalloc.h
ksw2_exts2_neon.o:ksw2_exts2_sse.c ksw2.h kalloc.h
$(CC) -c $(CFLAGS) $(CPPFLAGS) -DKSW_SSE2_ONLY -D__SSE2__ $(INCLUDES) $< -o $@

# no-SIMD version

ksw2_ll_nosimd.o:ksw2_ll_sse.c ksw2.h kalloc.h
$(CC) -c $(CFLAGS) $(CPPFLAGS) -DSIMDE_NO_NATIVE $(INCLUDES) $< -o $@

ksw2_extz2_nosimd.o:ksw2_extz2_sse.c ksw2.h kalloc.h
$(CC) -c $(CFLAGS) $(CPPFLAGS) -DKSW_SSE2_ONLY -D__SSE2__ -DSIMDE_NO_NATIVE $(INCLUDES) $< -o $@

ksw2_extd2_nosimd.o:ksw2_extd2_sse.c ksw2.h kalloc.h
$(CC) -c $(CFLAGS) $(CPPFLAGS) -DKSW_SSE2_ONLY -D__SSE2__ -DSIMDE_NO_NATIVE $(INCLUDES) $< -o $@

ksw2_exts2_nosimd.o:ksw2_exts2_sse.c ksw2.h kalloc.h
$(CC) -c $(CFLAGS) $(CPPFLAGS) -DKSW_SSE2_ONLY -D__SSE2__ -DSIMDE_NO_NATIVE $(INCLUDES) $< -o $@

# other non-file targets

clean:
Expand Down
4 changes: 4 additions & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -82,6 +82,10 @@ to disable SSE4 code, which will make minimap2 slightly slower.
Minimap2 also works with ARM CPUs supporting the NEON instruction sets. To
compile for 32 bit ARM architectures (such as ARMv7), use `make arm_neon=1`. To compile for for 64 bit ARM architectures (such as ARMv8), use `make arm_neon=1 aarch64=1`.

Minimap2 can use [SIMD Everywhere (SIMDe)](https://github.com/nemequ/simde) library for porting implementation
to the different SIMD instruction sets. To compile using SIMDe, use `make simde=1`. To compile for ARM CPUs, add `simde=1` to the commands given above.
SIMDe also enables non-SIMD implementation using `make no_simd=1`.

### <a name="general"></a>General usage

Without any options, minimap2 takes a reference database and a query sequence
Expand Down
8 changes: 8 additions & 0 deletions ksw2_extd2_sse.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,15 +4,23 @@
#include "ksw2.h"

#ifdef __SSE2__
#ifdef USE_SIMDE
#include <simde/x86/sse2.h>
#else
#include <emmintrin.h>
#endif
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the update!

I think this part could like like this. And other parts are as well.
Because it's natural for me such as "if simde is used, include the simde header, otherwise include the default header.".

But let's wait @lh3 's response.

#ifdef USE_SIMDE
#include <simde/x86/sse2.h>
#else
#include <emmintrin.h>
#endif

Copy link
Contributor Author

@mbrcic mbrcic Apr 22, 2020 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

As there was no other response, I have changed C preprocessor conditionals to the above stated form.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks. It looks good to me.

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

FYI, I personally always use #define SIMDE_ENABLE_NATIVE_ALIASES + #include <simde/x86/sse2.h> as the SIMDe library has fallback that get compiled away to the plain intrinsic calls when on x86 with the regular compilation flag (-msse4.1, etc).


#ifdef KSW_SSE2_ONLY
#undef __SSE4_1__
#endif

#ifdef __SSE4_1__
#ifdef USE_SIMDE
#include <simde/x86/sse4.1.h>
#else
#include <smmintrin.h>
#endif
#endif

#ifdef KSW_CPU_DISPATCH
#ifdef __SSE4_1__
Expand Down
9 changes: 8 additions & 1 deletion ksw2_exts2_sse.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,15 +4,22 @@
#include "ksw2.h"

#ifdef __SSE2__
#ifdef USE_SIMDE
#include <simde/x86/sse2.h>
#else
#include <emmintrin.h>

#endif
#ifdef KSW_SSE2_ONLY
#undef __SSE4_1__
#endif

#ifdef __SSE4_1__
#ifdef USE_SIMDE
#include <simde/x86/sse4.1.h>
#else
#include <smmintrin.h>
#endif
#endif

#ifdef KSW_CPU_DISPATCH
#ifdef __SSE4_1__
Expand Down
8 changes: 8 additions & 0 deletions ksw2_extz2_sse.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,15 +3,23 @@
#include "ksw2.h"

#ifdef __SSE2__
#ifdef USE_SIMDE
#include <simde/x86/sse2.h>
#else
#include <emmintrin.h>
#endif

#ifdef KSW_SSE2_ONLY
#undef __SSE4_1__
#endif

#ifdef __SSE4_1__
#ifdef USE_SIMDE
#include <simde/x86/sse4.1.h>
#else
#include <smmintrin.h>
#endif
#endif

#ifdef KSW_CPU_DISPATCH
#ifdef __SSE4_1__
Expand Down
7 changes: 6 additions & 1 deletion ksw2_ll_sse.c
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,9 +1,14 @@
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <emmintrin.h>
#include "ksw2.h"

#ifdef USE_SIMDE
#include <simde/x86/sse2.h>
#else
#include <emmintrin.h>
#endif

#ifdef __GNUC__
#define LIKELY(x) __builtin_expect((x),1)
#define UNLIKELY(x) __builtin_expect((x),0)
Expand Down
1 change: 1 addition & 0 deletions lib/simde
Open in desktop
Submodule simde added at b30129