Hi!
On 2022-01-04T15:12:58+0100, Tobias Burnus <tob...@codesourcery.com> wrote:
> This commit r12-6209 now makes the testcases iterate over all devices
> (including the initial/host device).
>
> Hence, with multiple non-host devices and this test, the error had been
> found before ... ;-)
Yay for test cases! :-)
... but we now run into issues if Intel MIC (emulated) offloading is
(additionally) enabled, because that one still doesn't properly implement
device-side 'omp_get_device_num'. ;-)
Thus pushed to master branch
commit d97364aab1af361275b87713154c366ce2b9029a
"Improve Intel MIC offloading XFAILing for 'omp_get_device_num'", see
attached.
(It wasn't obvious to me how to implement that; very incomplete
"[WIP] Intel MIC 'omp_get_device_num'" attached, not planning on working
on this any further.)
Grüße
Thomas
-----------------
Siemens Electronic Design Automation GmbH; Anschrift: Arnulfstraße 201, 80634
München; Gesellschaft mit beschränkter Haftung; Geschäftsführer: Thomas
Heurung, Frank Thürauf; Sitz der Gesellschaft: München; Registergericht
München, HRB 106955
>From d97364aab1af361275b87713154c366ce2b9029a Mon Sep 17 00:00:00 2001
From: Thomas Schwinge <tho...@codesourcery.com>
Date: Wed, 5 Jan 2022 19:52:25 +0100
Subject: [PATCH] Improve Intel MIC offloading XFAILing for
'omp_get_device_num'
After recent commit be661959a6b6d8f9c3c8608a746789e7b2ec3ca4
"libgomp/testsuite: Improve omp_get_device_num() tests", we're now iterating
over all OpenMP target devices. Intel MIC (emulated) offloading still doesn't
properly implement device-side 'omp_get_device_num', and we thus regress:
PASS: libgomp.c/../libgomp.c-c++-common/target-45.c (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.c/../libgomp.c-c++-common/target-45.c execution test
PASS: libgomp.c++/../libgomp.c-c++-common/target-45.c (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.c++/../libgomp.c-c++-common/target-45.c execution test
PASS: libgomp.fortran/target10.f90 -O0 (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -O0 execution test
PASS: libgomp.fortran/target10.f90 -O1 (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -O1 execution test
PASS: libgomp.fortran/target10.f90 -O2 (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -O2 execution test
PASS: libgomp.fortran/target10.f90 -O3 -fomit-frame-pointer -funroll-loops -fpeel-loops -ftracer -finline-functions (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -O3 -fomit-frame-pointer -funroll-loops -fpeel-loops -ftracer -finline-functions execution test
PASS: libgomp.fortran/target10.f90 -O3 -g (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -O3 -g execution test
PASS: libgomp.fortran/target10.f90 -Os (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -Os execution test
Improve the XFAILing added in commit bb75b22aba254e8ff144db27b1c8b4804bad73bb
"Allow matching Intel MIC in OpenMP 'declare variant'" for the case that *any*
Intel MIC offload device is available.
libgomp/
* testsuite/libgomp.c-c++-common/on_device_arch.h
(any_device_arch, any_device_arch_intel_mic): New.
* testsuite/lib/libgomp.exp
(check_effective_target_offload_device_any_intel_mic): New.
* testsuite/libgomp.c-c++-common/target-45.c: Use it.
* testsuite/libgomp.fortran/target10.f90: Likewise.
---
libgomp/testsuite/lib/libgomp.exp | 12 +++++++++-
.../libgomp.c-c++-common/on_device_arch.h | 23 +++++++++++++++++++
.../libgomp.c-c++-common/target-45.c | 2 +-
.../testsuite/libgomp.fortran/target10.f90 | 2 +-
4 files changed, 36 insertions(+), 3 deletions(-)
diff --git a/libgomp/testsuite/lib/libgomp.exp b/libgomp/testsuite/lib/libgomp.exp
index 57fb6b068f3..8c5ecfff0ac 100644
--- a/libgomp/testsuite/lib/libgomp.exp
+++ b/libgomp/testsuite/lib/libgomp.exp
@@ -451,7 +451,6 @@ proc check_effective_target_openacc_nvidia_accel_selected { } {
# Return 1 if using Intel MIC offload device.
proc check_effective_target_offload_device_intel_mic { } {
return [check_runtime_nocache offload_device_intel_mic {
- #include <omp.h>
#include "testsuite/libgomp.c-c++-common/on_device_arch.h"
int main ()
{
@@ -460,6 +459,17 @@ proc check_effective_target_offload_device_intel_mic { } {
} ]
}
+# Return 1 if any Intel MIC offload device is available.
+proc check_effective_target_offload_device_any_intel_mic { } {
+ return [check_runtime_nocache offload_device_any_intel_mic {
+ #include "testsuite/libgomp.c-c++-common/on_device_arch.h"
+ int main ()
+ {
+ return !any_device_arch_intel_mic ();
+ }
+ } ]
+}
+
# Return 1 if the OpenACC 'host' device type is selected.
proc check_effective_target_openacc_host_selected { } {
diff --git a/libgomp/testsuite/libgomp.c-c++-common/on_device_arch.h b/libgomp/testsuite/libgomp.c-c++-common/on_device_arch.h
index ee541dd2260..f92743b04d7 100644
--- a/libgomp/testsuite/libgomp.c-c++-common/on_device_arch.h
+++ b/libgomp/testsuite/libgomp.c-c++-common/on_device_arch.h
@@ -1,3 +1,4 @@
+#include <omp.h>
#include <gomp-constants.h>
/* static */ int
@@ -41,3 +42,25 @@ on_device_arch_intel_mic ()
{
return on_device_arch (GOMP_DEVICE_INTEL_MIC);
}
+
+static int
+any_device_arch (int d)
+{
+ int nd = omp_get_num_devices ();
+ for (int i = 0; i < nd; ++i)
+ {
+ int d_cur;
+ #pragma omp target device(i) map(from:d_cur)
+ d_cur = device_arch ();
+ if (d_cur == d)
+ return 1;
+ }
+
+ return 0;
+}
+
+int
+any_device_arch_intel_mic ()
+{
+ return any_device_arch (GOMP_DEVICE_INTEL_MIC);
+}
diff --git a/libgomp/testsuite/libgomp.c-c++-common/target-45.c b/libgomp/testsuite/libgomp.c-c++-common/target-45.c
index 837503996d7..27bbeddf7fd 100644
--- a/libgomp/testsuite/libgomp.c-c++-common/target-45.c
+++ b/libgomp/testsuite/libgomp.c-c++-common/target-45.c
@@ -1,4 +1,4 @@
-/* { dg-xfail-run-if TODO { offload_device_intel_mic } } */
+/* { dg-xfail-run-if TODO { offload_device_any_intel_mic } } */
#include <omp.h>
#include <stdlib.h>
diff --git a/libgomp/testsuite/libgomp.fortran/target10.f90 b/libgomp/testsuite/libgomp.fortran/target10.f90
index f6951fc9057..31452554d67 100644
--- a/libgomp/testsuite/libgomp.fortran/target10.f90
+++ b/libgomp/testsuite/libgomp.fortran/target10.f90
@@ -1,5 +1,5 @@
! { dg-do run }
-! { dg-xfail-run-if TODO { offload_device_intel_mic } }
+! { dg-xfail-run-if TODO { offload_device_any_intel_mic } }
program main
use omp_lib
--
2.34.1
>From cea0fd2d10f7a9fd060543e59e142c8c9d06abb0 Mon Sep 17 00:00:00 2001
From: Thomas Schwinge <tho...@codesourcery.com>
Date: Wed, 5 Jan 2022 19:15:06 +0100
Subject: [PATCH] [WIP] Intel MIC 'omp_get_device_num'
See commit 0bac793ed6bad2c0c13cd1e93a1aa5808467afc8 "openmp: Implement omp_get_device_num routine"
P commit fbb592407c9dd244b4cea086cbb90d7bd0bf60bb "libgomp: Fix GOMP_DEVICE_NUM_VAR stringification during offload image load"
P commit be661959a6b6d8f9c3c8608a746789e7b2ec3ca4 "libgomp/testsuite: Improve omp_get_device_num() tests"
With Intel MIC (emulated) offloading enabled:
PASS: libgomp.c/../libgomp.c-c++-common/target-45.c (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.c/../libgomp.c-c++-common/target-45.c execution test
PASS: libgomp.c++/../libgomp.c-c++-common/target-45.c (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.c++/../libgomp.c-c++-common/target-45.c execution test
PASS: libgomp.fortran/target10.f90 -O0 (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -O0 execution test
PASS: libgomp.fortran/target10.f90 -O1 (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -O1 execution test
PASS: libgomp.fortran/target10.f90 -O2 (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -O2 execution test
PASS: libgomp.fortran/target10.f90 -O3 -fomit-frame-pointer -funroll-loops -fpeel-loops -ftracer -finline-functions (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -O3 -fomit-frame-pointer -funroll-loops -fpeel-loops -ftracer -finline-functions execution test
PASS: libgomp.fortran/target10.f90 -O3 -g (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -O3 -g execution test
PASS: libgomp.fortran/target10.f90 -Os (test for excess errors)
[-PASS:-]{+FAIL:+} libgomp.fortran/target10.f90 -Os execution test
TODO Not sure if this is the right approach for Intel MIC, over all.
TODO This is, at least, still missing corresponding changes to 'liboffloadmic/plugin/libgomp-plugin-intelmic.cpp': 'GOMP_OFFLOAD_load_image'/'offload_image'.
TODO I don't understand how 'libgomp/icv-device.c'/'libgomp/config/*/icv-device.c' are split over 'liboffloadmic/plugin/offload_target_main.cpp' vs. 'liboffloadmic/runtime/offload_omp_host.cpp'/'liboffloadmic/runtime/offload_omp_target.cpp'.
---
liboffloadmic/plugin/Makefile.am | 2 +-
liboffloadmic/plugin/Makefile.in | 4 +++-
liboffloadmic/plugin/offload_target_main.cpp | 24 +++++++++++++++++++-
3 files changed, 27 insertions(+), 3 deletions(-)
diff --git a/liboffloadmic/plugin/Makefile.am b/liboffloadmic/plugin/Makefile.am
index 7caea7894ac..840940cb760 100644
--- a/liboffloadmic/plugin/Makefile.am
+++ b/liboffloadmic/plugin/Makefile.am
@@ -57,7 +57,7 @@ if PLUGIN_HOST
else # PLUGIN_TARGET
plugin_includedir = $(libsubincludedir)
plugin_include_HEADERS = main_target_image.h
- AM_CPPFLAGS = $(CPPFLAGS) -DLINUX -DCOI_LIBRARY_VERSION=2 -DOFFLOAD_DEBUG=1 -DSEP_SUPPORT -DTIMING_SUPPORT -DHOST_LIBRARY=0 -I$(coi_inc_dir) -I$(liboffload_src_dir) -I$(libgomp_dir)
+ AM_CPPFLAGS = $(CPPFLAGS) -DLINUX -DCOI_LIBRARY_VERSION=2 -DOFFLOAD_DEBUG=1 -DSEP_SUPPORT -DTIMING_SUPPORT -DHOST_LIBRARY=0 -I$(coi_inc_dir) -I$(liboffload_src_dir) -I$(libgomp_src_dir) -I$(libgomp_dir)
AM_CXXFLAGS = $(CXXFLAGS)
AM_LDFLAGS = -L$(liboffload_dir)/.libs -L$(libgomp_dir)/.libs -loffloadmic_target -lcoi_device -lgomp -rdynamic
endif
diff --git a/liboffloadmic/plugin/Makefile.in b/liboffloadmic/plugin/Makefile.in
index 8d5ad0025c2..45dcd01bab3 100644
--- a/liboffloadmic/plugin/Makefile.in
+++ b/liboffloadmic/plugin/Makefile.in
@@ -401,7 +401,7 @@ target_install_dir = $(accel_search_dir)/lib/gcc/$(accel_target)/$(gcc_version)$
@PLUGIN_HOST_TRUE@libgomp_plugin_intelmic_la_LDFLAGS = -L$(liboffload_dir)/.libs -loffloadmic_host -version-info 1:0:0
@PLUGIN_HOST_FALSE@plugin_includedir = $(libsubincludedir)
@PLUGIN_HOST_FALSE@plugin_include_HEADERS = main_target_image.h
-@PLUGIN_HOST_FALSE@AM_CPPFLAGS = $(CPPFLAGS) -DLINUX -DCOI_LIBRARY_VERSION=2 -DOFFLOAD_DEBUG=1 -DSEP_SUPPORT -DTIMING_SUPPORT -DHOST_LIBRARY=0 -I$(coi_inc_dir) -I$(liboffload_src_dir) -I$(libgomp_dir)
+@PLUGIN_HOST_FALSE@AM_CPPFLAGS = $(CPPFLAGS) -DLINUX -DCOI_LIBRARY_VERSION=2 -DOFFLOAD_DEBUG=1 -DSEP_SUPPORT -DTIMING_SUPPORT -DHOST_LIBRARY=0 -I$(coi_inc_dir) -I$(liboffload_src_dir) -I$(libgomp_src_dir) -I$(libgomp_dir)
@PLUGIN_HOST_FALSE@AM_CXXFLAGS = $(CXXFLAGS)
@PLUGIN_HOST_FALSE@AM_LDFLAGS = -L$(liboffload_dir)/.libs -L$(libgomp_dir)/.libs -loffloadmic_target -lcoi_device -lgomp -rdynamic
@@ -838,3 +838,5 @@ maintainer-clean-local: maintainer-clean-multi
# Tell versions [3.59,3.63) of GNU make to not export all variables.
# Otherwise a system limit (for SysV at least) may be exceeded.
.NOEXPORT:
+
+#TODO Properly regenerate.
diff --git a/liboffloadmic/plugin/offload_target_main.cpp b/liboffloadmic/plugin/offload_target_main.cpp
index ccf7240e286..3b7e06e249c 100644
--- a/liboffloadmic/plugin/offload_target_main.cpp
+++ b/liboffloadmic/plugin/offload_target_main.cpp
@@ -31,6 +31,7 @@
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
+#include "libgomp-plugin.h"
#include "compiler_if_target.h"
@@ -91,7 +92,10 @@ static void *last_var_ptr = NULL;
static int last_var_size = 0;
-/* Override the corresponding functions from libgomp. */
+/* Override functions from 'libgomp/icv-device.c' (see
+ 'libgomp/config/[...]/icv-device.c') as well as the corresponding
+ 'libgomp/fortran.c' wrapper routines. */
+
extern "C" int
omp_is_initial_device (void) __GOMP_NOTHROW
{
@@ -104,6 +108,24 @@ omp_is_initial_device_ (void)
return omp_is_initial_device ();
}
+#ifndef GOMP_DEVICE_NUM_VAR
+# error
+#endif
+
+static volatile int GOMP_DEVICE_NUM_VAR;
+
+extern "C" int
+omp_get_device_num (void) __GOMP_NOTHROW
+{
+ return GOMP_DEVICE_NUM_VAR;
+}
+
+extern "C" int32_t
+omp_get_device_num_ (void)
+{
+ return omp_get_device_num ();
+}
+
/* Dummy function needed for the initialization of target process during the
first call to __offload_offload1. */
--
2.34.1
