|
Packit |
577717 |
2012-02-13
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/net/linux-net.c: Repairing more coverity warnings.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-02-11
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/windows-common.c: Missed an instance of CPUs yesterday.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: papi_internal.c, threads.c: This changes fixes two race
|
|
Packit |
577717 |
conditions that are probably the cause of the pthrtough
|
|
Packit |
577717 |
double-free error.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
When freeing a thread, we remove and free all eventsets belonging
|
|
Packit |
577717 |
to that thread. This could race with the thread itself removing
|
|
Packit |
577717 |
the evenset, causing some ESI fields to be freed twice.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
The problem was found by using the Valgrind 3.8 Helgrind tool
|
|
Packit |
577717 |
|
|
Packit |
577717 |
valgrind --tool=helgrind --free-is-write=yes ctests/pthrtough
|
|
Packit |
577717 |
|
|
Packit |
577717 |
In order for Helgrind to work, I had to temporarily modify PAPI
|
|
Packit |
577717 |
to use POSIX pthread mutexes for locking. Is there any reason we
|
|
Packit |
577717 |
don't use these all the time?
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-02-10
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/utils/: avail.c, component.c, event_chooser.c,
|
|
Packit |
577717 |
native_avail.c: ix one more case of "CPU's" in the print header
|
|
Packit |
577717 |
code.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Also remove the extraneous The following correspond to fields
|
|
Packit |
577717 |
in the PAPI_event_info_t structure. message
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: testlib/papi_test.h, testlib/test_utils.c,
|
|
Packit |
577717 |
ctests/all_native_events.c, ctests/calibrate.c,
|
|
Packit |
577717 |
ctests/code2name.c, ctests/hwinfo.c: Fix one more case of "CPU's"
|
|
Packit |
577717 |
in the print header code.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Also remove the extraneous The following correspond to fields
|
|
Packit |
577717 |
in the PAPI_event_info_t structure. message
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/buildbot_configure_with_components.sh: take infiniband out of
|
|
Packit |
577717 |
the buildbot test.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: x86_cache_info.c, components/coretemp/linux-coretemp.c,
|
|
Packit |
577717 |
components/lmsensors/linux-lmsensors.c,
|
|
Packit |
577717 |
components/lustre/linux-lustre.c, components/net/linux-net.c,
|
|
Packit |
577717 |
utils/event_chooser.c: Fix coverity errors reported by Will
|
|
Packit |
577717 |
Cohen.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: aix.c, any-proc-null.c, linux-common.c, papi.c, papi.h,
|
|
Packit |
577717 |
papivi.h, solaris-niagara2.c, solaris-ultra.c,
|
|
Packit |
577717 |
ctests/clockres_pthreads.c: Address Redhat bug 785975. The
|
|
Packit |
577717 |
plural of CPU appears to be CPUs
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/Makefile.inc: Patch to cleanup dependencies, allowing for
|
|
Packit |
577717 |
parallel makes. Patch due to Will Cohen from redhat
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-02-09
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/buildbot_configure_with_components.sh: Add infiniband and mx
|
|
Packit |
577717 |
component to buildbot component tests.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/net/tests/: net_values_by_code.c,
|
|
Packit |
577717 |
net_values_by_name.c: Apply patch suggested by Will Cohen to
|
|
Packit |
577717 |
check for system return values.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/lmsensors/linux-lmsensors.h: Added missing string
|
|
Packit |
577717 |
header
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-02-08
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* man/... : update man pages one more time for 4.2.1
|
|
Packit |
577717 |
release
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* release_procedure.txt: Make sure generated html has papi group
|
|
Packit |
577717 |
id.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-02-07
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/multiplex.c: Fix the @file matching multiple files warning.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/README: Cleanup doxygen errors.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* doc/Doxyfile-html: Typo introduced by the last commit.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* doc/Doxyfile-html: Exclude linux-bgp.c from doxygen.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* doc/Doxyfile-html: Make sure the component README file gets
|
|
Packit |
577717 |
included in doxygen.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/coretemp_freebsd/coretemp_freebsd.c: Cleanup
|
|
Packit |
577717 |
doxygen warnings in freebsd coretemp component.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi.h: Cleanup some doxygen warnings related to the
|
|
Packit |
577717 |
groupings.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/example.c: fix doxygen warning in the
|
|
Packit |
577717 |
example component
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* doc/Doxyfile-html: Remove some cruft from doxygen config file.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
This addresses the warning about dot not found at /sw/bin/dot .
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/: infiniband/linux-infiniband.c,
|
|
Packit |
577717 |
infiniband/linux-infiniband.h, cuda/linux-cuda.c,
|
|
Packit |
577717 |
cuda/linux-cuda.h: Cleaned up some doxygen issues
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/lmsensors/linux-lmsensors.c: Removed long
|
|
Packit |
577717 |
forgotten debug outputs
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_libpfm4_events.c: Fix minor doxygen typos.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/vmware/vmware.c: Add params for doxygen
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* man/... : update man pages
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-02-06
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* doc/Doxyfile-man1: Fix a typo in a doxygen config file.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-02-03
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* release_procedure.txt, doc/Doxyfile, doc/Doxyfile-everything,
|
|
Packit |
577717 |
doc/Doxyfile-html, doc/Doxyfile.utils, doc/Doxyfile-man1,
|
|
Packit |
577717 |
doc/Doxyfile-man3, doc/Makefile, doc/doxygen_procedure.txt:
|
|
Packit |
577717 |
Rework the doxygen configuration files.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* RELEASENOTES.txt: Update for the impending release.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* ChangeLogP421.txt, RELEASENOTES.txt: Updates for the impending
|
|
Packit |
577717 |
release.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-02-02
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: papi.c, papi.h: Minor tweaks for doxygen errors
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-02-01
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/lmsensors/: Rules.lmsensors, configure.in: Fixed
|
|
Packit |
577717 |
configure error message and rules link error for shared object
|
|
Packit |
577717 |
linking. Thanks Will Cohen.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/appio/Rules.appio: Correct pathing
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/ctests/api.c: One minor tiny fix to check for PAPI_ENOEVNT
|
|
Packit |
577717 |
when testing PAPI_flops. If PAPI_FP_OPS does not exist on the
|
|
Packit |
577717 |
processor (like many of em), then this tests fails.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-31
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/ctests/multiattach.c: Increase acceptance criteria for
|
|
Packit |
577717 |
cycles.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/Makefile.in, src/configure, src/configure.in, src/papi.h,
|
|
Packit |
577717 |
doc/Doxyfile, doc/Doxyfile-everything, doc/Doxyfile.utils,
|
|
Packit |
577717 |
papi.spec: Update version number to 4.2.1 in preparation for
|
|
Packit |
577717 |
release.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/ctests/prof_utils.c: Correct a warning on 32bit builds about
|
|
Packit |
577717 |
casting caddr_t to (long long)
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Specifically: prof_utils.c:234: warning: cast from pointer to
|
|
Packit |
577717 |
integer of different size prof_utils.c:248: warning: cast from
|
|
Packit |
577717 |
pointer to integer of different size prof_utils.c:262: warning:
|
|
Packit |
577717 |
cast from pointer to integer of different size
|
|
Packit |
577717 |
|
|
Packit |
577717 |
We first cast to unsigned long and then on to long long. ( This
|
|
Packit |
577717 |
maybe overkill, but its for a printf format string )
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-30
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* release_procedure.txt: Add the correct path for doxygen on ICL
|
|
Packit |
577717 |
machines.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_events.csv: Modify Intel Sandybridge PAPI_FP_OPS and
|
|
Packit |
577717 |
PAPI_FP_INS events to not count x87 fp instructions.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
The problem is that the current predefines were made by adding 5
|
|
Packit |
577717 |
events. With the NMI watchdog stealing an event and/or
|
|
Packit |
577717 |
hyperthreading reducing the numbr of available counters by half,
|
|
Packit |
577717 |
we just couldn't fit.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
This now raises the potential for people using x87-compiled
|
|
Packit |
577717 |
floating point on Sandybridge and getting 0 FP_OPS. This is only
|
|
Packit |
577717 |
likely if running a 32-bit kernel and *not* compiling your code
|
|
Packit |
577717 |
with -msse.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
A long-term solution might be trying to find a better set of FP
|
|
Packit |
577717 |
predefines for sandybridge.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/: lustre/linux-lustre.c, mx/linux-mx.c: Some
|
|
Packit |
577717 |
really minor cleanups to the lustre and mx components.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-28
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/: example.c, tests/example_basic.c: Update
|
|
Packit |
577717 |
example component
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Cleans up code, adds some more documentation, adds counter write
|
|
Packit |
577717 |
support.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-27
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_user_events.c: Minor cleanups for user events.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/libpfm4/: README, include/perfmon/pfmlib.h, lib/Makefile,
|
|
Packit |
577717 |
lib/pfmlib_amd64.c, lib/pfmlib_common.c, lib/pfmlib_priv.h: Fix
|
|
Packit |
577717 |
"conflicts" in git import of libpfm4.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/libpfm4/lib/: pfmlib_amd64_fam11h.c,
|
|
Packit |
577717 |
events/amd64_events_fam11h.h: Initial revision
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-26
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_fwrappers.c: Escape the include directives in the
|
|
Packit |
577717 |
documentation.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
(Cleans up doxygen )
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/README: Adding vmware to component README
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/vmware/: Makefile.vmware.in,
|
|
Packit |
577717 |
PAPI-VMwareComponentDocument.pdf, Rules.vmware,
|
|
Packit |
577717 |
VMwareComponentDocument.txt, configure, configure.in, vmware.c,
|
|
Packit |
577717 |
vmware.h: merge vmware branch to head
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/perf_events.c: Set fast_counter_read back to 0 on x86/x86_64
|
|
Packit |
577717 |
perf_events, as currently rdpmc counter access is not supported.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
There are patches floating around that enable this (although
|
|
Packit |
577717 |
performance is still a long way from perfctr) but they will not
|
|
Packit |
577717 |
likely be merged for a while now, and the perf_events substrate
|
|
Packit |
577717 |
will require a lot of extra code to support it once it does make
|
|
Packit |
577717 |
it into a shipping kernel.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/buildbot_configure_with_components.sh: Remove acpi from the
|
|
Packit |
577717 |
buildbot configure script.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-25
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/mx/: Makefile.mx.in, Rules.mx, configure,
|
|
Packit |
577717 |
configure.in, linux-mx.c, linux-mx.h, tests/Makefile,
|
|
Packit |
577717 |
tests/mx_basic.c, tests/mx_elapsed.c, utils/fake_mx_counters.c,
|
|
Packit |
577717 |
utils/sample_output: Re-write of the MX component
|
|
Packit |
577717 |
|
|
Packit |
577717 |
+ Add tests + Modernize code + Remove the need to run ./configure
|
|
Packit |
577717 |
in the mx directory + Add fake mx_counters program that lets you
|
|
Packit |
577717 |
test component on machine without myrinet installed
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/: README, acpi/Rules.acpi,
|
|
Packit |
577717 |
acpi/linux-acpi-memory.c, acpi/linux-acpi.c, acpi/linux-acpi.h:
|
|
Packit |
577717 |
Remove the ACPI component.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
It was one of the oldest components and needed a lot of cleanup
|
|
Packit |
577717 |
work, and it turns out that the main useful event it provided
|
|
Packit |
577717 |
(temperature) isn't available on modern machines/kernels
|
|
Packit |
577717 |
(coretemp should be used instead).
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-23
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/perf_events.c: Restored Phil's changes that I inadvertently
|
|
Packit |
577717 |
clobbered with my last commit :(
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/perf_events.c: Remove a warning about an uninitialized
|
|
Packit |
577717 |
variable.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/utils/: component.c, event_info.c, native_avail.c: Update the
|
|
Packit |
577717 |
Doxygen comments on these utilities to have the command line
|
|
Packit |
577717 |
options listed in a list like the other utils.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/perf_events.c: More improvements to the read path for
|
|
Packit |
577717 |
multiplexed counters. Now the case for bad kernel behavior is
|
|
Packit |
577717 |
built in, and is not required with a #define.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Basically, there are situations when either enabled or running is
|
|
Packit |
577717 |
zero but not both. This could result in a divide by 0 in the
|
|
Packit |
577717 |
worst case, as was observed by Tushar Mohan in papiex. You could
|
|
Packit |
577717 |
trigger it by doing a read immediately after doing a start with
|
|
Packit |
577717 |
perf events and use a FORMAT_SCALE argument.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Now the logic goes, assuming mpxing.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
1) if (running=enabled) return raw counter 2) if (running
|
|
Packit |
577717 |
&& enabled) scale counter by ratio 3) else warn in debug mode
|
|
Packit |
577717 |
return raw counter
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Apparently we need a test case that does a read immediately after
|
|
Packit |
577717 |
a start. That's a hole.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Tested on brutus, core2 2.6.36
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Here's the original report. ------------------- Model string and
|
|
Packit |
577717 |
code : Intel(R) Pentium(R) M processor 1600MHz (9) Linux
|
|
Packit |
577717 |
thinkpad 2.6.38-02063808-generic #201106040910 SMP Sat Jun 4
|
|
Packit |
577717 |
10:51:30 UTC 2011 i686 GNU/Linux PAPI Version: 4.2.0.0
|
|
Packit |
577717 |
|
|
Packit |
577717 |
I think I ran into a bug similar to what we ran with MIPS.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
With the latest PAPI (from CVS), on an x86 (32-bit machine), when
|
|
Packit |
577717 |
using papiex with multiplex with anything more than two events, I
|
|
Packit |
577717 |
get a floating point exception in PAPI during the PAPI_read call.
|
|
Packit |
577717 |
On enabling debugging in the substrate, I think the problem is
|
|
Packit |
577717 |
the same (namely a division by zero, because some event had a
|
|
Packit |
577717 |
zero time of running):
|
|
Packit |
577717 |
|
|
Packit |
577717 |
libpapiex debug: 24625,0x0,papiex_thread_init_routine Starting
|
|
Packit |
577717 |
counters with PAPI_start
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:pe_enable_counters:953:24625
|
|
Packit |
577717 |
ioctl(enable): ctx: 0x96a4bc8, fd: 3
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:pe_enable_counters:953:24625
|
|
Packit |
577717 |
ioctl(enable): ctx: 0x96a4bc8, fd: 5 libpapiex debug:
|
|
Packit |
577717 |
24625,0x0,papiex_thread_init_routine Calling PAPI_lock before
|
|
Packit |
577717 |
critical section libpapiex debug:
|
|
Packit |
577717 |
24625,0x0,papiex_thread_init_routine Released PAPI lock libpapiex
|
|
Packit |
577717 |
debug: 24625,0x0,papiex_start START POINT 0 LABEL libpapiex
|
|
Packit |
577717 |
debug: 24625,0x0,papiex_start Reading counters (PAPI_read) to get
|
|
Packit |
577717 |
initial counts SUBSTRATE:perf_events.c:_papi_pe_read:1147:24625
|
|
Packit |
577717 |
read: fd: 3, tid: 0, cpu: -1, ret: 56
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1148:24625 read: 2 1341021
|
|
Packit |
577717 |
1341021 SUBSTRATE:perf_events.c:_papi_pe_read:1181:24625
|
|
Packit |
577717 |
(papi_pe_buffer[3] 33405 * tot_time_enabled 1341021) /
|
|
Packit |
577717 |
tot_time_running 1341021
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1181:24625
|
|
Packit |
577717 |
(papi_pe_buffer[5] 44552 * tot_time_enabled 1341021) /
|
|
Packit |
577717 |
tot_time_running 1341021
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1147:24625 read: fd: 5,
|
|
Packit |
577717 |
tid: 0, cpu: -1, ret: 40
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1148:24625 read: 1 214777 0
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1181:24625
|
|
Packit |
577717 |
(papi_pe_buffer[3] 0 * tot_time_enabled 214777) /
|
|
Packit |
577717 |
tot_time_running 0
|
|
Packit |
577717 |
|
|
Packit |
577717 |
The above debug log is for three events: PAPI_TOT_CYC,
|
|
Packit |
577717 |
PAPI_TOT_INS and PAPI_L1_DCM. Multiplexing works with two events.
|
|
Packit |
577717 |
Adding the third (any event), gives this error. Basically, the
|
|
Packit |
577717 |
floating point exception kills the program, and PAPI_read never
|
|
Packit |
577717 |
returns.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
I think I know why papiex always hits this bug: It's because
|
|
Packit |
577717 |
right after starting the counters with PAPI_start, papiex does a
|
|
Packit |
577717 |
PAPI_read to store the initial values of the counters in a tmp
|
|
Packit |
577717 |
variable. These are then subtracted from the final counter
|
|
Packit |
577717 |
values. Should we put a deliberate delay? Of course, the real bug
|
|
Packit |
577717 |
should be fixed in PAPI. ----
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/utils/event_info.c: Major re-write of the papi_xml_event_info
|
|
Packit |
577717 |
program. + Remove event code numbers, as they are not stable
|
|
Packit |
577717 |
run-to-run + Add some Doxygen comments + Remove some wrong
|
|
Packit |
577717 |
assumptions that could cause potential buffer overflows + Improve
|
|
Packit |
577717 |
usage information
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-20
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/lustre/: Rules.lustre, linux-lustre.c,
|
|
Packit |
577717 |
linux-lustre.h,
|
|
Packit |
577717 |
fake_proc/fs/lustre/llite/hpcdata-ffff81022a732800/read_ahead_stats,
|
|
Packit |
577717 |
fake_proc/fs/lustre/llite/hpcdata-ffff81022a732800/stats,
|
|
Packit |
577717 |
tests/Makefile, tests/lustre_basic.c: Finish the re-write of the
|
|
Packit |
577717 |
lustre component.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
It would be nice if someone with access to a machine with a
|
|
Packit |
577717 |
lustre filesystem could test this for us.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: papi_internal.c, components/lustre/linux-lustre.c: Update
|
|
Packit |
577717 |
the component initialization code so that it can handle a PAPI
|
|
Packit |
577717 |
ERROR return gracefully. Previously there was no way to indicate
|
|
Packit |
577717 |
initialization failure besides just setting num_native_events to
|
|
Packit |
577717 |
0.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-19
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/lustre/: linux-lustre.c, linux-lustre.h: First
|
|
Packit |
577717 |
pass at cleaning up the lustre component.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
It should now properly report no events when no lustre
|
|
Packit |
577717 |
filesystems are available.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-11
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_events.csv: Add AMD fam12h support to the events file.
|
|
Packit |
577717 |
Right now it is just an alias to the similar fam10h event list;
|
|
Packit |
577717 |
this can be split out if necessary once we find a tester with the
|
|
Packit |
577717 |
hardware.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/libpfm4/: README, docs/man3/pfm_get_event_next.3,
|
|
Packit |
577717 |
docs/man3/pfm_get_pmu_info.3, include/perfmon/perf_event.h,
|
|
Packit |
577717 |
include/perfmon/pfmlib.h, lib/Makefile, lib/pfmlib_amd64.c,
|
|
Packit |
577717 |
lib/pfmlib_amd64_priv.h, lib/pfmlib_common.c,
|
|
Packit |
577717 |
lib/pfmlib_perf_event.c, lib/pfmlib_priv.h,
|
|
Packit |
577717 |
lib/events/intel_coreduo_events.h, lib/events/perf_events.h,
|
|
Packit |
577717 |
perf_examples/Makefile, perf_examples/perf_util.c,
|
|
Packit |
577717 |
perf_examples/perf_util.h, perf_examples/self.c,
|
|
Packit |
577717 |
perf_examples/task_smpl.c, perf_examples/x86/bts_smpl.c: Fix
|
|
Packit |
577717 |
"merge" conflicts with libpfm4 merge.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/libpfm4/lib/: pfmlib_amd64_fam12h.c,
|
|
Packit |
577717 |
events/amd64_events_fam12h.h: Initial revision
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_libpfm4_events.c: Properly use the pfm_get_event_next()
|
|
Packit |
577717 |
iterator to find next event.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Without this, on AMD Fam10h some events are missed.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Some events are still missed due to libpfm4 bug, this will be
|
|
Packit |
577717 |
fixed once I update the libpfm4 tree included with PAPI.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Note, enumeration fixes like this often break things, so please
|
|
Packit |
577717 |
test if possible.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_events.csv: Update the coreduo (not core2) events. Most
|
|
Packit |
577717 |
notably the FP events were wrong.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
This, along with a forthcoming libpfm4 update, make all the
|
|
Packit |
577717 |
CTESTS pass on an old Yonah coreduo laptop I have.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-05
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/ctests/api.c: Make the api test actually test PAPI_flops() as
|
|
Packit |
577717 |
it claims to do, rather than PAPI_flips().
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Patch thanks to: Emilio De Camargo Francesquini
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_hl.c: Fix some copy-and-paste documentation remnants in
|
|
Packit |
577717 |
the papi_hl.c file, mostly where it said FLIPS where it meant
|
|
Packit |
577717 |
FLOPS.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2012-01-04
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/utils/native_avail.c: Update papi_native_avail to *not* print
|
|
Packit |
577717 |
the event codes, as these are not guaranteed to be stable from
|
|
Packit |
577717 |
run to run.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Also fix up the formatting and print some component info too.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Please try and let me know if you don't like the new output.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: configure, configure.in: Respect a FORCED option in
|
|
Packit |
577717 |
configure.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-12-22
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/Rules.pfm4_pe: Remove perfmon.h from MISCHDRS.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-12-20
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: Rules.perfctr, Rules.perfctr-pfm, Rules.pfm, Rules.pfm4_pe,
|
|
Packit |
577717 |
Rules.pfm_pe, linux-lock.h, mb.h: Merry Christmas ARM users.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
This patch fixes the SMP ARM issues reported by Harald Servat.
|
|
Packit |
577717 |
Also, adds proper header dependency checking in the Rules files.
|
|
Packit |
577717 |
People, please when you add headers, please add them to the
|
|
Packit |
577717 |
dependency lines so everything gets rebuilt properly.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
New implementation of SMP locks are very pedantic, that is, they
|
|
Packit |
577717 |
are nost the fastest, but they do use atomics and avoid kernel
|
|
Packit |
577717 |
intervention.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Passed on our 2 core ARM v7. All pthreads tests now pass, except
|
|
Packit |
577717 |
the ones that also fail in the single processor case usually due
|
|
Packit |
577717 |
to a missing event.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Samples:
|
|
Packit |
577717 |
|
|
Packit |
577717 |
mucci@panda:~/papi.head/src$ uname -a Linux panda 3.0.0 #2 SMP
|
|
Packit |
577717 |
Fri Jul 29 16:23:54 EDT 2011 armv7l GNU/Linux
|
|
Packit |
577717 |
|
|
Packit |
577717 |
mucci@panda:~/papi.head/src$ hostname panda
|
|
Packit |
577717 |
|
|
Packit |
577717 |
mucci@panda:~/papi.head/src$ cat /proc/cpuinfo Processor: ARMv7
|
|
Packit |
577717 |
Processor rev 2 (v7l) processor: 0 BogoMIPS: 2007.19
|
|
Packit |
577717 |
|
|
Packit |
577717 |
processor: 1 BogoMIPS: 1965.18
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Features: swp half thumb fastmult vfp edsp thumbee neon vfpv3 CPU
|
|
Packit |
577717 |
implementer: 0x41 CPU architecture: 7 CPU variant: 0x1 CPU part:
|
|
Packit |
577717 |
0xc09 CPU revision: 2
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Hardware: OMAP4 Panda board Revision: 0020 Serial:
|
|
Packit |
577717 |
0000000000000000
|
|
Packit |
577717 |
|
|
Packit |
577717 |
mucci@panda:~/papi.head/src$ ./ctests/locks_pthreads Creating 2
|
|
Packit |
577717 |
threads 10000 iterations took 13489 us. Running 44480 iterations
|
|
Packit |
577717 |
Expected: 88960 Received: 88960 locks_pthreads.c
|
|
Packit |
577717 |
PASSED
|
|
Packit |
577717 |
|
|
Packit |
577717 |
mucci@panda:~/papi.head/src$ ./ctests/pthrtough Creating 2
|
|
Packit |
577717 |
threads for 1000 iterations each of: register create_eventset
|
|
Packit |
577717 |
destroy_eventset unregister pthrtough.c
|
|
Packit |
577717 |
PASSED
|
|
Packit |
577717 |
|
|
Packit |
577717 |
mucci@panda:~/papi.head/src$ ./ctests/pthrtough2 Creating 2000
|
|
Packit |
577717 |
threads for 1 iterations each of: register create_eventset
|
|
Packit |
577717 |
destroy_eventset unregister Failed to create thread: 238
|
|
Packit |
577717 |
Continuing test with 237 threads. pthrtough2.c
|
|
Packit |
577717 |
PASSED
|
|
Packit |
577717 |
|
|
Packit |
577717 |
mucci@panda:~/papi.head/src$ ./ctests/thrspecific Thread
|
|
Packit |
577717 |
0x40ae1470 started, specific data is at 0xbea9c6d4 Thread
|
|
Packit |
577717 |
0x40021000 started, specific data is at 0xbea9c6c4 Thread
|
|
Packit |
577717 |
0x4244d470 started, specific data is at 0xbea9c6c8 Thread
|
|
Packit |
577717 |
0x4138d470 started, specific data is at 0xbea9c6d0 Thread
|
|
Packit |
577717 |
0x41c4d470 started, specific data is at 0xbea9c6cc Entry 0,
|
|
Packit |
577717 |
Thread 0x41c4d470, Data Pointer 0xbea9c6cc, Value 4000000 Entry
|
|
Packit |
577717 |
1, Thread 0x40021000, Data Pointer 0xbea9c6c4, Value 500000 Entry
|
|
Packit |
577717 |
2, Thread 0x40ae1470, Data Pointer 0xbea9c6d4, Value 1000000
|
|
Packit |
577717 |
Entry 3, Thread 0x4244d470, Data Pointer 0xbea9c6c8, Value
|
|
Packit |
577717 |
8000000 Entry 4, Thread 0x4138d470, Data Pointer 0xbea9c6d0,
|
|
Packit |
577717 |
Value 2000000 thrspecific.c PASSED
|
|
Packit |
577717 |
|
|
Packit |
577717 |
mucci@panda:~/papi.head/src$ ./ctests/krentel_pthreads
|
|
Packit |
577717 |
program_time = 6, threshold = 20000000, num_threads = 3
|
|
Packit |
577717 |
|
|
Packit |
577717 |
launched timer in thread 0 launched timer in thread 1 launched
|
|
Packit |
577717 |
timer in thread 3 launched timer in thread 2 [1] time = 1, count
|
|
Packit |
577717 |
= 7, iter = 5, rate = 1400.0/Kiter [2] time = 1, count = 7, iter
|
|
Packit |
577717 |
= 5, rate = 1400.0/Kiter [0] time = 1, count = 7, iter = 5, rate
|
|
Packit |
577717 |
= 1400.0/Kiter [3] time = 1, count = 7, iter = 5, rate =
|
|
Packit |
577717 |
1400.0/Kiter [1] time = 2, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [0] time = 2, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [3] time = 2, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [2] time = 2, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [1] time = 3, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [2] time = 3, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [0] time = 3, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [3] time = 3, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [1] time = 4, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [0] time = 4, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [3] time = 4, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [2] time = 4, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [3] time = 5, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [0] time = 5, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [2] time = 5, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [1] time = 5, count = 26, iter = 17, rate =
|
|
Packit |
577717 |
1529.4/Kiter [2] time = 6, count = 25, iter = 16, rate =
|
|
Packit |
577717 |
1562.5/Kiter [0] time = 6, count = 27, iter = 17, rate =
|
|
Packit |
577717 |
1588.2/Kiter done krentel_pthreads.c PASSED
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-12-15
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_libpfm_presets.c: Change PAPI_PERFMON_EVENT_FILE
|
|
Packit |
577717 |
environment variable name to PAPI_CSV_EVENT_FILE since it's not
|
|
Packit |
577717 |
just for perfmon anymore.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: configure, configure.in: Open mouth, insert foot; fix
|
|
Packit |
577717 |
perfctr configure by not testing a library we have not built yet.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-12-14
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: configure, configure.in: Missed one more place where we
|
|
Packit |
577717 |
tested perfctr != "no"
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: configure, configure.in: Fix a typo in the perfctr section;
|
|
Packit |
577717 |
it was causing a machine to default to perfctr when it had no
|
|
Packit |
577717 |
performance interface. ( a centos vm image with a 2.6.18 kernel
|
|
Packit |
577717 |
)
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Also checks that we actually have perfctr if we specify
|
|
Packit |
577717 |
--with-perfctr.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-12-08
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/cuda/: Makefile.cuda.in, Rules.cuda, configure,
|
|
Packit |
577717 |
configure.in, linux-cuda.c, linux-cuda.h: Added auto-detection of
|
|
Packit |
577717 |
CUDA version to PAPI CUDA Component. Reason is, the interface has
|
|
Packit |
577717 |
changed between CUDA/CUPTI 4.0 and 4.1. PAPI now supports both
|
|
Packit |
577717 |
CUDA versions without any exposure to the users. Configure step
|
|
Packit |
577717 |
is unchanged and no additional knowledge of which CUDA version is
|
|
Packit |
577717 |
installed is required.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-12-03
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/appio/: CHANGES, README, Rules.appio, appio.c,
|
|
Packit |
577717 |
appio.h, tests/Makefile, tests/appio_list_events.c,
|
|
Packit |
577717 |
tests/appio_values_by_code.c, tests/appio_values_by_name.c: [no
|
|
Packit |
577717 |
log message]
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-25
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/linux-timer.c: Fix compilation warning if you specify
|
|
Packit |
577717 |
--with-walltime=gettimeofday
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/linux-timer.c: Fix the build on Linux systems using mmtimer
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/linux-common.c: Update the linux MHz detection code to use
|
|
Packit |
577717 |
bogoMIPS when there is no MHz field available in /proc/cpuinfo.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
This gives roughly correct MHz on ARM, and the MIPS workaround
|
|
Packit |
577717 |
should also still work.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-23
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/net/linux-net.c: Fix compile errors in a debug
|
|
Packit |
577717 |
message. (pathname didn't exist but we are working on
|
|
Packit |
577717 |
NET_PROC_FILE)
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-22
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/net/: linux-net.c, tests/net_values_by_code.c,
|
|
Packit |
577717 |
tests/net_values_by_name.c: Change the ping command in the net
|
|
Packit |
577717 |
tests to not use &> to redirect to NULL.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
This would work on a system with csh, but on systems with a bash
|
|
Packit |
577717 |
shell this runs ping in the background instead, so the test
|
|
Packit |
577717 |
finishes before ping can generate any packets.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/net/linux-net.c: Fix slight bug in the net
|
|
Packit |
577717 |
component, where a memset() had the wrong arguments. This made
|
|
Packit |
577717 |
for weird results in the case where we start/stop quickly enough
|
|
Packit |
577717 |
that we return the initial data.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/net/: CHANGES, Makefile.net.in, README, Rules.net,
|
|
Packit |
577717 |
configure, configure.in, linux-net.c, linux-net.h,
|
|
Packit |
577717 |
tests/Makefile, tests/net_list_events.c,
|
|
Packit |
577717 |
tests/net_values_by_code.c, tests/net_values_by_name.c: Replace
|
|
Packit |
577717 |
net component with updated version written by Jose Pedro
|
|
Packit |
577717 |
Oliveira
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* Dynamically detects the network interfaces
|
|
Packit |
577717 |
(i.e. the ones listed in /proc/net/dev)
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* No longer needs to fork/exec the external ifconfig command and
|
|
Packit |
577717 |
parse its output. It now reads the Linux kernel network
|
|
Packit |
577717 |
statistics directly from /proc/net/dev.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* Each network interface now has 16 events instead of 13
|
|
Packit |
577717 |
(all counters in /proc/net/dev).
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* Adds support for PAPI_event_name_to_code()
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* Adds a couple of small tests/examples
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-16
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* doc/Doxyfile-everything: Fix the exclude libpfm/perfctr config.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-10
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/perf_events.c: Only scale when running != enabled.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Now verified on ig, brutus and the malta
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/perf_events.c: Further tuneups for mpx'ing.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Previous commit broke systems with valid return values from
|
|
Packit |
577717 |
perf_events for running & enabled. My attempt at scaling in long
|
|
Packit |
577717 |
long world caused an overflow which led to a negative number when
|
|
Packit |
577717 |
passed up the chain.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Also consolidated types... best way to avoid this stuff is to
|
|
Packit |
577717 |
start as the type you are ending as.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Now we use some better integer scaling...guaranteed within +-0.5%
|
|
Packit |
577717 |
of the actual scaled value of enabled / running.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
New results on brutus: multiplex1
|
|
Packit |
577717 |
|
|
Packit |
577717 |
case1: Does PAPI_multiplex_init() not break regular operation?
|
|
Packit |
577717 |
Added PAPI_TOT_CYC Added PAPI_FP_INS case1: PAPI_TOT_CYC
|
|
Packit |
577717 |
PAPI_FP_INS case1: 2739865106 600002876
|
|
Packit |
577717 |
|
|
Packit |
577717 |
case2: Does setmpx/add work? Added PAPI_TOT_CYC Added PAPI_FP_INS
|
|
Packit |
577717 |
case2: PAPI_TOT_CYC PAPI_FP_INS case2: 2739678237
|
|
Packit |
577717 |
600002258
|
|
Packit |
577717 |
|
|
Packit |
577717 |
case3: Does add/setmpx work? Added PAPI_TOT_CYC Added PAPI_FP_INS
|
|
Packit |
577717 |
case3: PAPI_TOT_CYC PAPI_FP_INS case3: 2739847832
|
|
Packit |
577717 |
600002298
|
|
Packit |
577717 |
|
|
Packit |
577717 |
case4: Does add/setmpx/add work? Added PAPI_TOT_CYC Added
|
|
Packit |
577717 |
PAPI_FP_INS case4: PAPI_TOT_CYC PAPI_FP_INS case4:
|
|
Packit |
577717 |
2737832980 600013404
|
|
Packit |
577717 |
|
|
Packit |
577717 |
case5: Does setmpx/add/add/start/read work? Added PAPI_TOT_CYC
|
|
Packit |
577717 |
Added PAPI_FP_INS read @start counter[0]: 7106 read @stop
|
|
Packit |
577717 |
counter[0]: 2740387017 difference counter[0]: 2740379911 read
|
|
Packit |
577717 |
@start counter[1]: 0 read @stop counter[1]: 600017169 difference
|
|
Packit |
577717 |
counter[1]: 600017169 multiplex1.c
|
|
Packit |
577717 |
PASSED
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-09
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/cuda/linux-cuda.c: For the CUDA Component,
|
|
Packit |
577717 |
PAPI_read() now accumulates event values. This has to be
|
|
Packit |
577717 |
explicitly done in PAPI because CUPTI automatically resets all
|
|
Packit |
577717 |
counter values to 0 after a read. (PAPI_start()/stop() continues
|
|
Packit |
577717 |
to reset the values to 0)
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/perf_events.c: Last of the multiplex fixes to perf events.
|
|
Packit |
577717 |
The root of all evil was this:
|
|
Packit |
577717 |
|
|
Packit |
577717 |
counts[i] = ( uint64_t )
|
|
Packit |
577717 |
( ( double ) buffer[count_idx] * ( double )
|
|
Packit |
577717 |
|
|
Packit |
577717 |
buffer[get_total_time_enabled_idx( )] /
|
|
Packit |
577717 |
( double )
|
|
Packit |
577717 |
|
|
Packit |
577717 |
buffer[get_total_time_running_idx( )] ) ; In addition to
|
|
Packit |
577717 |
improper casting to uints... (papi returns int64s), using
|
|
Packit |
577717 |
floating point arith is a no-no. Plus this resulted in divide by
|
|
Packit |
577717 |
zeros...
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Before:
|
|
Packit |
577717 |
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1155:12218 read: fd: 3,
|
|
Packit |
577717 |
tid: 0, cpu: -1, buffer[0-2]: 0x6cba, 0x0, 0x0, ret: 24
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1155:12218 read: fd: 4,
|
|
Packit |
577717 |
tid: 0, cpu: -1, buffer[0-2]: 0x23, 0x0, 0x0, ret: 24
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1155:12218 read: fd: 3,
|
|
Packit |
577717 |
tid: 0, cpu: -1, buffer[0-2]: 0x6de72b5d, 0x8ae0fa80, 0x8ae0fa80,
|
|
Packit |
577717 |
ret: 24 SUBSTRATE:perf_events.c:_papi_pe_read:1155:12218 read:
|
|
Packit |
577717 |
fd: 4, tid: 0, cpu: -1, buffer[0-2]: 0x4c4b46b, 0x8ae0fa80,
|
|
Packit |
577717 |
0x8ae0fa80, ret: 24
|
|
Packit |
577717 |
|
|
Packit |
577717 |
So kernel is good, but errors in multiplexed scaling.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
case5: Does setmpx/add/add/start/read work? Added PAPI_TOT_CYC
|
|
Packit |
577717 |
Added PAPI_FP_INS read @start counter[0]: 9223372034707292159
|
|
Packit |
577717 |
read @stop counter[0]: 1843791732 difference counter[0]:
|
|
Packit |
577717 |
-9223372032863500427 multiplex1.c
|
|
Packit |
577717 |
FAILED Line # 389
|
|
Packit |
577717 |
|
|
Packit |
577717 |
With fix:
|
|
Packit |
577717 |
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1151:12821 read: fd: 3,
|
|
Packit |
577717 |
tid: 0, cpu: -1, buffer[0-2]: 0x6782, 0x0, 0x0, ret: 24
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1151:12821 read: fd: 4,
|
|
Packit |
577717 |
tid: 0, cpu: -1, buffer[0-2]: 0x0, 0x0, 0x0, ret: 24
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_read:1151:12821 read: fd: 3,
|
|
Packit |
577717 |
tid: 0, cpu: -1, buffer[0-2]: 0x6de725dc, 0x8ae0fa80, 0x8ae0fa80,
|
|
Packit |
577717 |
ret: 24 SUBSTRATE:perf_events.c:_papi_pe_read:1151:12821 read:
|
|
Packit |
577717 |
fd: 4, tid: 0, cpu: -1, buffer[0-2]: 0x4c4b400, 0x8ae0fa80,
|
|
Packit |
577717 |
0x8ae0fa80, ret: 24 read @start counter[0]: 26498 read @stop
|
|
Packit |
577717 |
counter[0]: 1843865052 difference counter[0]: 1843838554 read
|
|
Packit |
577717 |
@start counter[1]: 0 read @stop counter[1]: 80000000 difference
|
|
Packit |
577717 |
counter[1]: 80000000
|
|
Packit |
577717 |
SUBSTRATE:perf_events.c:_papi_pe_update_control_state:1288:12821
|
|
Packit |
577717 |
Called with count == 0
|
|
Packit |
577717 |
SUBSTRATE:papi_libpfm4_events.c:_papi_libpfm_shutdown:1178:12821
|
|
Packit |
577717 |
shutdown multiplex1.c PASSED
|
|
Packit |
577717 |
|
|
Packit |
577717 |
New code is vastly simpler and smaller and checks for bad kernel
|
|
Packit |
577717 |
behavior:
|
|
Packit |
577717 |
|
|
Packit |
577717 |
int64_t tot_time_running =
|
|
Packit |
577717 |
papi_pe_buffer[get_total_time_running_idx( )];
|
|
Packit |
577717 |
int64_t tot_time_enabled =
|
|
Packit |
577717 |
papi_pe_buffer[get_total_time_enabled_idx( )];
|
|
Packit |
577717 |
#ifdef BRAINDEAD_MULTIPLEXING if (tot_time_enabled == 0)
|
|
Packit |
577717 |
tot_time_enabled = 1; if (tot_time_running == 0)
|
|
Packit |
577717 |
tot_time_running = 1; #else /* If we are convinced this
|
|
Packit |
577717 |
platform's kernel is fully operational, then this stuff will
|
|
Packit |
577717 |
never happen. If it does, then BRAINDEAD_MULTIPLEXING
|
|
Packit |
577717 |
needs to be enabled. */ if ((tot_time_running == 0) &&
|
|
Packit |
577717 |
(papi_pe_buffer[count_idx])) { PAPIERROR("This platform
|
|
Packit |
577717 |
has a kernel bug in multiplexing, count is %lld (not 0), but time
|
|
Packit |
577717 |
running is 0.\n",papi_pe_buffer[count_idx]); return
|
|
Packit |
577717 |
PAPI_EBUG; } if ((tot_time_enabled == 0) &&
|
|
Packit |
577717 |
(papi_pe_buffer[count_idx])) { PAPIERROR("This platform
|
|
Packit |
577717 |
has a kernel bug in multiplexing, count is %lld (not 0), but time
|
|
Packit |
577717 |
enabled is 0.\n",papi_pe_buffer[count_idx]); return PAPI_EBUG;
|
|
Packit |
577717 |
} #endif pe_ctl->counts[i] =
|
|
Packit |
577717 |
(papi_pe_buffer[count_idx] * tot_time_enabled) /
|
|
Packit |
577717 |
tot_time_running;
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Also, renamed all instances of 'buffer' to papi_pe_buffer because
|
|
Packit |
577717 |
buffer is a global variable on MIPS/Linux/libc. Yikes! (gdb)
|
|
Packit |
577717 |
whatis buffer type = struct utmp *
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/ctests/multiplex1.c: Made sure that PAPI_TOT_CYC is the first
|
|
Packit |
577717 |
event added to multiplexing event set.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
This will demonstrate the bug in perf_event multiplexing
|
|
Packit |
577717 |
arithmetic in case5 on MIPS and other perf_event subsystems that
|
|
Packit |
577717 |
likely have some breakage in the kernels handling of
|
|
Packit |
577717 |
multiplexing. The common bug is that the perf_event subsystem
|
|
Packit |
577717 |
does not fill in the second and third elements of the 24 byte
|
|
Packit |
577717 |
read that gets returned from the kernel. These values are
|
|
Packit |
577717 |
time_enabled and time_running. MIPS as of 3.0.3 just fills this
|
|
Packit |
577717 |
in after a HZ tick has happened. Workarounds are pretty simple in
|
|
Packit |
577717 |
the low level layer...
|
|
Packit |
577717 |
|
|
Packit |
577717 |
A buggy output looks like this (3.0.3 MIPS/Linux Big Endian)
|
|
Packit |
577717 |
|
|
Packit |
577717 |
-bash-4.1$ ./ctests/multiplex1 case1: Does PAPI_multiplex_init()
|
|
Packit |
577717 |
not break regular operation? Added PAPI_TOT_CYC Added PAPI_FP_INS
|
|
Packit |
577717 |
case1: PAPI_TOT_CYC PAPI_FP_INS case1: 1843775252
|
|
Packit |
577717 |
80000000
|
|
Packit |
577717 |
|
|
Packit |
577717 |
case2: Does setmpx/add work? Added PAPI_TOT_CYC Added PAPI_FP_INS
|
|
Packit |
577717 |
case2: PAPI_TOT_CYC PAPI_FP_INS case2: 1843773254
|
|
Packit |
577717 |
80000037
|
|
Packit |
577717 |
|
|
Packit |
577717 |
case3: Does add/setmpx work? Added PAPI_TOT_CYC Added PAPI_FP_INS
|
|
Packit |
577717 |
case3: PAPI_TOT_CYC PAPI_FP_INS case3: 1843772919
|
|
Packit |
577717 |
80000037
|
|
Packit |
577717 |
|
|
Packit |
577717 |
case4: Does add/setmpx/add work? Added PAPI_TOT_CYC Added
|
|
Packit |
577717 |
PAPI_FP_INS case4: PAPI_TOT_CYC PAPI_FP_INS case4:
|
|
Packit |
577717 |
1843773959 80000037
|
|
Packit |
577717 |
|
|
Packit |
577717 |
case5: Does setmpx/add/add/start/read work? Added PAPI_TOT_CYC
|
|
Packit |
577717 |
Added PAPI_FP_INS read @start counter[0]: 9223372034707292159
|
|
Packit |
577717 |
read @stop counter[0]: 1843784577 difference counter[0]:
|
|
Packit |
577717 |
-9223372032863507582 multiplex1.c
|
|
Packit |
577717 |
FAILED Line # 389 Error: Difference in start and stop resulted in
|
|
Packit |
577717 |
negative value!
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-08
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/cuda/: linux-cuda.c, linux-cuda.h: Updated CUDA
|
|
Packit |
577717 |
component for CUPTI 4.1 (RC1). Note, SetCudaDevice() should now
|
|
Packit |
577717 |
work with the latest CUDA 4.1 version.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-07
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/coretemp/linux-coretemp.c: Update coretemp to
|
|
Packit |
577717 |
better handle sparse numbering of the inputs.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* doc/Doxyfile-everything: Exclude the libpfm* and perfctr-*
|
|
Packit |
577717 |
directories from consideration when generating Doxygen docs.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: papi.h, components/acpi/linux-acpi.h,
|
|
Packit |
577717 |
components/coretemp_freebsd/coretemp_freebsd.c,
|
|
Packit |
577717 |
components/cuda/linux-cuda.h,
|
|
Packit |
577717 |
components/infiniband/linux-infiniband.h,
|
|
Packit |
577717 |
components/mx/linux-mx.h, components/net/linux-net.h: Place a
|
|
Packit |
577717 |
space in < your name here > to cleanup doxygen warnings.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/perf_events.c: Only perf event systems that have FAST counter
|
|
Packit |
577717 |
reads and FAST hw timer access are x86...
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/linux-common.c: MIPS clock and Linux fixup code
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/example.c: A little more documentation on
|
|
Packit |
577717 |
which of the component vector function pointers are relevant.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_vector.c: Tested the dummy get_{real,virt}_{cyc,usec}
|
|
Packit |
577717 |
functions on zeus, they appear to work.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/tests/example_multiple_components.c:
|
|
Packit |
577717 |
Another fix to properly skip the multiple component case if CPU
|
|
Packit |
577717 |
component not available.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/tests/example_multiple_components.c: Skip
|
|
Packit |
577717 |
the test if no CPU component enabled, rather than fail.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-04
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/example.c: Free example_native_table with
|
|
Packit |
577717 |
papi_free, glibc didn't like it if we just called free. (we
|
|
Packit |
577717 |
allocate it with papi_calloc)
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* man/...: Version number bump. (since the pages are
|
|
Packit |
577717 |
quantifiably different from those released in 4.2.0 )
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* doc/: Doxyfile, Doxyfile-everything, Doxyfile.utils: Bump version
|
|
Packit |
577717 |
number in the doxygen config files.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/example.c:
|
|
Packit |
577717 |
_papi_example_shutdown_substrate does not have any arguments.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/net/linux-net.c: Include ctype.h for isspace().
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* release_procedure.txt: release_procedure now reflects the correct
|
|
Packit |
577717 |
version of doxygen to use.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/buildbot_configure_with_components.sh: Do not always
|
|
Packit |
577717 |
configure with not cpu counters, allow this to be passed in.
|
|
Packit |
577717 |
Allows us to use one script for both types of builds we test.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* delete_before_release.sh,
|
|
Packit |
577717 |
src/buildbot_configure_with_components.sh: Create a script for
|
|
Packit |
577717 |
buildbot to configure with several components.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Buildbot runs all commandline arguments through a sanitization
|
|
Packit |
577717 |
before passing them to sh. Thus --with-configure="a b c" =>
|
|
Packit |
577717 |
'--with-configure="a b c"' which is bad.
|
|
Packit |
577717 |
delete_before_release.sh has been instructed to remove this file.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* man/...: Rebuild the manpages with doxygen 1.7.4 to
|
|
Packit |
577717 |
remove the 's at the end of sentances.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
The html output looks clean.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-03
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: multiplex.c, papi.c: Fix some gcc-4.6 compile warnings
|
|
Packit |
577717 |
complaining that retval was being set but not used.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi.c: Add some extra comments to the PAPI_num_cmp_hwctrs()
|
|
Packit |
577717 |
code that describe its limitations a bit better.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-02
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: ctests/overflow_allcounters.c, testlib/test_utils.c: Add
|
|
Packit |
577717 |
lots of debugging to make results of overflow_allcounters test a
|
|
Packit |
577717 |
bit more clear.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/coretemp/tests/coretemp_pretty.c: coretemp_pretty
|
|
Packit |
577717 |
wasn't printing the description for fan inputs.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
The result on an apple MacBook Pro (running Linux) now looks like
|
|
Packit |
577717 |
this:
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Trying all coretemp events Found coretemp component at cid 2
|
|
Packit |
577717 |
hwmon0.temp1_input value: 33.50 degrees C, applesmc
|
|
Packit |
577717 |
module, label TB0T hwmon0.temp2_input value: 33.50 degrees C,
|
|
Packit |
577717 |
applesmc module, label TB1T hwmon0.temp3_input value: 32.00
|
|
Packit |
577717 |
degrees C, applesmc module, label TB2T hwmon0.temp4_input value:
|
|
Packit |
577717 |
0.00 degrees C, applesmc module, label TB3T hwmon0.temp5_input
|
|
Packit |
577717 |
value: 62.25 degrees C, applesmc module, label TC0D
|
|
Packit |
577717 |
hwmon0.temp6_input value: 54.25 degrees C, applesmc module,
|
|
Packit |
577717 |
label TC0F hwmon0.temp7_input value: 57.25 degrees C, applesmc
|
|
Packit |
577717 |
module, label TC0P hwmon0.temp8_input value: 69.00 degrees C,
|
|
Packit |
577717 |
applesmc module, label TG0D hwmon0.temp9_input value: 58.00
|
|
Packit |
577717 |
degrees C, applesmc module, label TG0F hwmon0.temp10_input
|
|
Packit |
577717 |
value: 51.25 degrees C, applesmc module, label TG0H
|
|
Packit |
577717 |
hwmon0.temp11_input value: 58.25 degrees C, applesmc
|
|
Packit |
577717 |
module, label TG0P hwmon0.temp12_input value: 60.75
|
|
Packit |
577717 |
degrees C, applesmc module, label TG0T hwmon0.temp13_input
|
|
Packit |
577717 |
value: 62.25 degrees C, applesmc module, label TN0D
|
|
Packit |
577717 |
hwmon0.temp14_input value: 59.25 degrees C, applesmc
|
|
Packit |
577717 |
module, label TN0P hwmon0.temp15_input value: 49.00
|
|
Packit |
577717 |
degrees C, applesmc module, label TTF0 hwmon0.temp16_input
|
|
Packit |
577717 |
value: 54.00 degrees C, applesmc module, label Th2H
|
|
Packit |
577717 |
hwmon0.temp17_input value: 58.75 degrees C, applesmc
|
|
Packit |
577717 |
module, label Tm0P hwmon0.temp18_input value: 31.50
|
|
Packit |
577717 |
degrees C, applesmc module, label Ts0P hwmon0.temp19_input
|
|
Packit |
577717 |
value: 44.25 degrees C, applesmc module, label Ts0S
|
|
Packit |
577717 |
hwmon0.fan1_input value: 1999 RPM, applesmc module, label Left
|
|
Packit |
577717 |
side hwmon0.fan2_input value: 2003 RPM, applesmc module,
|
|
Packit |
577717 |
label Right side coretemp_pretty.c PASSED
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/coretemp/: linux-coretemp.c, linux-coretemp.h,
|
|
Packit |
577717 |
tests/coretemp_pretty.c: Make the coretemp code a bit pickier
|
|
Packit |
577717 |
about which events it supports. Add descriptions to the events.
|
|
Packit |
577717 |
Also add support for Voltage (in*) events.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
On an amd14h machine I have access to, coretemp_pretty now
|
|
Packit |
577717 |
prints:
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Trying all coretemp events Found coretemp component at cid 2
|
|
Packit |
577717 |
hwmon0.in1_input value: 1.31 V, it8721 module, label ?
|
|
Packit |
577717 |
hwmon0.in2_input value: 2.22 V, it8721 module, label ?
|
|
Packit |
577717 |
hwmon0.in3_input value: 3.34 V, it8721 module, label +3.3V
|
|
Packit |
577717 |
hwmon0.in4_input value: 1.02 V, it8721 module, label ?
|
|
Packit |
577717 |
hwmon0.in5_input value: 1.52 V, it8721 module, label ?
|
|
Packit |
577717 |
hwmon0.in6_input value: 1.13 V, it8721 module, label ?
|
|
Packit |
577717 |
hwmon0.in7_input value: 3.26 V, it8721 module, label 3VSB
|
|
Packit |
577717 |
hwmon0.in8_input value: 3.17 V, it8721 module, label Vbat
|
|
Packit |
577717 |
hwmon0.temp1_input value: 28.00 degrees C, it8721 module, label ?
|
|
Packit |
577717 |
hwmon0.temp2_input value: -128.00 degrees C, it8721 module, label
|
|
Packit |
577717 |
? hwmon0.temp3_input value: -128.00 degrees C, it8721 module,
|
|
Packit |
577717 |
label ? hwmon0.fan1_input value: 0 RPM hwmon0.fan2_input value:
|
|
Packit |
577717 |
1320 RPM hwmon1.temp1_input value: 33.00 degrees C, jc42 module,
|
|
Packit |
577717 |
label ? hwmon2.temp1_input value: 31.75 degrees C, jc42 module,
|
|
Packit |
577717 |
label ? hwmon3.temp1_input value: 53.00 degrees C, radeon module,
|
|
Packit |
577717 |
label ? hwmon4.temp1_input value: 53.12 degrees C, k10temp
|
|
Packit |
577717 |
module, label ? coretemp_pretty.c PASSED
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/coretemp/: linux-coretemp.c,
|
|
Packit |
577717 |
tests/coretemp_pretty.c: Cut and paste error slipped in to that
|
|
Packit |
577717 |
last commit. Fixes a build issue.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/coretemp/: linux-coretemp.c, tests/Makefile,
|
|
Packit |
577717 |
tests/coretemp_pretty.c: Clean up coretemp with same cleanups
|
|
Packit |
577717 |
done in example component.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Add a new test, "coretemp_pretty" that prints coretemp results in
|
|
Packit |
577717 |
a more user-friendly way.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* man/:... Rebuild the man pages with a newer version of
|
|
Packit |
577717 |
doxygen. ( older versions of doxygen had a nasty bug in man
|
|
Packit |
577717 |
output. )
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Also reworked the utilities documentation to remove pages for the
|
|
Packit |
577717 |
files. Thanks to Jose Pedre Oliveria for pointing this out.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/tests/: Makefile,
|
|
Packit |
577717 |
example_multiple_components.c: Add a test that makes sure you can
|
|
Packit |
577717 |
have active EventSets on multiple components at the same time.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* release_procedure.txt: Change PATH specification to include tcsh
|
|
Packit |
577717 |
syntax; other minor syntax corrections.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/example.c: More cleanups and documentation
|
|
Packit |
577717 |
for the example component.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-11-01
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/example.c: Some more major overhaul of the
|
|
Packit |
577717 |
example component. A lot more documentation, plus make is behave
|
|
Packit |
577717 |
a lot more like a real component would.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* doc/Doxyfile.utils: Turn off undocumented warnings for the utils.
|
|
Packit |
577717 |
doxygen run.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/utils/: avail.c, command_line.c, cost.c, event_chooser.c,
|
|
Packit |
577717 |
multiplex_cost.c: Add spaces to the comments so doxygen doesn't
|
|
Packit |
577717 |
think <event> is an xml tag.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-10-31
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/utils/: avail.c, clockres.c, command_line.c, component.c,
|
|
Packit |
577717 |
cost.c, decode.c, error_codes.c, event_chooser.c, mem_info.c,
|
|
Packit |
577717 |
multiplex_cost.c, native_avail.c: Remove the @file directive from
|
|
Packit |
577717 |
the doxygen comment blocks for the utilities. This cleans up the
|
|
Packit |
577717 |
generated man pages. ( we nolonger build *.c.1 )
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/: example.c, tests/example_basic.c:
|
|
Packit |
577717 |
Clarify in the example component that ->reset only gets called if
|
|
Packit |
577717 |
an eventset is currently running.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Extend the example_basic test to test PAPI_reset()
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* release_procedure.txt: Fix a maketarget typo.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* release_procedure.txt: We now have a good version of doxygen
|
|
Packit |
577717 |
installed on most icl run machines. (
|
|
Packit |
577717 |
/mnt/scratch/sw/doxygen-1.7.5.1 )
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* doc/doxygen_procedure.txt: [no log message]
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* release_procedure.txt: Update release_procedure to inform how to
|
|
Packit |
577717 |
update the website documentation link.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-10-28
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* RELEASENOTES.txt: Correct the RELEASENOTES for some things I
|
|
Packit |
577717 |
missed when reviewing it.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
It's Offcore events that we don't support on
|
|
Packit |
577717 |
Nehalem/Westmere/Sandybridge.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Also the power6 libpfm4 bug that was listed as an outstanding bug
|
|
Packit |
577717 |
was fixed a long time ago.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/coretemp/linux-coretemp.c: Have coretemp set the
|
|
Packit |
577717 |
num_native_events field.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/tests/example_basic.c: Update example test
|
|
Packit |
577717 |
to print num_native_events, to help debug issues with other
|
|
Packit |
577717 |
components not updating the value.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/coretemp/: linux-coretemp.c, linux-coretemp.h: Fix
|
|
Packit |
577717 |
typo enent -> event Also remove residual LMSENSOR mentions from
|
|
Packit |
577717 |
the coretemp header.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/papi_libpfm4_events.c: Fix two memory leak locations.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
The attached patch reduces the number of lost memory blocks
|
|
Packit |
577717 |
reported by valgrind from 234 to 39. It frees the memory
|
|
Packit |
577717 |
allocated by the 4 strdups and the calloc functions in
|
|
Packit |
577717 |
papi_libpfm4_events.c:allocate_native_event().
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Patch by: José Pedro Oliveira
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/cuda/tests/Makefile: The change to pass the PAPI
|
|
Packit |
577717 |
CC/CFLAGS to the component tests broke the nvidia test as it
|
|
Packit |
577717 |
wants CC to be nvcc. So update that Makefile to use nvcc
|
|
Packit |
577717 |
instead.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
2011-10-27
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/tests/example_basic.c: Improve the
|
|
Packit |
577717 |
example_basic component test to be much more comprehensive.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/example/: example.c, tests/HelloWorld.c,
|
|
Packit |
577717 |
tests/Makefile, tests/example_basic.c: Cleanup the example test.
|
|
Packit |
577717 |
Fix various mistakes in the comments as well as add better error
|
|
Packit |
577717 |
checking.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Also rename the "HelloWorld" test to "example_basic"
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/coretemp/tests/Makefile: The coretemp_test target
|
|
Packit |
577717 |
was example_test due to cut-and-paste error.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Patch from Jose Pedro Oliveira
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/Makefile.inc: Add a component_tests dependency so that the
|
|
Packit |
577717 |
component_tests are made during a make -j build
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/Makefile.inc: Make sure the component test makefiles get
|
|
Packit |
577717 |
passed the CC and CFLAGS definitions.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/components/coretemp/: linux-coretemp.c, tests/Makefile,
|
|
Packit |
577717 |
tests/coretemp_basic.c: Fix up the coretemp component some more.
|
|
Packit |
577717 |
Make sure the enumerate function returns PAPI_ENOEVNT if no
|
|
Packit |
577717 |
events are available.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Update the Makefile so it has proper dependencies.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Update the test so it prints the first event available. (The
|
|
Packit |
577717 |
latter based on a patch from Jose Pedro Oliveira)
|
|
Packit |
577717 |
|
|
Packit |
577717 |
* src/: solaris-ultra.c, ctests/all_native_events.c: The
|
|
Packit |
577717 |
solaris-ultra substrate was still broken. This is because recent
|
|
Packit |
577717 |
changes to component bind time explictly used the ->set_domain()
|
|
Packit |
577717 |
call, and this vector was not set up in solaris_ultra.
|
|
Packit |
577717 |
|
|
Packit |
577717 |
Also made the all_native_events test report the returned error
|
|
Packit |
577717 |
value to aid in debugging problems like this in the future.
|