| Message ID | 1472228046-9002-1-git-send-email-mike.holmes@linaro.org |
|---|---|
| State | New |
| Headers | show |
On Fri, Aug 26, 2016 at 11:14 AM, Mike Holmes <mike.holmes@linaro.org> wrote: > A recent bug showed the limitations of using powers of two, add a prime > number of threads as a new test case. > > Signed-off-by: Mike Holmes <mike.holmes@linaro.org> > Reviewed-by: Bill Fischofer <bill.fischofer@linaro.org> > --- > v2 > Test with max cores (Maxim) > > DEPENDENCIES | 4 ++++ > test/common_plat/performance/odp_scheduling_run.sh | 3 +++ > test/linux-generic/performance/odp_scheduling_run_proc.sh | 3 +++ > 3 files changed, 10 insertions(+) > > diff --git a/DEPENDENCIES b/DEPENDENCIES > index f1f0edd..3c74d6d 100644 > --- a/DEPENDENCIES > +++ b/DEPENDENCIES > @@ -260,3 +260,7 @@ The tested version of doxygen is 1.8.8 > To have this tool also check spelling you need codespell. > # Debian/Ubuntu > #sudo apt install codespell > + > +7.0 Validation tests > + Some tests use nproc to determine how many CPUs are available. > + $ apt-get install coreutils > diff --git a/test/common_plat/performance/odp_scheduling_run.sh > b/test/common_plat/performance/odp_scheduling_run.sh > index 755b0c1..d1480dc 100755 > --- a/test/common_plat/performance/odp_scheduling_run.sh > +++ b/test/common_plat/performance/odp_scheduling_run.sh > @@ -20,6 +20,9 @@ run() > } > > run 1 > +run 5 > run 8 > +run 11 > +run ${nproc} > > exit $ret > diff --git a/test/linux-generic/performance/odp_scheduling_run_proc.sh > b/test/linux-generic/performance/odp_scheduling_run_proc.sh > index b3ef26f..591f86f 100755 > --- a/test/linux-generic/performance/odp_scheduling_run_proc.sh > +++ b/test/linux-generic/performance/odp_scheduling_run_proc.sh > @@ -21,6 +21,9 @@ run() > } > > run 1 > +run 5 > run 8 > +run 11 > +run ${nproc} > > exit $ret > -- > 2.7.4 > >
On 08/26/16 19:14, Mike Holmes wrote: > A recent bug showed the limitations of using powers of two, add a prime > number of threads as a new test case. > > Signed-off-by: Mike Holmes <mike.holmes@linaro.org> > --- > v2 > Test with max cores (Maxim) > > DEPENDENCIES | 4 ++++ > test/common_plat/performance/odp_scheduling_run.sh | 3 +++ > test/linux-generic/performance/odp_scheduling_run_proc.sh | 3 +++ > 3 files changed, 10 insertions(+) > > diff --git a/DEPENDENCIES b/DEPENDENCIES > index f1f0edd..3c74d6d 100644 > --- a/DEPENDENCIES > +++ b/DEPENDENCIES > @@ -260,3 +260,7 @@ The tested version of doxygen is 1.8.8 > To have this tool also check spelling you need codespell. > # Debian/Ubuntu > #sudo apt install codespell > + > +7.0 Validation tests > + Some tests use nproc to determine how many CPUs are available. > + $ apt-get install coreutils > diff --git a/test/common_plat/performance/odp_scheduling_run.sh b/test/common_plat/performance/odp_scheduling_run.sh > index 755b0c1..d1480dc 100755 > --- a/test/common_plat/performance/odp_scheduling_run.sh > +++ b/test/common_plat/performance/odp_scheduling_run.sh > @@ -20,6 +20,9 @@ run() > } > > run 1 > +run 5 > run 8 > +run 11 > +run ${nproc} I don't think that we need additional external dependency only for that case. Use run 0 to run on all available cores. Maxim. > > exit $ret > diff --git a/test/linux-generic/performance/odp_scheduling_run_proc.sh b/test/linux-generic/performance/odp_scheduling_run_proc.sh > index b3ef26f..591f86f 100755 > --- a/test/linux-generic/performance/odp_scheduling_run_proc.sh > +++ b/test/linux-generic/performance/odp_scheduling_run_proc.sh > @@ -21,6 +21,9 @@ run() > } > > run 1 > +run 5 > run 8 > +run 11 > +run ${nproc} > > exit $ret
On 29 August 2016 at 04:08, Maxim Uvarov <maxim.uvarov@linaro.org> wrote: > On 08/26/16 19:14, Mike Holmes wrote: > >> A recent bug showed the limitations of using powers of two, add a prime >> number of threads as a new test case. >> >> Signed-off-by: Mike Holmes <mike.holmes@linaro.org> >> --- >> v2 >> Test with max cores (Maxim) >> >> DEPENDENCIES | 4 ++++ >> test/common_plat/performance/odp_scheduling_run.sh | 3 +++ >> test/linux-generic/performance/odp_scheduling_run_proc.sh | 3 +++ >> 3 files changed, 10 insertions(+) >> >> diff --git a/DEPENDENCIES b/DEPENDENCIES >> index f1f0edd..3c74d6d 100644 >> --- a/DEPENDENCIES >> +++ b/DEPENDENCIES >> @@ -260,3 +260,7 @@ The tested version of doxygen is 1.8.8 >> To have this tool also check spelling you need codespell. >> # Debian/Ubuntu >> #sudo apt install codespell >> + >> +7.0 Validation tests >> + Some tests use nproc to determine how many CPUs are available. >> + $ apt-get install coreutils >> diff --git a/test/common_plat/performance/odp_scheduling_run.sh >> b/test/common_plat/performance/odp_scheduling_run.sh >> index 755b0c1..d1480dc 100755 >> --- a/test/common_plat/performance/odp_scheduling_run.sh >> +++ b/test/common_plat/performance/odp_scheduling_run.sh >> @@ -20,6 +20,9 @@ run() >> } >> run 1 >> +run 5 >> run 8 >> +run 11 >> +run ${nproc} >> > > I don't think that we need additional external dependency only for that > case. Use run 0 to run on all available cores. Ok, perhaps the help needs to specify that behavior and we can use 0 Currently there is no mention of a default or the use of 0 Usage: ./odp_example [options] Options: -c, --count <number> CPU count -h, --help this help -f, --fair collect fairness statistics > > > Maxim. > > > > exit $ret >> diff --git a/test/linux-generic/performance/odp_scheduling_run_proc.sh >> b/test/linux-generic/performance/odp_scheduling_run_proc.sh >> index b3ef26f..591f86f 100755 >> --- a/test/linux-generic/performance/odp_scheduling_run_proc.sh >> +++ b/test/linux-generic/performance/odp_scheduling_run_proc.sh >> @@ -21,6 +21,9 @@ run() >> } >> run 1 >> +run 5 >> run 8 >> +run 11 >> +run ${nproc} >> exit $ret >> > > -- Mike Holmes Program Manager - Linaro Networking Group Linaro.org <http://www.linaro.org/> *│ *Open source software for ARM SoCs "Work should be fun and collaborative, the rest follows"
On Mon, Aug 29, 2016 at 7:40 AM, Mike Holmes <mike.holmes@linaro.org> wrote: > On 29 August 2016 at 04:08, Maxim Uvarov <maxim.uvarov@linaro.org> wrote: > > > On 08/26/16 19:14, Mike Holmes wrote: > > > >> A recent bug showed the limitations of using powers of two, add a prime > >> number of threads as a new test case. > >> > >> Signed-off-by: Mike Holmes <mike.holmes@linaro.org> > >> --- > >> v2 > >> Test with max cores (Maxim) > >> > >> DEPENDENCIES | 4 ++++ > >> test/common_plat/performance/odp_scheduling_run.sh | 3 +++ > >> test/linux-generic/performance/odp_scheduling_run_proc.sh | 3 +++ > >> 3 files changed, 10 insertions(+) > >> > >> diff --git a/DEPENDENCIES b/DEPENDENCIES > >> index f1f0edd..3c74d6d 100644 > >> --- a/DEPENDENCIES > >> +++ b/DEPENDENCIES > >> @@ -260,3 +260,7 @@ The tested version of doxygen is 1.8.8 > >> To have this tool also check spelling you need codespell. > >> # Debian/Ubuntu > >> #sudo apt install codespell > >> + > >> +7.0 Validation tests > >> + Some tests use nproc to determine how many CPUs are available. > >> + $ apt-get install coreutils > >> diff --git a/test/common_plat/performance/odp_scheduling_run.sh > >> b/test/common_plat/performance/odp_scheduling_run.sh > >> index 755b0c1..d1480dc 100755 > >> --- a/test/common_plat/performance/odp_scheduling_run.sh > >> +++ b/test/common_plat/performance/odp_scheduling_run.sh > >> @@ -20,6 +20,9 @@ run() > >> } > >> run 1 > >> +run 5 > >> run 8 > >> +run 11 > >> +run ${nproc} > >> > > > > I don't think that we need additional external dependency only for that > > case. Use run 0 to run on all available cores. > > > Ok, perhaps the help needs to specify that behavior and we can use 0 > > Currently there is no mention of a default or the use of 0 > > Usage: ./odp_example [options] > Options: > -c, --count <number> CPU count > -h, --help this help > -f, --fair collect fairness statistics > I like the convention that 0 == use all available cores as we use that convention elsewhere in ODP (e.g., in memory limits associated with capabilities). Agree that the doc needs to reflect that convention. > > > > > > > > > Maxim. > > > > > > > > exit $ret > >> diff --git a/test/linux-generic/performance/odp_scheduling_run_proc.sh > >> b/test/linux-generic/performance/odp_scheduling_run_proc.sh > >> index b3ef26f..591f86f 100755 > >> --- a/test/linux-generic/performance/odp_scheduling_run_proc.sh > >> +++ b/test/linux-generic/performance/odp_scheduling_run_proc.sh > >> @@ -21,6 +21,9 @@ run() > >> } > >> run 1 > >> +run 5 > >> run 8 > >> +run 11 > >> +run ${nproc} > >> exit $ret > >> > > > > > > > -- > Mike Holmes > Program Manager - Linaro Networking Group > Linaro.org <http://www.linaro.org/> *│ *Open source software for ARM SoCs > "Work should be fun and collaborative, the rest follows" >
diff --git a/DEPENDENCIES b/DEPENDENCIES index f1f0edd..3c74d6d 100644 --- a/DEPENDENCIES +++ b/DEPENDENCIES @@ -260,3 +260,7 @@ The tested version of doxygen is 1.8.8 To have this tool also check spelling you need codespell. # Debian/Ubuntu #sudo apt install codespell + +7.0 Validation tests + Some tests use nproc to determine how many CPUs are available. + $ apt-get install coreutils diff --git a/test/common_plat/performance/odp_scheduling_run.sh b/test/common_plat/performance/odp_scheduling_run.sh index 755b0c1..d1480dc 100755 --- a/test/common_plat/performance/odp_scheduling_run.sh +++ b/test/common_plat/performance/odp_scheduling_run.sh @@ -20,6 +20,9 @@ run() } run 1 +run 5 run 8 +run 11 +run ${nproc} exit $ret diff --git a/test/linux-generic/performance/odp_scheduling_run_proc.sh b/test/linux-generic/performance/odp_scheduling_run_proc.sh index b3ef26f..591f86f 100755 --- a/test/linux-generic/performance/odp_scheduling_run_proc.sh +++ b/test/linux-generic/performance/odp_scheduling_run_proc.sh @@ -21,6 +21,9 @@ run() } run 1 +run 5 run 8 +run 11 +run ${nproc} exit $ret
A recent bug showed the limitations of using powers of two, add a prime number of threads as a new test case. Signed-off-by: Mike Holmes <mike.holmes@linaro.org> --- v2 Test with max cores (Maxim) DEPENDENCIES | 4 ++++ test/common_plat/performance/odp_scheduling_run.sh | 3 +++ test/linux-generic/performance/odp_scheduling_run_proc.sh | 3 +++ 3 files changed, 10 insertions(+) -- 2.7.4