diff mbox series

test-memcpy-support.h: Double TIMEOUT to (16 * 60)

Message ID 20211204175951.4028115-1-hjl.tools@gmail.com
State Superseded
Headers show
Series test-memcpy-support.h: Double TIMEOUT to (16 * 60) | expand

Checks

Context Check Description
dj/TryBot-apply_patch success Patch applied to master at the time it was sent
dj/TryBot-32bit success Build for i686

Commit Message

H.J. Lu Dec. 4, 2021, 5:59 p.m. UTC
On Ice Lake and Tiger Lake laptops, test-memcpy and test-mempcpy timeout
when there are 3 "make check -j8" runs in parallel. Double TIMEOUT to
(16 * 60) to avoid timeout.
---
 string/test-memcpy-support.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Florian Weimer Dec. 4, 2021, 6:08 p.m. UTC | #1
* H. J. Lu via Libc-alpha:

> On Ice Lake and Tiger Lake laptops, test-memcpy and test-mempcpy timeout
> when there are 3 "make check -j8" runs in parallel. Double TIMEOUT to
> (16 * 60) to avoid timeout.
> ---
>  string/test-memcpy-support.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/string/test-memcpy-support.h b/string/test-memcpy-support.h
> index 419158a420..892fa71449 100644
> --- a/string/test-memcpy-support.h
> +++ b/string/test-memcpy-support.h
> @@ -26,7 +26,7 @@
>  #define MIN_PAGE_SIZE 131072
>  #define TEST_MAIN
>  #define TEST_NAME "memcpy"
> -#define TIMEOUT (8 * 60)
> +#define TIMEOUT (16 * 60)
>  #include "test-string.h"
>  
>  char *simple_memcpy (char *, const char *, size_t);

Shouldn't you be using TIMEOUTFACTOR if running the test suite in
parallel?  There are likely other issues as well.

Thanks,
Florian
H.J. Lu Dec. 4, 2021, 6:41 p.m. UTC | #2
On Sat, Dec 4, 2021 at 10:08 AM Florian Weimer <fweimer@redhat.com> wrote:
>
> * H. J. Lu via Libc-alpha:
>
> > On Ice Lake and Tiger Lake laptops, test-memcpy and test-mempcpy timeout
> > when there are 3 "make check -j8" runs in parallel. Double TIMEOUT to
> > (16 * 60) to avoid timeout.
> > ---
> >  string/test-memcpy-support.h | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/string/test-memcpy-support.h b/string/test-memcpy-support.h
> > index 419158a420..892fa71449 100644
> > --- a/string/test-memcpy-support.h
> > +++ b/string/test-memcpy-support.h
> > @@ -26,7 +26,7 @@
> >  #define MIN_PAGE_SIZE 131072
> >  #define TEST_MAIN
> >  #define TEST_NAME "memcpy"
> > -#define TIMEOUT (8 * 60)
> > +#define TIMEOUT (16 * 60)
> >  #include "test-string.h"
> >
> >  char *simple_memcpy (char *, const char *, size_t);
>
> Shouldn't you be using TIMEOUTFACTOR if running the test suite in
> parallel?  There are likely other issues as well.
>
> Thanks,
> Florian
>

I prefer a configure option over an environment variable.  I'd like to
add --with-timeoutfactor=.
Noah Goldstein Dec. 4, 2021, 7:23 p.m. UTC | #3
On Sat, Dec 4, 2021 at 12:42 PM H.J. Lu via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> On Sat, Dec 4, 2021 at 10:08 AM Florian Weimer <fweimer@redhat.com> wrote:
> >
> > * H. J. Lu via Libc-alpha:
> >
> > > On Ice Lake and Tiger Lake laptops, test-memcpy and test-mempcpy timeout
> > > when there are 3 "make check -j8" runs in parallel. Double TIMEOUT to
> > > (16 * 60) to avoid timeout.

Maybe instead of increasing the timeout we should make the extra memcpy
tests only run if explicitly set to. We only really need the run them
when making
changes to memcpy, otherwise its basically adding several minutes to the test
time.

> > > ---
> > >  string/test-memcpy-support.h | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/string/test-memcpy-support.h b/string/test-memcpy-support.h
> > > index 419158a420..892fa71449 100644
> > > --- a/string/test-memcpy-support.h
> > > +++ b/string/test-memcpy-support.h
> > > @@ -26,7 +26,7 @@
> > >  #define MIN_PAGE_SIZE 131072
> > >  #define TEST_MAIN
> > >  #define TEST_NAME "memcpy"
> > > -#define TIMEOUT (8 * 60)
> > > +#define TIMEOUT (16 * 60)
> > >  #include "test-string.h"
> > >
> > >  char *simple_memcpy (char *, const char *, size_t);
> >
> > Shouldn't you be using TIMEOUTFACTOR if running the test suite in
> > parallel?  There are likely other issues as well.
> >
> > Thanks,
> > Florian
> >
>
> I prefer a configure option over an environment variable.  I'd like to
> add --with-timeoutfactor=.
>
> --
> H.J.
H.J. Lu Dec. 4, 2021, 7:37 p.m. UTC | #4
On Sat, Dec 4, 2021 at 11:24 AM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
>
> On Sat, Dec 4, 2021 at 12:42 PM H.J. Lu via Libc-alpha
> <libc-alpha@sourceware.org> wrote:
> >
> > On Sat, Dec 4, 2021 at 10:08 AM Florian Weimer <fweimer@redhat.com> wrote:
> > >
> > > * H. J. Lu via Libc-alpha:
> > >
> > > > On Ice Lake and Tiger Lake laptops, test-memcpy and test-mempcpy timeout
> > > > when there are 3 "make check -j8" runs in parallel. Double TIMEOUT to
> > > > (16 * 60) to avoid timeout.
>
> Maybe instead of increasing the timeout we should make the extra memcpy
> tests only run if explicitly set to. We only really need the run them
> when making
> changes to memcpy, otherwise its basically adding several minutes to the test
> time.

I think we should always run all tests.  I am working on the v2 patch to add
--with-timeoutfactor=NUM to specify an integer to scale the timeout.
diff mbox series

Patch

diff --git a/string/test-memcpy-support.h b/string/test-memcpy-support.h
index 419158a420..892fa71449 100644
--- a/string/test-memcpy-support.h
+++ b/string/test-memcpy-support.h
@@ -26,7 +26,7 @@ 
 #define MIN_PAGE_SIZE 131072
 #define TEST_MAIN
 #define TEST_NAME "memcpy"
-#define TIMEOUT (8 * 60)
+#define TIMEOUT (16 * 60)
 #include "test-string.h"
 
 char *simple_memcpy (char *, const char *, size_t);