* [PATCH] gprofng: test suite problems
@ 2022-04-27 9:36 vladimir.mezentsev
2022-04-28 0:07 ` Nick Alcock
2022-05-25 9:08 ` Luis Machado
0 siblings, 2 replies; 6+ messages in thread
From: vladimir.mezentsev @ 2022-04-27 9:36 UTC (permalink / raw)
To: binutils
From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
gprofng/ChangeLog
2022-04-27 Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
PR gprofng/29065
* testsuite/lib/Makefile.skel: set LD_LIBRARY_PATH
---
gprofng/testsuite/lib/Makefile.skel | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/gprofng/testsuite/lib/Makefile.skel b/gprofng/testsuite/lib/Makefile.skel
index 7134c2715b5..7810ba7ed7a 100644
--- a/gprofng/testsuite/lib/Makefile.skel
+++ b/gprofng/testsuite/lib/Makefile.skel
@@ -39,8 +39,8 @@ DISPLAY = $(GPROFNG) display text
EXPERIMENT = test.er
DISPLAY_LOG = display.log
-
-export LD_LIBRARY_PATH := $(shell dirname $$(find ../root -name libgprofng.so.0 | head -1))
+gprofng_dir := $(shell dirname $$(find ../root -name libgprofng.so.0 | head -1))
+export LD_LIBRARY_PATH := $(gprofng_dir):$(gprofng_dir)/..
.PHONY: all collect compare clobber clean
--
2.27.0
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] gprofng: test suite problems
2022-04-27 9:36 [PATCH] gprofng: test suite problems vladimir.mezentsev
@ 2022-04-28 0:07 ` Nick Alcock
2022-04-28 3:22 ` Vladimir Mezentsev
2022-05-25 9:08 ` Luis Machado
1 sibling, 1 reply; 6+ messages in thread
From: Nick Alcock @ 2022-04-28 0:07 UTC (permalink / raw)
To: vladimir.mezentsev; +Cc: binutils
On 27 Apr 2022, Vladimir Mezentsev via Binutils verbalised:
> From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
>
> gprofng/ChangeLog
> 2022-04-27 Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
>
> PR gprofng/29065
> * testsuite/lib/Makefile.skel: set LD_LIBRARY_PATH
This seems like a misleading description to me. LD_LIBRARY_PATH was
already being set: you're just changing *how* it's set.
Perhaps
> * testsuite/lib/Makefile.skel: Search parent dir for libs too.
(assuming that's actually what it's doing? "root" is a rather odd
directory name, and I say that even though I quite possibly wrote this
code.)
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] gprofng: test suite problems
2022-04-28 0:07 ` Nick Alcock
@ 2022-04-28 3:22 ` Vladimir Mezentsev
0 siblings, 0 replies; 6+ messages in thread
From: Vladimir Mezentsev @ 2022-04-28 3:22 UTC (permalink / raw)
To: Nick Alcock; +Cc: binutils
On 4/27/22 17:07, Nick Alcock wrote:
> On 27 Apr 2022, Vladimir Mezentsev via Binutils verbalised:
>
>> From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
>>
>> gprofng/ChangeLog
>> 2022-04-27 Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
>>
>> PR gprofng/29065
>> * testsuite/lib/Makefile.skel: set LD_LIBRARY_PATH
> This seems like a misleading description to me. LD_LIBRARY_PATH was
> already being set: you're just changing *how* it's set.
>
> Perhaps
>
>> * testsuite/lib/Makefile.skel: Search parent dir for libs too.
Nick, thanks for review.
I will change gprofng/ChangeLog.
> (assuming that's actually what it's doing? "root" is a rather odd
> directory name, and I say that even though I quite possibly wrote this
> code.)
yes. You wrote this code.
-Vladimir
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] gprofng: test suite problems
2022-04-27 9:36 [PATCH] gprofng: test suite problems vladimir.mezentsev
2022-04-28 0:07 ` Nick Alcock
@ 2022-05-25 9:08 ` Luis Machado
2022-05-25 17:19 ` Vladimir Mezentsev
1 sibling, 1 reply; 6+ messages in thread
From: Luis Machado @ 2022-05-25 9:08 UTC (permalink / raw)
To: vladimir.mezentsev, binutils
Hi,
Just wondering, how long is the gprofng testsuite run suppose to take?
It seems to take a long time to execute (aarch64-linux Ubuntu 22.04), to the
point where it makes validation of other components a bit annoying. I clocked it at 7m 53s, on a reasonably
powerful machine.
Ideally one would "make check" everything, but with this particular suite taking
so long, it is tempting to skip it.
Does it parallelize based on -j input?
On 4/27/22 10:36, Vladimir Mezentsev via Binutils wrote:
> From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
>
> gprofng/ChangeLog
> 2022-04-27 Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
>
> PR gprofng/29065
> * testsuite/lib/Makefile.skel: set LD_LIBRARY_PATH
> ---
> gprofng/testsuite/lib/Makefile.skel | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/gprofng/testsuite/lib/Makefile.skel b/gprofng/testsuite/lib/Makefile.skel
> index 7134c2715b5..7810ba7ed7a 100644
> --- a/gprofng/testsuite/lib/Makefile.skel
> +++ b/gprofng/testsuite/lib/Makefile.skel
> @@ -39,8 +39,8 @@ DISPLAY = $(GPROFNG) display text
> EXPERIMENT = test.er
> DISPLAY_LOG = display.log
>
> -
> -export LD_LIBRARY_PATH := $(shell dirname $$(find ../root -name libgprofng.so.0 | head -1))
> +gprofng_dir := $(shell dirname $$(find ../root -name libgprofng.so.0 | head -1))
> +export LD_LIBRARY_PATH := $(gprofng_dir):$(gprofng_dir)/..
>
> .PHONY: all collect compare clobber clean
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] gprofng: test suite problems
2022-05-25 9:08 ` Luis Machado
@ 2022-05-25 17:19 ` Vladimir Mezentsev
2022-05-26 7:26 ` Luis Machado
0 siblings, 1 reply; 6+ messages in thread
From: Vladimir Mezentsev @ 2022-05-25 17:19 UTC (permalink / raw)
To: Luis Machado, binutils
On 5/25/22 02:08, Luis Machado wrote:
> Hi,
>
> Just wondering, how long is the gprofng testsuite run suppose to take?
There are ~15 gprofng tests. Each test runs ~ 1 minute (does not
depend on the machine).
>
> It seems to take a long time to execute (aarch64-linux Ubuntu 22.04),
> to the
> point where it makes validation of other components a bit annoying. I
> clocked it at 7m 53s, on a reasonably
> powerful machine.
>
> Ideally one would "make check" everything, but with this particular
> suite taking
> so long, it is tempting to skip it.
>
> Does it parallelize based on -j input?
If you run only gprofng tests : `(cd BLD_DIR/gprofng; make check -j ...
) `, it doesn't parallelize.
If you run all binutils tests: `(cd BLD_DIR/; make check -j ... ) `,
each testsuite runs in parallel.
Thank you,
-Vladimir
>
> On 4/27/22 10:36, Vladimir Mezentsev via Binutils wrote:
>> From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
>>
>> gprofng/ChangeLog
>> 2022-04-27 Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
>>
>> PR gprofng/29065
>> * testsuite/lib/Makefile.skel: set LD_LIBRARY_PATH
>> ---
>> gprofng/testsuite/lib/Makefile.skel | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/gprofng/testsuite/lib/Makefile.skel
>> b/gprofng/testsuite/lib/Makefile.skel
>> index 7134c2715b5..7810ba7ed7a 100644
>> --- a/gprofng/testsuite/lib/Makefile.skel
>> +++ b/gprofng/testsuite/lib/Makefile.skel
>> @@ -39,8 +39,8 @@ DISPLAY = $(GPROFNG) display text
>> EXPERIMENT = test.er
>> DISPLAY_LOG = display.log
>> -
>> -export LD_LIBRARY_PATH := $(shell dirname $$(find ../root -name
>> libgprofng.so.0 | head -1))
>> +gprofng_dir := $(shell dirname $$(find ../root -name libgprofng.so.0
>> | head -1))
>> +export LD_LIBRARY_PATH := $(gprofng_dir):$(gprofng_dir)/..
>> .PHONY: all collect compare clobber clean
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] gprofng: test suite problems
2022-05-25 17:19 ` Vladimir Mezentsev
@ 2022-05-26 7:26 ` Luis Machado
0 siblings, 0 replies; 6+ messages in thread
From: Luis Machado @ 2022-05-26 7:26 UTC (permalink / raw)
To: Vladimir Mezentsev, binutils
On 5/25/22 18:19, Vladimir Mezentsev wrote:
>
>
> On 5/25/22 02:08, Luis Machado wrote:
>> Hi,
>>
>> Just wondering, how long is the gprofng testsuite run suppose to take?
>
> There are ~15 gprofng tests. Each test runs ~ 1 minute (does not depend on the machine).
>
>
>
>>
>> It seems to take a long time to execute (aarch64-linux Ubuntu 22.04), to the
>> point where it makes validation of other components a bit annoying. I clocked it at 7m 53s, on a reasonably
>> powerful machine.
>>
>> Ideally one would "make check" everything, but with this particular suite taking
>> so long, it is tempting to skip it.
>>
>> Does it parallelize based on -j input?
>
> If you run only gprofng tests : `(cd BLD_DIR/gprofng; make check -j ... ) `, it doesn't parallelize.
> If you run all binutils tests: `(cd BLD_DIR/; make check -j ... ) `, each testsuite runs in parallel.
I see.
Trying a "make check-gprofng -j$(nproc)" with 64 cores still gives me the same 7 minutes or so for the testsuite run.
A single-process "make check-gprofng" gives me around 8 minutes. It doesn't look that much different.
>>
>> On 4/27/22 10:36, Vladimir Mezentsev via Binutils wrote:
>>> From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
>>>
>>> gprofng/ChangeLog
>>> 2022-04-27 Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
>>>
>>> PR gprofng/29065
>>> * testsuite/lib/Makefile.skel: set LD_LIBRARY_PATH
>>> ---
>>> gprofng/testsuite/lib/Makefile.skel | 4 ++--
>>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/gprofng/testsuite/lib/Makefile.skel b/gprofng/testsuite/lib/Makefile.skel
>>> index 7134c2715b5..7810ba7ed7a 100644
>>> --- a/gprofng/testsuite/lib/Makefile.skel
>>> +++ b/gprofng/testsuite/lib/Makefile.skel
>>> @@ -39,8 +39,8 @@ DISPLAY = $(GPROFNG) display text
>>> EXPERIMENT = test.er
>>> DISPLAY_LOG = display.log
>>> -
>>> -export LD_LIBRARY_PATH := $(shell dirname $$(find ../root -name libgprofng.so.0 | head -1))
>>> +gprofng_dir := $(shell dirname $$(find ../root -name libgprofng.so.0 | head -1))
>>> +export LD_LIBRARY_PATH := $(gprofng_dir):$(gprofng_dir)/..
>>> .PHONY: all collect compare clobber clean
>>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2022-05-26 7:26 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-04-27 9:36 [PATCH] gprofng: test suite problems vladimir.mezentsev
2022-04-28 0:07 ` Nick Alcock
2022-04-28 3:22 ` Vladimir Mezentsev
2022-05-25 9:08 ` Luis Machado
2022-05-25 17:19 ` Vladimir Mezentsev
2022-05-26 7:26 ` Luis Machado
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).