From: Lipeng Zhu <lipeng.zhu@intel.com>
To: "Richard Earnshaw (lists)" <Richard.Earnshaw@arm.com>, jakub@redhat.com
Cc: fortran@gcc.gnu.org, gcc-patches@gcc.gnu.org,
hongjiu.lu@intel.com, pan.deng@intel.com, rep.dot.nop@gmail.com,
tianyou.li@intel.com, tkoenig@netcologne.de,
wangyang.guo@intel.com
Subject: Re: [PATCH v7] libgfortran: Replace mutex with rwlock
Date: Fri, 15 Dec 2023 19:31:53 +0800 [thread overview]
Message-ID: <8192480a-1dfc-4bc6-b141-0254364d3fba@intel.com> (raw)
In-Reply-To: <523c6fbf-344c-48b8-8508-7e2acb088606@arm.com>
On 2023/12/14 23:50, Richard Earnshaw (lists) wrote:
> On 09/12/2023 15:39, Lipeng Zhu wrote:
>> This patch try to introduce the rwlock and split the read/write to
>> unit_root tree and unit_cache with rwlock instead of the mutex to
>> increase CPU efficiency. In the get_gfc_unit function, the percentage
>> to step into the insert_unit function is around 30%, in most instances,
>> we can get the unit in the phase of reading the unit_cache or unit_root
>> tree. So split the read/write phase by rwlock would be an approach to
>> make it more parallel.
>>
>> BTW, the IPC metrics can gain around 9x in our test
>> server with 220 cores. The benchmark we used is
>> https://github.com/rwesson/NEAT
>>
>> libgcc/ChangeLog:
>>
>> * gthr-posix.h (__GTHREAD_RWLOCK_INIT): New macro.
>> (__gthrw): New function.
>> (__gthread_rwlock_rdlock): New function.
>> (__gthread_rwlock_tryrdlock): New function.
>> (__gthread_rwlock_wrlock): New function.
>> (__gthread_rwlock_trywrlock): New function.
>> (__gthread_rwlock_unlock): New function.
>>
>> libgfortran/ChangeLog:
>>
>> * io/async.c (DEBUG_LINE): New macro.
>> * io/async.h (RWLOCK_DEBUG_ADD): New macro.
>> (CHECK_RDLOCK): New macro.
>> (CHECK_WRLOCK): New macro.
>> (TAIL_RWLOCK_DEBUG_QUEUE): New macro.
>> (IN_RWLOCK_DEBUG_QUEUE): New macro.
>> (RDLOCK): New macro.
>> (WRLOCK): New macro.
>> (RWUNLOCK): New macro.
>> (RD_TO_WRLOCK): New macro.
>> (INTERN_RDLOCK): New macro.
>> (INTERN_WRLOCK): New macro.
>> (INTERN_RWUNLOCK): New macro.
>> * io/io.h (struct gfc_unit): Change UNIT_LOCK to UNIT_RWLOCK in
>> a comment.
>> (unit_lock): Remove including associated internal_proto.
>> (unit_rwlock): New declarations including associated internal_proto.
>> (dec_waiting_unlocked): Use WRLOCK and RWUNLOCK on unit_rwlock
>> instead of __gthread_mutex_lock and __gthread_mutex_unlock on
>> unit_lock.
>> * io/transfer.c (st_read_done_worker): Use WRLOCK and RWUNLOCK on
>> unit_rwlock instead of LOCK and UNLOCK on unit_lock.
>> (st_write_done_worker): Likewise.
>> * io/unit.c: Change UNIT_LOCK to UNIT_RWLOCK in 'IO locking rules'
>> comment. Use unit_rwlock variable instead of unit_lock variable.
>> (get_gfc_unit_from_unit_root): New function.
>> (get_gfc_unit): Use RDLOCK, WRLOCK and RWUNLOCK on unit_rwlock
>> instead of LOCK and UNLOCK on unit_lock.
>> (close_unit_1): Use WRLOCK and RWUNLOCK on unit_rwlock instead of
>> LOCK and UNLOCK on unit_lock.
>> (close_units): Likewise.
>> (newunit_alloc): Use RWUNLOCK on unit_rwlock instead of UNLOCK on
>> unit_lock.
>> * io/unix.c (find_file): Use RDLOCK and RWUNLOCK on unit_rwlock
>> instead of LOCK and UNLOCK on unit_lock.
>> (flush_all_units): Use WRLOCK and RWUNLOCK on unit_rwlock instead
>> of LOCK and UNLOCK on unit_lock.
>>
>
> It looks like this has broken builds on arm-none-eabi when using newlib:
>
> In file included from /work/rearnsha/gnusrc/nightly/gcc-cross/master/libgfortran
> /runtime/error.c:27:
> /work/rearnsha/gnusrc/nightly/gcc-cross/master/libgfortran/io/io.h: In function
> ‘dec_waiting_unlocked’:
> /work/rearnsha/gnusrc/nightly/gcc-cross/master/libgfortran/io/io.h:1023:3: error
> : implicit declaration of function ‘WRLOCK’ [-Wimplicit-function-declaration]
> 1023 | WRLOCK (&unit_rwlock);
> | ^~~~~~
> /work/rearnsha/gnusrc/nightly/gcc-cross/master/libgfortran/io/io.h:1025:3: error
> : implicit declaration of function ‘RWUNLOCK’ [-Wimplicit-function-declaration]
> 1025 | RWUNLOCK (&unit_rwlock);
> | ^~~~~~~~
>
>
> R.
Hi Richard,
The root cause is that the macro WRLOCK and RWUNLOCK are not defined in
io.h. The reason of x86 platform not failed is that
HAVE_ATOMIC_FETCH_ADD is defined then caused above macros were never
been used. Code logic show as below:
#ifdef HAVE_ATOMIC_FETCH_ADD
(void) __atomic_fetch_add (&u->waiting, -1, __ATOMIC_RELAXED);
#else
WRLOCK (&unit_rwlock);
u->waiting--;
RWUNLOCK (&unit_rwlock);
#endif
I just draft a patch try to fix this bug, because I didn't have arm
platform, would you help to validate if it was fixed on arm platform?
diff --git a/libgfortran/io/io.h b/libgfortran/io/io.h
index 15daa0995b1..c7f0f7d7d9e 100644
--- a/libgfortran/io/io.h
+++ b/libgfortran/io/io.h
@@ -1020,9 +1020,15 @@ dec_waiting_unlocked (gfc_unit *u)
#ifdef HAVE_ATOMIC_FETCH_ADD
(void) __atomic_fetch_add (&u->waiting, -1, __ATOMIC_RELAXED);
#else
- WRLOCK (&unit_rwlock);
+#ifdef __GTHREAD_RWLOCK_INIT
+ __gthread_rwlock_wrlock (&unit_rwlock);
+ u->waiting--;
+ __gthread_rwlock_unlock (&unit_rwlock);
+#else
+ __gthread_mutex_lock (&unit_rwlock);
u->waiting--;
- RWUNLOCK (&unit_rwlock);
+ __gthread_mutex_unlock (&unit_rwlock);
+#endif
#endif
}
Lipeng Zhu
next prev parent reply other threads:[~2023-12-15 11:31 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-09 2:32 [PATCH v4] " Zhu, Lipeng
2023-05-16 7:08 ` Zhu, Lipeng
2023-05-23 2:53 ` Zhu, Lipeng
2023-05-24 19:18 ` Thomas Koenig
2023-08-18 3:06 ` Zhu, Lipeng
2023-09-14 8:33 ` Zhu, Lipeng
2023-10-23 1:21 ` Zhu, Lipeng
2023-10-23 5:52 ` Thomas Koenig
2023-10-23 23:59 ` Zhu, Lipeng
2023-11-01 10:14 ` Zhu, Lipeng
2023-11-02 9:58 ` Bernhard Reutner-Fischer
2023-11-23 9:36 ` Zhu, Lipeng
2023-12-07 5:18 ` Zhu, Lipeng
2023-08-18 3:18 ` [PATCH v6] " Zhu, Lipeng
2023-12-08 10:19 ` Jakub Jelinek
2023-12-09 15:13 ` Zhu, Lipeng
2023-12-09 15:39 ` [PATCH v7] " Lipeng Zhu
2023-12-09 15:23 ` Jakub Jelinek
2023-12-10 3:25 ` Zhu, Lipeng
2023-12-11 17:45 ` H.J. Lu
2023-12-12 2:05 ` Zhu, Lipeng
2023-12-13 20:52 ` Thomas Schwinge
2023-12-14 2:28 ` Zhu, Lipeng
2023-12-14 12:29 ` Thomas Schwinge
2023-12-14 12:39 ` Jakub Jelinek
2023-12-15 5:43 ` Zhu, Lipeng
2023-12-21 11:42 ` Thomas Schwinge
2023-12-22 6:48 ` Lipeng Zhu
2024-01-03 9:14 ` Lipeng Zhu
2024-01-17 13:25 ` Lipeng Zhu
2023-12-14 15:50 ` Richard Earnshaw (lists)
2023-12-15 11:31 ` Lipeng Zhu [this message]
2023-12-15 19:23 ` Richard Earnshaw
2024-01-02 11:57 ` Vaseeharan Vinayagamoorthy
2024-01-03 1:02 ` Lipeng Zhu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=8192480a-1dfc-4bc6-b141-0254364d3fba@intel.com \
--to=lipeng.zhu@intel.com \
--cc=Richard.Earnshaw@arm.com \
--cc=fortran@gcc.gnu.org \
--cc=gcc-patches@gcc.gnu.org \
--cc=hongjiu.lu@intel.com \
--cc=jakub@redhat.com \
--cc=pan.deng@intel.com \
--cc=rep.dot.nop@gmail.com \
--cc=tianyou.li@intel.com \
--cc=tkoenig@netcologne.de \
--cc=wangyang.guo@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).