Content-Type: multipart/alternative;
 boundary="------------w8mXMYuDd0s7lf0LOO6aA0md"
Message-ID:
 <OS0P286MB016392E0B2017DDB6709A11C82EB2@OS0P286MB0163.JPNP286.PROD.OUTLOOK.COM>
Date: Wed, 22 May 2024 17:27:44 +0800
User-Agent: Mozilla Thunderbird
Content-Language: en-US
To: pgsql-general@postgresql.org
From: XChy <xxs_chy@outlook.com>
Subject: Missed compiler optimization issue in function
 select_rtable_names_for_explain
MIME-Version: 1.0
X-MS-Exchange-MessageSentRepresentingType: 1
X-MS-PublicTrafficType: Email
X-MS-TrafficTypeDiagnostic: OS0P286MB0163:EE_|TYCP286MB3819:EE_
X-MS-Office365-Filtering-Correlation-Id: 9557a589-ac99-4151-a2a1-08dc7a417254
X-Microsoft-Antispam:
	BCL:0;ARA:14566002|9400799015|461199019|440099019|3412199016|1602099003;
X-Microsoft-Antispam-Message-Info:
	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
X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1
X-MS-Exchange-AntiSpam-MessageData-0:
	=?utf-8?B?VmVxMElNMW9wdWNoMFhrVGlUMFZyVW9xRFVGZEg0WkhYYjYveEhaUi9PLzZU?=
 =?utf-8?B?a2MxVzhDMTJXUU1uUkdNVUVSWmowS0xUQjNkaldnSGVMTzZRWUJTdi8yVUps?=
 =?utf-8?B?aW9nWVhuaHQxa2U2cUVxeUxuNTFrZUpIbHpmSXpTa0tDRXR3VGpJaEhWaUx3?=
 =?utf-8?B?RTRMWlV2WGp1bjRDb3ZnWWRUL1JkcHNtRFUwTThJMGhSQW5BZ3VrWWplSmZC?=
 =?utf-8?B?UmtNMnZhckNoazFEZnV4VlpIbHltSUo4eGFmS0ZacjlvaXZ3MTZraXFkd0hD?=
 =?utf-8?B?V01pRSsvQnRUUkJQdmhOdXYvSER6MEFOQVc1K0FXQXRxTVVjeXBWYkJocTFy?=
 =?utf-8?B?SDhDTUdJQnBRRnQ2Ky9uc1dFRWd6MWdOSnpsQUJyZmgrcnUyY250Q0N0dzd3?=
 =?utf-8?B?Z0NEeWFXR0RBWUVmUTNEVUxwOFFKcXJENUppQTM0cjNKelkwdWk2OHlHbzl4?=
 =?utf-8?B?Y0N4RWR2V0xZVTc3aWlEUitiaDVrZ1dadzBRSk1uSnY5TWRQbWlPVVRkLzlQ?=
 =?utf-8?B?RHFsVVlJcUNBRzJIb1NtckJyOUFYMGl0blk1QlFRMjFUek9MeU9zbmxVRklh?=
 =?utf-8?B?L3gzdmlSUUtKUHF0ZlEvczVSWnEydTNCWlB4V2tlN2gvNHFxS1d3dHJoVzcw?=
 =?utf-8?B?Ykl6MzFCM0pXOFFoc2JMcnVWYU4yZDJ3L0ErNkE3Tk9aUzhQc1hRZ2NrTnhH?=
 =?utf-8?B?dDkyR3BqOCtyMTJiajFCckt6WjRqSHl6UlhFSVBvdy9WRUUrWW5USytPQmpx?=
 =?utf-8?B?TDViUWJwc2dnV3dhc1FkODFsV3BZc0g3bFNQK093dW52a0M1OW0xVDRIQ2Jt?=
 =?utf-8?B?Rk92SUFUVXJEN0JyanBOOHd5Q3V2UFk4QVRKNThhWUVHWTM1VXdnaXBrVm9C?=
 =?utf-8?B?bWJHT25zRWVvKzdENm9FdWhyQ00zMElUbFEya0pjSFR2Wi81eUE0SWxTSDRW?=
 =?utf-8?B?K1hDM09Scmp2czJETTIyME9UTUlCeHB6UkRWSk1BS0VYYWpENm9BYnJlNytn?=
 =?utf-8?B?VGRPS2liWkhlQVVwcUpWTGJNRi8zZ3oxTk9ZOWMwOXFrMkxIejVpZkpXZEls?=
 =?utf-8?B?c29qeTRYblM2aGgwaUJIWTBmOHFlMzVFUVMxMXNtbVR4ZEpJK1BoaGpwTXpE?=
 =?utf-8?B?UVhIMlE1ZWQ0WW5sdy9ESGxKMkVKQmN1Q3YwT3RURmx1WVFPN3BWVXlVMEVl?=
 =?utf-8?B?WXpiQUxPVHRxdmRuSjExWjhNTjJKTVZFVVVQZHdVQ3ZNd0lWNFJRU1pJNlZC?=
 =?utf-8?B?c2tUQ1lDQktWd1NYWEZmT0pFY1dQcGUzdkY1MFB2a2lwTC9rRDF5TTZRRVdN?=
 =?utf-8?B?TUJaUENsNUxVRyt2VFhkTVVOWUZ4Mzh0YnpTdjhncHNCcXJmbGVTeXpTYXlM?=
 =?utf-8?B?eDJ5WE9PSVdkdDVJUitiYkJydlNDNlZOeFhtdXo0T1BiRWk0a3J2NCthdVht?=
 =?utf-8?B?QmJSNHFGRUZlc0RGaFJESGJnRkMyNHVkWTBVL1pXUVcyMWM5dUZ1cFptU3Rp?=
 =?utf-8?B?MW45dHNpNjlzYUc0czdaZkQwd3FxOVdLYk90bnVPLzNoZXZ2YUU5bTlUTlA1?=
 =?utf-8?B?ampoOGlwZE0wQnUyRDI5SmZDWXVzQzY4L3BnaldkRmJNVFcwNy9LY3hEaGEv?=
 =?utf-8?Q?kLkkU3dgeG+br0xKDZDhG4+Z/QRCHWGdsWmObyyT6T0w=3D?=
X-OriginatorOrg: outlook.com
X-MS-Exchange-CrossTenant-Network-Message-Id: 9557a589-ac99-4151-a2a1-08dc7a417254
X-MS-Exchange-CrossTenant-AuthSource: OS0P286MB0163.JPNP286.PROD.OUTLOOK.COM
X-MS-Exchange-CrossTenant-AuthAs: Internal
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 22 May 2024 09:27:49.4314
 (UTC)
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 84df9e7f-e9f6-40af-b435-aaaaaaaaaaaa
X-MS-Exchange-CrossTenant-RMS-PersistedConsumerOrg:
	00000000-0000-0000-0000-000000000000
X-MS-Exchange-Transport-CrossTenantHeadersStamped: TYCP286MB3819
List-Id: <pgsql-general.lists.postgresql.org>
List-Help: <https://lists.postgresql.org/manage/>
List-Subscribe: <https://lists.postgresql.org/manage/>
List-Post: <mailto:pgsql-general@lists.postgresql.org>
List-Owner: <mailto:pgsql-general-owner@lists.postgresql.org>
List-Archive: <https://www.postgresql.org/list/pgsql-general>
Archived-At: <https://www.postgresql.org/message-id/OS0P286MB016392E0B2017DDB6709A11C82EB2%40OS0P286MB0163.JPNP286.PROD.OUTLOOK.COM>
Precedence: bulk

--------------w8mXMYuDd0s7lf0LOO6aA0md
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Hi everyone,

I'm a compiler developer working on detecting missed optimization in 
real-world applications. Recently, we found that LLVM missed a dead 
store elimination optimization in the PostgreSQL code 
<https://github.com/postgres/postgres/blob/c37267162e889fe783786b9e28d1b65b82365a00/src/backend/utils/adt/ruleutils.c#L3794> 
(https://github.com/postgres/postgres/blob/master/src/backend/utils/adt/ruleutils.c#L3794) 
in the master branch.

For the example below:

```

int dst[128];

memset(dst, 0, 128);

*unrelated = some_value;

dst[1] = 0;

dst[2] = 0;

```

LLVM cannot eliminate the useless stores after memset since the store to 
"unrelated" clobbers. But if we put the stores to "dst" ahead of the 
store to "unrelated", we could prevent confusing the compiler. See also 
the compiler explorer link: https://godbolt.org/z/P9jnKod3v and the 
issue of LLVM: https://github.com/llvm/llvm-project/issues/88632

To improve the codegen quality, I think it's also possible to modify the 
source code, changing the order of initialization of the member, to get 
better optimization. But I don't know whether this can be considered as 
a bug, thus post the issue here.

If anyone could confirm this problem or post a patch for it, let me know 
please. Thanks!


Best regards. Hongyu.

--------------w8mXMYuDd0s7lf0LOO6aA0md
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: 7bit

<!DOCTYPE html><html><head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body>
    <p>Hi everyone,</p>
    <p>I'm a compiler developer working on detecting missed optimization
      in real-world applications. Recently, we found that LLVM missed a
      dead store elimination optimization in the PostgreSQL <a href="https://github.com/postgres/postgres/blob/c37267162e889fe783786b9e28d1b65b82365a00/src/backend/utils/adt/ruleutils.c#L3794">code</a>
(<a class="moz-txt-link-freetext" href="https://github.com/postgres/postgres/blob/master/src/backend/utils/adt/ruleutils.c#L3794">https://github.com/postgres/postgres/blob/master/src/backend/utils/adt/ruleutils.c#L3794</a>)
      in the master branch. <br>
    </p>
    <p>For the example below:</p>
    <p>```<br>
    </p>
    <p align="justify">int dst[128];<br>
    </p>
    <p align="justify">memset(dst, 0, 128);</p>
    <p align="justify">*unrelated = some_value;</p>
    <p align="justify">dst[1] = 0;</p>
    <p align="justify">dst[2] = 0;</p>
    <p align="justify">```<br>
    </p>
    <p>LLVM cannot eliminate the useless stores after memset since the
      store to &quot;unrelated&quot; clobbers. But if we put the stores to &quot;dst&quot;
      ahead of the store to &quot;unrelated&quot;, we could prevent confusing the
      compiler. See also the compiler explorer link:
      <a class="moz-txt-link-freetext" href="https://godbolt.org/z/P9jnKod3v">https://godbolt.org/z/P9jnKod3v</a> and the issue of LLVM:
      <a class="moz-txt-link-freetext" href="https://github.com/llvm/llvm-project/issues/88632">https://github.com/llvm/llvm-project/issues/88632</a></p>
    <p>To improve the codegen quality, I think it's also possible to
      modify the source code, changing the order of initialization of
      the member, to get better optimization. But I don't know whether
      this can be considered as a bug, thus post the issue here.<br>
    </p>
    <p>If anyone could confirm this problem or post a patch for it, let
      me know please. Thanks!<br>
    </p>
    <p><br>
    </p>
    <p>Best regards. Hongyu.<br>
    </p>
  </body>
</html>

--------------w8mXMYuDd0s7lf0LOO6aA0md--