<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
<meta name="Generator" content="Microsoft Exchange Server">
<!-- converted from text --><style><!-- .EmailQuote { margin-left: 1pt; padding-left: 4pt; border-left: #800000 2px solid; } --></style>
</head>
<body>
<meta content="text/html; charset=UTF-8">
<style type="text/css" style="">
<!--
p
{margin-top:0;
margin-bottom:0}
-->
</style>
<div dir="ltr">
<div id="x_divtagdefaultwrapper" dir="ltr" style="font-size:12pt; color:#000000; font-family:Calibri,Helvetica,sans-serif">
<p></p>
<div style="font-family:Calibri,Helvetica,sans-serif,serif,EmojiFont; font-size:16px; margin-top:0px; margin-bottom:0px">
Yes, I agree that an array can be zero sized.</div>
<div style="font-family:Calibri,Helvetica,sans-serif,serif,EmojiFont; font-size:16px; margin-top:0px; margin-bottom:0px">
<br>
</div>
<div style="font-family:Calibri,Helvetica,sans-serif,serif,EmojiFont; font-size:16px; margin-top:0px; margin-bottom:0px">
I still do not see a reason to check `data.length <= 0`.</div>
<div style="font-family:Calibri,Helvetica,sans-serif,serif,EmojiFont; font-size:16px; margin-top:0px; margin-bottom:0px">
`data.length u<= iters - 1`, which is the upper bounds check, should be</div>
<div style="font-family:Calibri,Helvetica,sans-serif,serif,EmojiFont; font-size:16px; margin-top:0px; margin-bottom:0px">
sufficient to determine whether <font size="3"><span style="font-size:12pt">we should trap.</span></font></div>
<br>
<p></p>
</div>
<hr tabindex="-1" style="display:inline-block; width:98%">
<div id="x_divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" color="#000000" style="font-size:11pt"><b>From:</b> Vladimir Kozlov <vladimir.kozlov@oracle.com><br>
<b>Sent:</b> Thursday, February 1, 2024 12:02:15 PM<br>
<b>To:</b> Cao, Joshua; hotspot-compiler-dev@openjdk.org<br>
<b>Subject:</b> RE: [EXTERNAL] [External] : Re: What is the purpose of loop predication lower bounds check</font>
<div> </div>
</div>
</div>
<font size="2"><span style="font-size:10pt;">
<div class="PlainText">CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.<br>
<br>
<br>
<br>
In general we usually generate different checks during parsing and<br>
optimize/collapse/remove/merge them later based on full information<br>
after parsing.<br>
<br>
On 2/1/24 11:55 AM, Vladimir Kozlov wrote:<br>
> If array is external it could be 0 sized.<br>
> With locally allocated array the check should collapse unless we missing<br>
> such optimization.<br>
><br>
> On 2/1/24 11:52 AM, Cao, Joshua wrote:<br>
>> If I use the same example but with negative stride:<br>
>><br>
>><br>
>> ```<br>
>><br>
>> for (int i = iters - 1; i >= 0; --i) {<br>
>><br>
>> blackhole(data[i]);<br>
>><br>
>> }<br>
>> ```<br>
>><br>
>> We still generate these two checks:<br>
>> ```<br>
>><br>
>> if (data.length <= 0) { trap() }<br>
>> if (data.length <= iters - 1) { trap() }<br>
>><br>
>> ```<br>
>><br>
>><br>
>> So even with negative stride, the upper bounds check is sufficient by<br>
>> itself.<br>
>><br>
>><br>
>> I am playing around with stride/scale/offset and I cannot come up with an<br>
>><br>
>> example where the lower bounds check is important.<br>
>><br>
>> ------------------------------------------------------------------------<br>
>> *From:* hotspot-compiler-dev <hotspot-compiler-dev-retn@openjdk.org><br>
>> on behalf of Vladimir Kozlov <vladimir.kozlov@oracle.com><br>
>> *Sent:* Thursday, February 1, 2024 11:44:29 AM<br>
>> *To:* hotspot-compiler-dev@openjdk.org<br>
>> *Subject:* RE: [EXTERNAL] What is the purpose of loop predication<br>
>> lower bounds check<br>
>> CAUTION: This email originated from outside of the organization. Do<br>
>> not click links or open attachments unless you can confirm the sender<br>
>> and know the content is safe.<br>
>><br>
>><br>
>><br>
>> Initial index value is variable.<br>
>><br>
>> Reverse (negative stride) processing of array?<br>
>><br>
>> On 2/1/24 11:31 AM, Cao, Joshua wrote:<br>
>>> Can I get some help with <a href="https://bugs.openjdk.org/browse/JDK-8325146?">
https://bugs.openjdk.org/browse/JDK-8325146?</a><br>
>> <<a href="https://bugs.openjdk.org/browse/JDK-8325146?">https://bugs.openjdk.org/browse/JDK-8325146?</a>><br>
>>> What is<br>
>>> the purpose of the lower bounds check for loop predication? It seems<br>
>>> like an<br>
>>> unnecessary if-check for any example I can come up with.<br>
>>><br>
>>> Whatever the reason might be, I think there are cases in which we can<br>
>>> ellide<br>
>>> the lower bounds check, like in the example in the JBS issue.<br>
>>><br>
>>><br>
</div>
</span></font>
</body>
</html>