RFR (XS): 8217371: C2: Incorrect LP64 guard in x86.ad after JDK-8210764 (Update avx512 implementation)

Deshpande, Vivek R vivek.r.deshpande at intel.com
Wed Jan 30 21:52:53 UTC 2019


Hi Vladimir

I tested the patch with Lucene tests.
Then I have pushed the patch:
http://hg.openjdk.java.net/jdk/jdk/rev/6121eee15c23
Thanks.

Regards,
Vivek

-----Original Message-----
From: Vladimir Kozlov [mailto:vladimir.kozlov at oracle.com] 
Sent: Monday, January 28, 2019 9:39 AM
To: Viswanathan, Sandhya <sandhya.viswanathan at intel.com>; hotspot-compiler-dev at openjdk.java.net
Cc: Deshpande, Vivek R <vivek.r.deshpande at intel.com>
Subject: Re: RFR (XS): 8217371: C2: Incorrect LP64 guard in x86.ad after JDK-8210764 (Update avx512 implementation)

Hi Sandhya,

Can you also run Lucene tests which hit previous avx512 issues on SKX and Knights?
This is spilling code and it is used when a lot of xmm registers are used and our jtreg tests may not use this code.

You can push after that.

Thanks,
Vladimir

On 1/28/19 9:22 AM, Viswanathan, Sandhya wrote:
> Thanks Vladimir, Tobias and Nils.
> 
> Could Vivek go ahead and push it?
> 
> Best Regards,
> Sandhya
> 
> 
> -----Original Message-----
> From: Vladimir Kozlov [mailto:vladimir.kozlov at oracle.com]
> Sent: Monday, January 28, 2019 9:17 AM
> To: hotspot-compiler-dev at openjdk.java.net; Viswanathan, Sandhya 
> <sandhya.viswanathan at intel.com>
> Subject: Re: RFR (XS): 8217371: C2: Incorrect LP64 guard in x86.ad 
> after JDK-8210764 (Update avx512 implementation)
> 
> +1
> 
> Thanks,
> Vladimir
> 
> On 1/28/19 2:00 AM, Tobias Hartmann wrote:
>> Hi Sandhya,
>>
>> looks good to me too.
>>
>> Best regards,
>> Tobias
>>
>> On 28.01.19 09:40, Nils Eliasson wrote:
>>> Hi Sandhya,
>>>
>>> Looks good,
>>>
>>> Regards,
>>>
>>> Nils
>>>
>>> On 2019-01-27 04:47, Viswanathan, Sandhya wrote:
>>>>
>>>> Hi All,
>>>>
>>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8217371
>>>> <https://bugs.openjdk.java.net/browse/JDK-8217371>
>>>>
>>>> Webrev: http://cr.openjdk.java.net/~sviswanathan/8217371/webrev.00/
>>>> <http://cr.openjdk.java.net/~sviswanathan/8217371/webrev.00/>
>>>>
>>>> The above webrev fixes the incorrect LP64 guard issue in x86.ad file.
>>>>
>>>> I have corrected the guard to _LP64 and updated the spill/fill 
>>>> instructions.  This bug only affected the Knights family where AVX512VL is not supported.
>>>>    
>>>> I have tested it on SKX and Knights family with compiler jtreg tests.
>>>> Please review.
>>>>
>>>>
>>>> Best Regards,
>>>> Sandhya
>>>>


More information about the hotspot-compiler-dev mailing list