[vector] mulAll for byte and short

[Deshpande, Vivek R]

Hi Vladimir

Thanks for your input.
Yes we can align the code in either direction later along with some more aggressive strategies.

Regards,
Vivek

-----Original Message-----
From: Vladimir Ivanov [mailto:vladimir.x.ivanov at oracle.com] 
Sent: Monday, May 21, 2018 11:00 AM
To: Deshpande, Vivek R <vivek.r.deshpande at intel.com>; panama-dev at openjdk.java.net
Subject: Re: [vector] mulAll for byte and short


> We started overloading the nodes AndReductionV onwards to reduce the number nodes.
> I followed the same idea further for this node also.
> Let me know what you think.

 From code/IR perspective, having dedicated node per operation + type pair looks clearer.

Anyway, we can decide and then align the code later all at once. If amount of node classes is a big deal, we can improve on that with more aggressive strategies.

Feel free to push it as it is now.

Best regards,
Vladimir Ivanov

> -----Original Message-----
> From: Vladimir Ivanov [mailto:vladimir.x.ivanov at oracle.com]
> Sent: Friday, May 18, 2018 5:52 PM
> To: Deshpande, Vivek R <vivek.r.deshpande at intel.com>; 
> panama-dev at openjdk.java.net
> Subject: Re: [vector] mulAll for byte and short
> 
> 
>> I would like to contribute a patch for intrinsics for mulAll for byte and short.
>> The webrev is here:
>> http://cr.openjdk.java.net/~vdeshpande/VectorAPI_mulAll/webrev.00/
> 
> Looks good.
> 
> Have you considered introducing MulReductionVB & MulReductionVS nodes instead of overloading MulReductionVI with additional behavior?
> It would simplify other places a bit (e.g., matcher rules).
> 
> Best regards,
> Vladimir Ivanov
> 
>> -----Original Message-----
>> From: Paul Sandoz [mailto:paul.sandoz at oracle.com]
>> Sent: Friday, May 18, 2018 3:19 PM
>> To: Deshpande, Vivek R <vivek.r.deshpande at intel.com>
>> Cc: panama-dev at openjdk.java.net; Viswanathan, Sandhya 
>> <sandhya.viswanathan at intel.com>
>> Subject: Re: [vector] mulAll for byte and short
>>
>> ok, thanks!
>> Paul.
>>
>>> On May 18, 2018, at 1:16 PM, Deshpande, Vivek R <vivek.r.deshpande at intel.com> wrote:
>>>
>>> Hi Paul
>>>
>>> mulAll for byte and short is supported in the API.
>>> I have implemented the intrinsic for mulAll for byte and short and would send the patch soon to review.
>>>
>>> Regards,
>>> Vivek	
>>>
>>> -----Original Message-----
>>> From: panama-dev [mailto:panama-dev-bounces at openjdk.java.net] On 
>>> Behalf Of Paul Sandoz
>>> Sent: Friday, May 18, 2018 12:09 PM
>>> To: panama-dev at openjdk.java.net
>>> Subject: [vector] mulAll for byte and short
>>>
>>> Hi,
>>>
>>> I cannot recall if this was discussed before and i did not find anything in prior emails.
>>>
>>> Out of all the reduction operations mulAll is an outlier and does not support byte and short? is than an oversight or deliberate? I suspect the latter but just wanted to double check.
>>>
>>> Thanks,
>>> Paul.
>>