RFR: 8309893: Integrate ReplicateB/S/I/L/F/D nodes to Replicate node [v2]

Eric Liu eliu at openjdk.org
Tue Oct 31 10:10:39 UTC 2023


On Wed, 25 Oct 2023 09:23:02 GMT, Emanuel Peter <epeter at openjdk.org> wrote:

>> Eric Liu has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains three commits:
>> 
>>  - small fix
>>    
>>    Change-Id: I121a8c1ad75e88c0af41f2c1ef00289a3d61f400
>>  - Merge jdk:master
>>    
>>    Change-Id: I4df10ba7290e9af02deeac5a931234a07be7c207
>>  - 8309893: Integrate ReplicateB/S/I/L/F/D nodes to Replicate node
>>    
>>    This patch creates ReplicateNode to replace ReplicateB/S/I/L/F/DNode,
>>    like other vector nodes introduced recently, e.g., PopulateIndexNode and
>>    ReverseVNode, etc. This refers from:
>>    https://mail.openjdk.org/pipermail/panama-dev/2020-April/008484.html
>>    
>>    After merging these nodes, code will be easier to maintain. E.g.,
>>    matching rules can be simplified.
>>    
>>    Besides AArch64, this patch tries to keep other ad files as the same
>>    before, only supplies some necessary predicate. E.g., for matching rules
>>    using ReplicateB before, they are now matching Replicate with a new
>>    predicate "Matcher::vector_element_basic_type(n) == T_BYTE". This would
>>    be easy for review and lower risks.
>>    
>>    [TEST]
>>    x86:     Tested with option "-XX:UseAVX=0/1/2/3".
>>    AArch64: Tested on SVE machine and Neon machine.
>>    
>>    Full jtreg passed without new issue.
>>    
>>    Change-Id: Icb16084e3909177ab302f0ba33c2216c860a5da6
>
> src/hotspot/share/opto/matcher.cpp line 2800:
> 
>> 2798:   BasicType bt = vector_element_basic_type(n);
>> 2799:   assert(bt != T_CHAR, "char is not allowed in vector");
>> 2800:   return is_subword_type(bt) || bt == T_INT;
> 
> I suggest you create a `is_non_long_integral_vector` in `src/hotspot/share/utilities/globalDefinitions.hpp`, next to `is_integral_type`.

This method is used only for matcher.  We can do this if  there were more usages of `is_subword_type(bt) || bt == T_INT`.

> test/hotspot/jtreg/compiler/vectorization/runner/ArrayInvariantFillTest.java line 75:
> 
>> 73:     @IR(applyIfCPUFeatureOr = {"asimd", "true", "sse2", "true"},
>> 74:         applyIf = {"OptimizeFill", "false"},
>> 75:         counts = {IRNode.REPLICATE, ">0"})
> 
> This may be a good time to try and convert `IRNode.REPLICATE` to a `vectorNode`. That would ensure that we still can use the type of the replicate in this test here.

Thanks, fixed.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/14830#discussion_r1377342505
PR Review Comment: https://git.openjdk.org/jdk/pull/14830#discussion_r1377343456


More information about the hotspot-compiler-dev mailing list