[aarch64-port-dev ] RFR(M): 8213134: AArch64: vector shift failed with MaxVectorSize=8

[Yang Zhang (Arm Technology China)]

-----Original Message-----
From: Andrew Haley <aph at redhat.com> 
Sent: Tuesday, November 13, 2018 5:11 PM
To: Yang Zhang (Arm Technology China) <Yang.Zhang at arm.com>
Cc: aarch64-port-dev at openjdk.java.net; nd <nd at arm.com>
Subject: Re: [aarch64-port-dev ] RFR(M): 8213134: AArch64: vector shift failed with MaxVectorSize=8

On 11/13/2018 01:27 AM, Yang Zhang (Arm Technology China) wrote:
>> Our test covers vector shift with imm, vector shift with vector(dump from scaler), vector shift with vector(from loading, this format isn’t supported by middle end). But we don’t test option MaxVectorSize=8.

>Sorry, I didn't see the test. Where did you put it?

It's my fault. I didn't explain it clearly.
Hotspt jtreg contains the following test cases that can cover all the patterns of vector shift.  But option "-XX:MaxVectorSize=8"(vecD 64-bit) isn't tested, so that only "-XX:MaxVectorSize=16"(vecX 128-bit, default) is tested.
compiler/c2/cr6340864/TestByteVect.java
compiler/c2/cr6340864/TestIntVect.java
compiler/c2/cr6340864/TestShortVect.java
compiler/codegen/TestCharVect2.java

Do I need to add test option "-XX:MaxVectorSize=8" to jtreg test cases in my patch?

Regards
Yang

-- 
Andrew Haley
Java Platform Lead Engineer
Red Hat UK Ltd. <https://www.redhat.com>
EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671