Maybe you could do a fold with  a 'filler' that fills multiple elements at a time to reduce the depth.

e.g. declare a bunch of:

void fillArray(Object[] arr, int startPos, Object a0,  Object a1, ...) {
    arr[startPos + 0] = a0;
    arr[startPos + 1] = a1;
    ...
}

With differing arities, and then use a series of collectArguments calls to collect the arguments into the final array in chunks (of e.g. size 10). Merge all the array parameters together using permuteArguments, and then finally stick an array constructor in that parameter slot using collectArguments to create the array.

See here a crude diagram:

result( Object,Object,Object,       Object,Object,Object)void
                                                   |      | |             |      |      |                  arity                             |      | |             |      |      |                    |                               |      | |             |      |      |                    V                               |      | |             |      |      |   Object[]::<init>(int)Object[]                    | |      |             |      |      |                         |            startPos=0    |      | |             |      |      |         |----------|------------|        |         |      | |             |      |      |         |          |            V        V         V      V V             |      |      |         |          | fillArray(Object[],int, Object,Object,Object)void     |      |      |         | |                                                   | |      |      |         |          | startPos=3                        |       |      |      |         |          |------------| |                             |       |      |      |         |                       V V                             |       V      V      V         | fillArray(Object[],int,                           | Object,Object,Object)void
| |                           |
        |       |---------- collectArguments --------------------------<---------------------------<
        V       V
target(Object[]  )void

Where Object[] and Object can be replaced with sharper types.

Although that would require writing a bunch of hard-coded fillers, one for each primitive type and then one for reference types X the chunk size you want to support.

HTH,
Jorn

On 02/04/2021 10:00, Charles Oliver Nutter wrote:
First attempt at a workaround seems to be a wash. I rolled back to my
older logic (that does not use a hand-crafted collector method) to
come up with a pure-MethodHandle workaround for asCollector. I came up
with this (using InvokeBinder):

```
MethodHandle constructArray = Binder.from(arrayType, Object[].class)
         .fold(MethodHandles.arrayLength(Object[].class))
         .dropLast()
         .newArray();

MethodHandle transmuteArray = Binder.from(arrayType, Object[].class)
         .fold(constructArray)
         .appendInts(0, 0, count)
         .permute(1, 2, 0, 3, 4)
         .cast(ARRAYCOPY.type().changeReturnType(arrayType))
         .fold(ARRAYCOPY)
         .permute(2)
         .cast(arrayType, arrayType)
         .identity();

MethodHandle collector = transmuteArray.asCollector(Object[].class,
count).asType(source.dropParameterTypes(0,
index).changeReturnType(arrayType));

return MethodHandles.collectArguments(target, index, collector);
```

Hopefully this is mostly readable. Basically I craft a chain of
handles that uses the normal Object[] collector and then simulates
what the pre-Jorn asCollector does: allocate the actual array we want
and arraycopy everything over. I figured this would be worth a try
since Jorn's comments on the PR hinted at the intermediate Object[]
going away for some collect forms. Unfortunately, reproducing the old
asCollector using MethodHandles does not appear to work any better...
or at least it still pales compared to a collector function.

I am open to suggestions because my next attempt will probably be to
chain a series of folds together that populate the target array
directly, but it will be array.length deep. Not ideal and not a good
general solution.

On Thu, Apr 1, 2021 at 6:44 PM Charles Oliver Nutter
<head...@headius.com> wrote:
Very nice! I will have a look at the pull request and perhaps it will lead me 
to a short-term work around as well.

On Thu, Apr 1, 2021, 12:04 Jorn Vernee <jorn.ver...@oracle.com> wrote:
Hi Charlie,

(Sorry for replying out of line like this, but I'm not currently
subscribed to the mlvm-dev mailing list, so I could not reply to your
earlier email thread directly.)

I have fixed the performance issue with asCollector you reported [1],
and with the patch the performance should be the same/similar for any
array type (as well as fixing a related issue with collectors that take
more than 10 arguments). The patch is out for review here:
https://urldefense.com/v3/__https://github.com/openjdk/jdk/pull/3306__;!!GqivPVa7Brio!NeZJUPlqKG3_Z1yUBVS5cX1Gtz0pQMuFcTiR0PMhz42KSgmJpMNj8zkru_YUYkM-$

Cheers,
Jorn

[1] : https://bugs.openjdk.java.net/browse/JDK-8264288

_______________________________________________
mlvm-dev mailing list
mlvm-dev@openjdk.java.net
https://mail.openjdk.java.net/mailman/listinfo/mlvm-dev

Reply via email to