gpu: cache transformed bounds

To avoid duplicate work when using macros and non-offset transforms,
cache also the new bounding boxes set up for them. The ops.Reader
already generates Keys for all operations, so use them in the cache.

Signed-off-by: Viktor <viktor.ogeman@gmail.com>

internal/ops/reader.go

@@ -54,10 +54,6 @@ type opMacroDef struct {
 	endpc pc
 }
 
-func (r *Reader) NewKey(pc int) Key {
-	return Key{ops: r.ops, pc: pc, version: r.ops.Version()}
-}
-
 // Reset start reading from the op list.
 func (r *Reader) Reset(ops *op.Ops) {
 	r.stack = r.stack[:0]