V8 Typed Array

There are 12 typed array supported. They are as follows.

Type	Value Range	Size in bytes	Java Type	Web IDL type
Int8Array	-128 to 127	1	byte	byte
Uint8Array	0 to 255	1	byte	octet
Uint8ClampedArray	0 to 255	1	byte	octet
Int16Array	-32768 to 32767	2	short	short
Uint16Array	0 to 65535	2	short	unsigned short
Int32Array	-2147483648 to 2147483647	4	int	long
Uint32Array	0 to 4294967295	4	int	unsigned long
Float16Array	-65504 to 65504	2	short	N/A
Float32Array	-3.4e38 to 3.4e38	4	int	unrestricted float
Float64Array	-1.8e308 to 1.8e308	8	double	unrestricted double
BigInt64Array	-2^63 to 2^63 - 1	8	long	bigint
BigUint64Array	0 to 2^64 - 1	8	long	bigint

Play with V8ValueTypedArray

All typed array share the same V8 value type V8ValueTypedArray. They can be differentiated by their getType() which is an enum indicating their actual types. There is a buffer associated with the V8ValueTypedArray. That buffer is the actual shared memory between JVM and V8. In other words, there is zero memory copy between JVM and V8 if V8ValueTypedArray is used. Many performance sensitive applications usually exchange data via V8ValueTypedArray. There are a set of from**** and to**** API for the data exchange.

try (V8ValueTypedArray v8ValueTypedArray = v8Runtime.createV8ValueTypedArray(
    V8ValueReferenceType.Int8Array, 4)) {
    assertEquals(4, v8ValueTypedArray.getLength());
    assertEquals(1, v8ValueTypedArray.getSizeInBytes());
    assertEquals(4, v8ValueTypedArray.getByteLength());
    assertEquals(0, v8ValueTypedArray.getByteOffset());
    assertEquals(V8ValueReferenceType.Int8Array, v8ValueTypedArray.getType());
    try (V8ValueArrayBuffer v8ValueArrayBuffer = v8ValueTypedArray.getBuffer()) {
        v8ValueArrayBuffer.fromBytes(new byte[]{ (byte) 1, (byte) 2, (byte) 3, (byte) 4,});
    }
}

Play with Float16Array

Float16Array was introduced to V8 in Mar, 2024 as an implementation of TC39 proposal-float16array. It's quite special because there is no float 16 in Java, so has to be mapped to short.

Javet borrows Float16 from the Android FP16 implementation (copyright preserved) to allow easy translation between short and float. Just call short toHalf(float f) or float toFloat(short h) to complete the translation.

Float16Array is not enabled by default. Please make sure the following code is executed before the first NodeRuntime or V8Runtime is created.

// Node.js mode
NodeRuntimeOptions.NODE_FLAGS.setJsFloat16Array(true);

// V8 mode
V8RuntimeOptions.V8_FLAGS.setJsFloat16Array(true);

Please review the test cases for more detail.