On Fri, 27 May 2022 18:12:59 GMT, Alexander Zuev <kiz...@openjdk.org> wrote:
> On 10000 iterations the difference is measurable in tens of milliseconds
> between test runs. On 100000 iterations test execution with BICUBIC
> approximation is noticeable (like test run takes more than a second longer).
Not an ideal benchmark but it shows that the bicubic is ~200 times slower.
public static void main(String[] args) {
var gc =
GraphicsEnvironment.getLocalGraphicsEnvironment()
.getDefaultScreenDevice()
.getDefaultConfiguration();
var vi = gc.createCompatibleVolatileImage(50, 50,
Transparency.TRANSLUCENT);
var bi = new BufferedImage(50,50,BufferedImage.TYPE_INT_ARGB);
Graphics2D g2d = vi.createGraphics();
g2d.scale(2,2);
test(bi, g2d);// warmup
test(bi, g2d);
g2d.setRenderingHint(RenderingHints.KEY_INTERPOLATION,
RenderingHints.VALUE_INTERPOLATION_BICUBIC);
test(bi, g2d);// warmup
test(bi, g2d);
}
private static void test(BufferedImage bi, Graphics2D g2d) {
long start = System.nanoTime();
for (int i = 0; i < 1000; ++i) {
g2d.drawImage(bi, null, 0, 0);
}
long time = System.nanoTime() - start;
System.out.println("Time: " + time);
}
Time: 21313400
Time: 4036200
Time: 1140613100
Time: 839499200
-------------
PR: https://git.openjdk.java.net/jdk/pull/7805
