Nodes drawing artifacts with Java 8 on Intel HD Graphics hardware

[Don-vip]

I am currently on a very low-end Windows laptop (Acer Extensa 2509 series, Intel Celeron N2930 1.83 GHz, 4 Gb RAM, Windows 8.1 x64), and face this ugly graphical bug:

This is clearly a Java bug: it works fine with Java 7 (7u80) but always fail with Java 8 (8u45).

It seems the Intel drivers cause many problems with Java 8. I'd like to see if we can detect & workaround the problem, as it would be a blocker for Java 8 migration.

The hardware involved here is an Intel HD graphics chip (2013), 0F31:
PCI\VEN_8086&DEV_0F31&SUBSYS_09051025&REV_0E\3&11583659&0&10

Driver is the default version: 10.18.10.3496 of Intel (March 2014).

Adding the JVM parameter -Dsun.java2d.d3d=false does not solve the problem.

[Don-vip]

Reported as ​https://bugs.openjdk.java.net/browse/JI-9021456

[mkoniecz]

On Java bug tracker:

To verify this, could you please supply -Dsun.java2d.trace=log option to the test program's JVM, and collect
the console output. This will give us a hint which pipeline (software or D3D) is actually used.

We need additional information as requested above. Therefore, making the bug status to Resolved/Incomplete for now.

[Don-vip]

I sent to Oracle this answer today:

I updated Java to latest version (8u51) and am still able to reproduce the problem.
I reproduced it in all these cases:

1. -Dsun.java2d.d3d=false -Dsun.java2d.trace=log
2. J2D_TRACE_LEVEL=4
3. -Dsun.java2d.opengl=True

The problem is the same in all cases. Please find attached the console output of each one of them. I think this output is worth noting:

[E] D3DPPLM::CheckForBadHardware: found matching hardware: VendorId=0x8086 DeviceId=0xffffffff
[E] D3DPPLM::CheckForBadHardware: bad hardware found, device disabled
[E] D3DPPLM::GDICheckForBadHardware: no suitable devices found

[asdqweasd]

I had this bug in 2015 on Windows 8. Since then I have installed Ubuntu 16.10 and the bug still reproduces on the same laptop. So it is definitely a hardware related problem in Java.
Processor: Intel® Celeron(R) CPU N2830 @ 2.16GHz × 2
Graphics: Intel® Bay Trail

Using JOSM is tough when this bug occurs. If it happens, you can try to download more OSM data or hide/delete GPX traces. Usually after some tweaking the bug disappears, but only for the current session.