OpenGL range_for support (stage 2.1.1) (#594)

* [skip ci] range for using GLSL threads

range for and extptr (TODO: extptr not work)

[skip ci] remove ext ptr fake support

 Conflicts:
	examples/opengl_backend.py
	taichi/codegen/codegen_opengl.cpp
	taichi/platform/opengl/opengl_kernel.cpp
	taichi/platform/opengl/opengl_kernel.h

 Conflicts:
	taichi/program/program.cpp

* [skip ci] fix typo in _thread_id_ > (should be >=)

 Conflicts:
	taichi/codegen/codegen_opengl.cpp

 Conflicts:
	taichi/codegen/codegen_opengl.cpp

* multi work groups & compile before launch to save time

 Conflicts:
	taichi/codegen/codegen_opengl.cpp
	taichi/platform/opengl/opengl_api.cpp
	taichi/platform/opengl/opengl_api.h
	taichi/platform/opengl/opengl_data_types.h

fix merge

 Conflicts:
	taichi/codegen/codegen_opengl.cpp

* [skip ci] non-GL build fix

examples/opengl_range_for.py

@@ -0,0 +1,14 @@
+import taichi as ti
+
+ti.init(arch=ti.opengl)
+
+x = ti.var(ti.i32, shape=(5, 5))
+
+@ti.kernel
+def func():
+  for i in range(5):
+    for j in range(5):
+      x[i, j] = i + j
+
+func()
+print(x[2, 3])