upstream @6b7fb59               upstream @HEAD
				no patches                      with your PR
				RAM [*], fps, view range        RAM [*], fps, range
----------------------------------------------------------------------
 no shaders     10 GB, 36 fps, 105.6m			10 GB, 35 fps, 104.3m
no aniso/mip

 shaders only   9.9 GB, 30 fps, 98.9m			10 GB, 28 fps, 98.1m
(no aniso,
 no mipmap,
 no bumps,
 no para)

shaders		8.3GB, 30 fps, 114m				N/A
+aniso
+mipmap

shaders		13 GB, 32 fps, 90.6m			14 GB, 28 fps, 93.0m
+bumpmaps

shaders		12 GB, 17 fps, 35m				N/A
 +bumpmaps
 +parallax
 +aniso
 +mipmap

shaders		14 GB, 30 fps, 89.7m			14 GB, 26 fps, 89.5m
 +bumpmaps
 +parallax
 +parallax_
  occlusion_
  iterations
   = default

shaders		N/A								14 GB, 39 fps, 74.7m
 +bumpmaps
 +parallax
 +parallax_
  occlusion_
  iterations=1

FPS and view range values taken at their apparent best values.

Using smooth lighting, particles enabled, 3d clouds, fancy trees,
connected glass.  All other settings turned off or set to "no"

No mipmap or anisotropic or other filtering was used in any test.

maximized window at 1600x1200 on one of my two screens,
AMD R9 280x video with fglrx driver
Using Dreambuilder, on my usual test map.
HDX 512px texture pack with normalmaps, total of 5113 texture files

To ensure each test is accurate and records the exactly-chosen modes,
Minetest was closed completely after changing settings and then
restarted to measure.

Only the map in front of me and perhapos 20 degrees to each sides was ever made
visible.

Screenshot of the area under test, taken with all-view-range turned on only
for the several seconds needed to take it, so one can see ALL of the terrain
that could theoretically appear:

http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/random/Screenshot%20-%2006032015%20-%2003%3a17%3a22%20PM.png


Map was allowed to load and settle for perhaps 30 seconds on each run before
adding the figures to the chart.


[*] RAM values measured with top(1), from the resident/"RES" column.