Updating texture paths with the string_replace operator


Arnold 6.0.2 includes a new string_replace that lets you do find and replace on string parameters.

For example, I got an ass procedural file from a user, but I didn’t get the textures. Normally I would just enable ignore textures, but this time I used string_replace to replace all textures with my trusty-old-Softimage noicon pic.

string_replace operator in MtoA (Maya)
  • Selection selects all Arnold image nodes. So this string replace operation applies to all string parameters of the image node (which includes name, uvset, color_space as well as filename).
  • Match is a regular expression that matches anything that looks like C:/project/sourceimages/example.tx or X:/temp/test/example.tx
    • [A-Z] matches any drive letter from A to Z
    • .* matches any string of characters
    • Note that match is a regular expression, so you cannot use glob wildcards like this: C:\sourceimages\*.tx (because * is not any string of characters, it is zero or more occurrences of the previous character)
    • \.tx matches a period followed by “tx”. The period has to be escaped with a backslash because ‘.’ matches any single character
    • Note that this match expression won’t handle something like example.txtra.tx
  • Replace is the string that replaces anything matched by the Match expression.

How to check what mip levels are used


You can use the Arnold log to check what mip levels are used. In the Image file statistics, there’ll be a MIP-COUNT for every texture

00:00:02 124MB | Image file statistics:
00:00:02 124MB |       opens tiles MB read --redundant-- I/O time res             File
00:00:02 124MB |     1   1   538   12.6                  2.2s     8192x8192x3.f16 debug.tx MIP-COUNT[0,13,314,147,40,15,4,1,1,1,1,0,0,1]
00:00:02 124MB |

For example, with my debug texture, I see

MIP-COUNT[0,13,314,147,40,15,4,1,1,1,1,0,0,1]

in the Arnold log for this render:

 

The case of the 15% rendering utilization


Or, “Understanding the Arnold log, part 23”

In this case, a client had very low (15%) CPU usage for a render. We got the Arnold log, and here’s the interesting part:

00:01:44  856MB   | OpenImageIO ImageCache statistics (0000000025850F20) ver 1.5.24
00:01:44  856MB   |   Images : 3 unique
00:01:44  856MB   |     ImageInputs : 299 created, 2 current, 3 peak
00:01:44  856MB   |     Total size of all images referenced : 192.4 MB
00:01:44  856MB   |     Read from disk : 12.0 GB
00:01:44  856MB   |     File I/O time : 45m 41.5s (48.1s average per thread)
00:01:44  856MB   |     File open time only : 0.0s
00:01:44  856MB   |   Tiles: 711523 created, 559 current, 640 peak
00:01:44  856MB   |     total tile requests : 249379431
00:01:44  856MB   |     micro-cache misses : 3353694 (1.34482%)
00:01:44  856MB   |     main cache misses : 711523 (0.285317%)
00:01:44  856MB   |     Peak cache memory : 10.0 MB
00:01:44  856MB   |   1 not tiled, 1 not MIP-mapped
00:01:44  856MB   | -----------------------------------------------------------------------------------
00:01:44  856MB   | performance warnings:
00:01:44  856MB   | Rendering utilization was only 15%. Your render may be bound by a single threaded process or I/O.
00:01:44  856MB   | -----------------------------------------------------------------------------------
  • File I/O time seems a bit high for three textures and a render that took less than 2 minutes
  • main cache* misses is pretty high. Normally you expect something less than 0.01%.

    0.285% means that texture tiles are loaded from disk (instead of from the in-memory texture cache) once out of every 350 texture lookups. That’s a little high.

    * The main cache is the cache of 64×64 texture tiles loaded from disk into the texture cache.

  • Peak cache memory is 10.0 MB !!! That explains the main cache misses: the texture cache is really, really small.

Other clues to the too-small texture cache size:

  • Read from disk is 12 GB but the total size of all images referenced is just 192.4 MB, and the peak cache memory was just 10.0 MB
    So the same texture data is constantly being unloaded from the cache and reloaded from disk.

The solution? Increase the size of the texture cache. The current default is 2048, which should be good in most cases.

Convert texture maps to .tx


When you’re rendering with Arnold, convert your texture maps to .tx files with maketx. Always 🙂

.tx textures are better because they are:

  • Tiled (usually the tiles are 64×64 pixels).

    Arnold loads one tile at a time, as needed, rather than loading the entire texture map in memory. So textures are loaded faster, because Arnold loads only what it needs to render the image.

    Arnold uses a texture cache system to efficiently manage texture memory usage: only the most recently used tiles are kept in memory. You can have hundreds, or even thousands, of 4k and 8k images, and Arnold will never use more runtime memory than the user-specified max cache size.

  • Mip-mapped.

    Mip-mapped textures are anti-aliased, even at low AA sample settings.

If you don’t convert your texture maps with maketx, the default Arnold behavior is to auto-tile and auto-map your textures for you, but this is very inefficient because it is done once for every rendered frame, rather than one-time-only with maketx.

Texture search paths


With Arnold there are two things in an ASS file that specify the location of a texture file:

  • The texture filename, which can be an absolute or relative path.
  • The texture_searchpath in the render options. The texture_searchpath is used only if the texture filename is a relative path (for example: “sourceimages/example.tiff”.
  • For example, if the texture filename was a relative path, you might have something like this in an ASS file:

    options
    {
     ...
     texture_searchpath "//server/projects/test"
     ...
    }
    
    MayaFile
    {
     ...
     filename "sourceimages/noicon.png"
     ...
    }