General user's guide --
https://drupal.star.bnl.gov/STAR/comp/simu/geometry0

Scrolling most of the way down, we get to some evaluation documents.
These look like they were an intermediate
report... some small questions were still open at the time.  I believe
we tracked these issues down to a small
overlap in the prompt hit region, and an error in AgSTAR description
(not reproduced in AgML) with respect to
the support rods.

   Geometry differentials relevant for tracking--

https://drupal.star.bnl.gov/STAR/comp/simu/geometry0/agml-vs-agstar-comparison

   Hit by hit comparisons for simulation--

https://drupal.star.bnl.gov/STAR/blog/jwebb/2011/jul/13/hit-hit-comparison-between-agml-and-agstar-geometries-tag-y2009c

   Track reconstruction comparison--

https://drupal.star.bnl.gov/STAR/system/files/AgML%20vs%20AGI%20Sti%20Tracks%20y2010%20geometry.pdf

The PDF below shows detailed comparison between AgML --> ROOT and AgSTAr
--> ROOT for the y2011 geometry.
    http://www.star.bnl.gov/~jwebb/AgML_Status/

    The PMD never did convert right.  (And actually, I suspect that AgML
produces the correct version of the
    detector, as I vaguely recall odd overlaps in the AgSTAR model which
were not produced by AgML).

I also presented a "readiness" document in November 2011--

https://drupal.star.bnl.gov/STAR/system/files/AgML%20Project%20Overview.pdf

    but I don't see any material budgets there.

Looking back at my blogs, I see "Milestone 1", claiming consistency
y2009 geometry to present--

https://drupal.star.bnl.gov/STAR/blog/jwebb/2011/mar/09/agml-release-milestone-1

Release "Milestone 2", claiming good agreement between *latest* y2005
geometry to present:

https://drupal.star.bnl.gov/STAR/blog/jwebb/2011/apr/28/agml-release-milestone-2