[Bug 82463] New: i965: Using EWA algorithm for isotropic case

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=82463

          Priority: medium
            Bug ID: 82463
          Assignee: idr at freedesktop.org
           Summary: i965: Using EWA algorithm for isotropic case
        QA Contact: intel-3d-bugs at lists.freedesktop.org
          Severity: normal
    Classification: Unclassified
                OS: All
          Reporter: pavel.e.popov at intel.com
          Hardware: x86-64 (AMD64)
            Status: NEW
           Version: 10.2
         Component: Drivers/DRI/i965
           Product: Mesa

Tests "MesaDemos tests/texfilt" and Test "MesaDemos tests/mipmap_tunnel"
started to show an expected results (the transition from one mipmap level to
another should be nice and regular/circular) for anisotropic case (2x - 16x)
after EWA algorithm was enabled in Mesa: 
http://cgit.freedesktop.org/mesa/mesa/commit/?id=f74cf5f80ea6ea16abb38c28c2b4826b8559017b

But for isotropic case (1x) something like a polar rose is still observed. The
screenshots are attached below. This is because EWA algorithm was enabled in
Mesa only for anisotropic case. 

But EWA algorithm also can be enabled for isotropic case. At least Haswell
programming notes (Volume 2d: Command Reference: Structures(Haswell), Page 271)
contains the following lines:
"When EWA used for non-anisotropic filtering and the coordinates have zero
derivative the computed LOD is 0 instead of -inf."

I enabled EWA algorithm for isotropic case too and expected results were
observed for both tests. The screenshots are attached below.

Also I run these tests on Nvidia and observed that they also work as expected.
The screenshots are attached below.

Could we use EWA for isotropic case? Or just say that this polar rose is ok for
isotropic case and performance is more important for us here.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20140811/7b9d09b9/attachment-0001.html>