On Thu, Apr 19, 2012 at 8:30 PM, Paul Berry <span dir="ltr"><<a href="mailto:stereotype441@gmail.com">stereotype441@gmail.com</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im">On 19 April 2012 06:41, Anuj Phogat <span dir="ltr"><<a href="mailto:anuj.phogat@gmail.com" target="_blank">anuj.phogat@gmail.com</a>></span> wrote:<br></div><div class="gmail_quote"><div class="im">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div>On Thu, Apr 19, 2012 at 12:09 AM, Paul Berry <span dir="ltr"><<a href="mailto:stereotype441@gmail.com" target="_blank">stereotype441@gmail.com</a>></span> wrote:<br></div><div class="gmail_quote"><div>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
This is the test I've been using to prototype MSAA on the i965<br>
driver. Unlike the existing EXT_framebuffer_multisample<br>
tests (which so far mostly check corner cases at the API level),<br>
this test verifies that correct pixels are rendered when MSAA is<br>
enabled. It's a bit long for a typical piglit test, since in<br>
order to verify that the MSAA color, depth, and stencil buffers<br>
are all configured correctly, it has to render some fairly<br>
complex scenes.<br>
---<br>
<br>
This test partially overlaps the functionality of the test Anuj<br>
sent to the list on Monday ([PATCH] Add test to switch on/off<br>
MSAA in FBO). Anuj: I'm interested in hearing how my test<br>
performs on your AMD system.</blockquote><div><br></div></div><div>Here is what I get on my AMD system:</div><div>./bin/ext_framebuffer_multisample-accuracy 4 color<div>Framebuffer not complete</div><div>PIGLIT: { 'result' : 'skip' }</div>
</div><div>similar error with other test_type options.</div><div><br></div></div></blockquote></div><div><br>Do you also get this result with the "depthstencil" flag? I added that flag so that the test would work on my nVidia test machine, which doesn't seem to support separating the depth and stencil buffers.<br>
</div></div></blockquote><div><div><br></div><div>Now I don't get "Framebuffer not complete"</div><div>$ ./bin/ext_framebuffer_multisample-accuracy 4 color depthstencil -auto<br>Pixels that should be unlit<br>
count = 214644<br> Perfect output<br>Pixels that should be totally lit<br> count = 28968<br> Perfect output<br>Pixels that should be partially lit<br> count = 18532<br> RMS error = 0.260300<br>The error threshold for this test is 0.119880<br>
PIGLIT: {'result': 'fail' }</div></div><div><br></div><div>I get similar RMS error value for all other cases. </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="gmail_quote"><div> </div><div class="im"><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div class="gmail_quote"><div></div><div>No such error is observed on SandyBridge. But MSAA was absent as expected.</div>
<div><br></div><div>I couldn't get MSAA working on AMD with my test case as well.</div>
<div>I've installed ATI's catalyst drivers (Version: 12.2) on Ubuntu 10.04</div><div>Couldn't successfully install catalyst drivers on latest Ubuntu version: 11.10.</div><div><br></div><div>Executing fglrxinfo outputs:</div>
<div>display: :0 screen: 0</div><div>OpenGL vendor string: Advanced Micro Devices, Inc.</div><div>OpenGL renderer string: AMD Radeon(TM) HD 6480G</div><div>OpenGL version string: 4.2.11554 Compatibility Profile Context</div>
<div><br></div><div>I'll try updating my system and see if that helps.</div></div></blockquote></div><div><br>I'd also be curious to see the output of glxinfo on your system--this will show which framebuffer configurations the driver claims to support MSAA with.</div>
</div></blockquote><div> Here is my glxinfo output: <a href="http://pastebin.com/7YeVMDC2">http://pastebin.com/7YeVMDC2</a></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="gmail_quote"><div class="im"><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div class="gmail_quote"><div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Everyone: I'm interested in hearing<br>
opinions about whether this test is too big, and if so, how to<br>
effectively split it up into smaller tests.<br></blockquote><div><br></div></div><div>This is a fairly comprehensive test to measure the accuracy of MSAA in FBO.</div><div>I liked the way you are generating reference image. In my opinion there are lot</div>
<div>of functions in this test which could be utilized by upcoming multisample test </div><div>cases. e.g. the current piglit test I'm working on (Test switching MSAA on/off in</div><div>FBO), can utilize draw_reference_image() and measure_accuracy() functions to</div>
<div>verify if MSAA is actually switched ON in FBO. Every new multisample test case</div><div>would require these functions to verify if MSAA is happening as expected.</div><div>So, it would be helpful to split it in to two files: common.cpp, accuracy.cpp</div>
</div></blockquote></div><div><br>That's a good thought. I'll try to figure out a sensible way to factor out the reusable stuff into a separate file.<br></div></div>
</blockquote></div><br>