TENTATIVELY MARKED FOR DEATH – THE ISSUES DESCRIBED MAY VERY WELL HAVE BEEN PURELY SIDE EFFECTS OF GRID-WIDE CONNECTIVITY ISSUES STARTING ON 4 APR EXTENDING INTO 6 APR. I finally installed what I believe were the latest ATI drivers (8.3? including OpenGL Version: 2.1.7412 Release) after 1.19.1(4) became the release viewer early on 4 April 2008. I've filed this as affecting the 1.19.1.4 release since many other residents seemed to report similar problems as soon as 1.19.1(4) was being installed by those unwilling to participate in the RC process. If there is a material difference between the 1.19.1(4) release candidate (which I was using when these behaviors were seen) please edit/update this issue as most appropriate. For clarity's sake, I had been using 1.19.1(4) as a release candidate since very shortly after its release on Saturday, 29 Mar 08. (http://blog.secondlife.com/2008/03/29/new-release-candidate-viewer-1191-rc4-available/) For most of the time I've used this RC, and even in some of its earlier builds (from 1.19.1(2) onward let's say), especially preceding the roll-out of the new grid server code, this RC was remarkably stable under the 7.11 drivers I was using (though I did not try to turn VBO during that time). Syrupy response, yes, but more stable than other RCs or even Release versions in recent memory. I had put off updating the ATI drivers for several months as the Jira reports, especially concerning the VBO artifacts, left much confusion as to whether any newer drivers were actually addressing VBO vertex artifacts seen and documented under VWR-3258 under issues linked as related. Discussion seemed to indicate that several updates to drivers were not addressing the problems – and, fwiw, I'd seen what seemed to be many related instability issues with the few other programs I use that rely on the OpenGL rendering system, most notably Starry Night, a sort of planetarium and telescope control program, which rarely runs more that a few hours with OpenGL enabled without a sudden and inconvenient crash, at least in the outdated versions I have. I can state with reasonable certainly, reading back over these issues, especially the reporting on VWR-3258, that the ATI drivers I was using up until 4 Apr 08 were version 7.11. Rationale for New Issue: I'm filing this as a new issue, though I suspect it's related to several others, because, for one thing, I have yet to see the VBO artifacts themselves reappear so far while VBO has been turned on in this viewer, while running the v8.3 ATI driver set. All potentially related issues that I could spot (and that remain active issues as of this reporting) have been linked. Observed: Two crashes in fairly quick succession came about when I enabled VBO rendering in this viewer at around 1100hrs PDT, 4 Apr 08. Average time in world before crashing was about 20-30 minutes (estimated). However, since installing the 8.3 Catalyst ATI driver package, I have yet to see the vertex artifacts that I and other residents have documented in detail, (and that many residents have mentioned seeing in world) under other ATI drivers. In past reports and in my few follow-up tests under the older drivers, using any of the recent viewers, particularly releases since WindLight First Look viewers became prevalent, vertex artifacts usually appeared in my viewer almost as soon as I enabled OpenGL VBO. Subsequent (third, fourth, fifth?) logins with VBO disabled allowed me to remain connected to the grid with only one further crash (and while also streaming audio to in world parcels via the WinAmp SHOUTcast plug-in) for about 5 uninterrupted hours. That crash took place some time after I had stopped streaming and had unloaded WinAmp and related programs. The single crash came at around 6 hours in world, and I was able to reconnect to the grid almost immediately, as the crash halted the SLCandidate process almost instantly, allowing a quick relog. The following session lasted around 3 to 5 hours and I was probably one of the last ones still connected to the grid when I logged out without incident at around 12:10am, PDT/SLT. Shortly thereafter the grid was shut down entirely to resolve top level routing issues mentioned in http://blog.secondlife.com/2008/04/04/logins-disabled-until-further-notice/ I have not tested in the 1.19.1(4) release itself, partly due to lack of time, and also working under the assumption there are no material differences between it and the RC I already have installed – and also selfishly wanting to keep 1.19.0.5 as a fall-back viewer should something make 1.19.1.4 untenable. And of course, given grid problems of Fri. 4 Apr 08, this may all merely be wildly hopeful speculation on my part that some part of the seemingly widespread crashiness in 1.19.1.4 is due to either outdated ATI drivers, or some persistent bug that only gets nasty when VBO and OpenGL are being used with the most current available ATI drivers.