Commit Graph

82 Commits (def46d24c2a4e226054b7c19fb9acdd17bb040fd)

Author SHA1 Message Date
Nicky 8c91463476 FIRE-12968; Even though the root cause is likely the shader memory overrun, preseed the VBO pools to take some strain from the driver. Remove the LLStrider changes while at it, those are not used in the Havok version and have not triggered anything in the OS version since month. 2014-03-10 00:45:41 +01:00
Tank_Master 33922af1fa Merge LL release 2013-11-01 19:52:14 -07:00
Tank_Master 0a1a351b86 Merge Viewer-Share 2013-10-21 23:51:34 -07:00
Tank_Master 5e2324b5d3 Backed out changeset: f632d5a3a4e3 - has unwanted side effects 2013-10-17 23:46:34 -07:00
Dave Parks 0e116741a0 MAINT-2968 Fix for crash on AMD hardware running current drivers. 2013-10-10 18:00:11 -05:00
Dave Parks ecbdb4948f MAINT-2968 Fix for crash on AMD hardware running current drivers. 2013-10-10 18:00:11 -05:00
Graham Linden 95e34d86b9 The unbearable lightness of being norspec 2013-09-18 18:24:16 -07:00
Tank_Master 05ad3a1c29 Merge Viewer-bear (merge of materials + fixes) to FS-exp 2013-09-21 11:44:46 -07:00
Tank_Master 52d9b07503 Merge viewer-bear 2013-09-12 21:15:34 -07:00
Dave Parks 79029149a4 Automated merge with file:///F:%5Cviewer-bear 2013-06-21 12:51:29 -05:00
Dave Parks c3f14b915c NORSPEC-229 Fix for bad binormals on mirrored surfaces (use tangent calculator instead of binormal calculator, convert binormal centric code to tangent centric) 2013-06-03 12:50:48 -05:00
simon ee2fce8790 Merge downstream code and viewer-beta 2013-05-09 14:10:45 -07:00
Graham Madarasz fd447189c5 Merge 3.5.1 into Materials 2013-04-25 21:00:16 -07:00
Dave Parks fc4869fb78 NORSPEC-65 Hook up texture offset/scale/rotation for normal and specular maps. 2013-03-29 16:13:36 -05:00
Graham Madarasz bf6182daa8 Update Mac and Windows breakpad builds to latest 2013-03-29 07:50:08 -07:00
Dave Parks 609ed855e1 MAINT-2371 More optimizations.
Reviewed by Graham
2013-03-04 18:01:42 -06:00
Dave Parks f5e5396c3a MAINT-2371 First set of profile guided optimizations.
Reviewed by Graham
2013-02-26 15:15:08 -06:00
Tank_Master 718f6fa0ff merge LL 3.6.1 2013-07-10 00:23:23 -07:00
Tank_Master d4640277d0 Merge LL 3.5.3 release. 2013-07-09 16:24:58 -07:00
Nicky 2b620b006b Disable caching of vbo memory in LLVBOPool. Had been there once and it was a memory hog, now it's back and it's again keeping massive amounts of memory 2012-11-02 00:05:28 +01:00
Tank_Master cb24ec2e0d merge LL 3.3.4 beta
autoreplace does not work (needs a hook in nearby chat), no cmd line option to add new stuff
spellcheck only works when packaged
2012-07-16 21:59:14 -07:00
Dave Parks b7cfd8c7f0 MAINT-616 Factor out calls to glGenFoo where possible, add setting to control synchronizing strategy WRT occlusion queries, add experimental transform feedback driven LoD update 2012-05-21 23:33:25 -05:00
Dave Parks 89b0b6ac71 MAINT-616 Different VBO Pooling scheme -- populate VBO pool with buffers that are likely to be requested, but never reuse a previously used buffer. 2012-05-15 16:42:04 -05:00
Dave Parks 2e1067dd32 MAINT-708 Reduce amount of video memory consumed by vertex buffer objects. 2012-03-13 17:34:11 -05:00
Nat Goodspeed 0249a0d1e7 Automated merge with http://hg.secondlife.com/viewer-release 2012-04-26 14:07:09 -04:00
Dave Parks 3953477ee5 MAINT-708 Reduce amount of video memory consumed by vertex buffer objects. 2012-03-13 17:34:11 -05:00
Leslie Linden ae7d475aeb SH-2794, resolved merge conflict and corrected logic in LLVertexBuffer::determineUsage() to work the way it used to. 2012-02-06 12:43:23 -08:00
Brad Payne (Vir Linden) 344c46a20c merge, conflicts in llvertexbuffer.{h,cpp} 2012-02-06 15:08:41 -05:00
Dave Parks 17b21fc197 Fix for windows build 2012-01-26 12:50:24 -06:00
Leslie Linden 71a660e006 * Fixed up shader compilation errors to get lighting and shadows working again on Mac.
* Cleaned up vertex buffer allocation to guarantee allocation and deallocation using the same allocation methods from the same pools.
* Added new shader feature for atmospheric helpers in the absence of lighting.

Reviewed by davep.
2012-01-26 10:17:41 -08:00
Dave Parks 655505d304 SH-2794 Fix for bad textures on some hardware when vbo mapping disabled (always use GL_STREAM_DRAW as the usage hint when mapping is disabled as geometry will be uploaded again and again) 2012-01-19 14:57:07 -06:00
Dave Parks 0c1fc78bd9 SH-2794 Map buffer now performs as well as buffersubdata 2012-01-18 12:40:11 -06:00
Dave Parks 5546041301 SH-2794 WIP -- work on getting glMapBuffer and friends to behave 2011-12-21 15:55:14 -06:00
Dave Parks 7b6723d1e0 SH-1650 Mitigate memory fragmentation by holding onto and reusing VBOs 2011-10-13 01:19:45 -05:00
Dave Parks ae0cd901f8 SH-2428 Make GCC 4.4 happy by using LLVector4a striders with LLAlphaObject subclasses -- also vectorize particle VBO updates 2011-10-04 17:25:55 -05:00
Dave Parks 54ba56dbe9 SH-2244 changes to run in a core context on AMD hardware without generating deprecation or performance warnings 2011-09-26 18:31:55 -05:00
Dave Parks 57620a9da9 SH-2244 Vertex buffer cleanup, fix for bad vertices in rigged attachments, added "RenderUseVAO" debug setting to control whether or not to use VAO's in non-core GL profiles. 2011-09-23 17:59:05 -05:00
Dave Parks b9926e8f57 SH-2244 Better VAO support -- still slower than non-VAO implementation for some reason 2011-09-23 02:29:53 -05:00
Dave Parks a2d08a6d80 SH-2244 Fix "RenderGLCoreProfile" actually make a core profile context and modify viewer to run under said context without generating errors. 2011-09-22 00:10:57 -05:00
Dave Parks c0ca2c62fd merge 2011-09-13 16:11:22 -05:00
Xiaohong Bao 749d9ebadc Merge 2011-08-31 10:48:56 -06:00
Dave Parks 2ee8154780 SH-2238, SH-2223, SH-SH-2242 glVertexAttrib throughout main render pipeline complete, preview renders and debug displays still pending. Also fixed a render glitch and a crash (JIRAs listed). 2011-08-10 13:01:14 -05:00
Dave Parks 14f6bbadef SH-2242 Work in progress migrating to glVertexAttrib everywhere 2011-08-08 15:29:23 -05:00
Xiaohong Bao d951267467 Merge from viewer-development 2011-07-15 12:14:34 -06:00
Dave Parks 268f3dbdea SH-2038 Fix for some compiler errors from the apple tweaks -- also add a fence API (disabled for now). 2011-07-10 02:29:39 -05:00
Dave Parks 6992dbc1e3 SH-1682 Use GL_ARB_map_buffer_range to mitigate impact of mapping larger vertex buffer objects. Limit number of textures per batch to 6 (prevents frame stalls on NVIDIA). 2011-06-01 23:46:04 -05:00
Dave Parks 1fd46831f0 SH-1682 Work in progress on using texture indexes to improve batch size (wow, super fast so far) 2011-05-29 00:40:57 -05:00
Dave Parks 8275820df0 SH-1301 Add some extra asserts around VBO usage to try and narrow down a crash. 2011-04-15 14:11:48 -05:00
Dave Parks 719610e68b SH-1225 Import convex hull data from .slm 2011-03-29 11:50:02 -05:00
Xiaohong Bao 3f6208b176 Merge 2011-02-24 16:20:43 -07:00