Re: [Mesa-dev] r600g: status of my work on the shader optimization
Hi! Amazing work! I see some 50 % speed ups in FlightGear and even more. While normally 3D clouds tear performance down to an unflyable stutter, with your branch I can fly in densly clouded conditions at usable framerates. I can now turn all shaders to maximum and enjoy the view. This makes a huge difference. Unfortunately there's a downside as well: [ 4644.872676] radeon :01:00.0: GPU lockup CP stall for more than 1msec [ 4644.872681] radeon :01:00.0: GPU lockup (waiting for 0x000a0c1d last fence id 0x000a0c16) [ 4644.873727] radeon :01:00.0: Saved 215 dwords of commands on ring 0. [ 4644.873730] radeon :01:00.0: GPU softreset: 0x0003 [ 4644.880943] radeon :01:00.0: GRBM_STATUS = 0xF4002828 [ 4644.880946] radeon :01:00.0: GRBM_STATUS_SE0 = 0xC005 [ 4644.880948] radeon :01:00.0: GRBM_STATUS_SE1 = 0x0007 [ 4644.880950] radeon :01:00.0: SRBM_STATUS = 0x20C0 [ 4644.880952] radeon :01:00.0: R_008674_CP_STALLED_STAT1 = 0x [ 4644.880955] radeon :01:00.0: R_008678_CP_STALLED_STAT2 = 0x4004 [ 4644.880957] radeon :01:00.0: R_00867C_CP_BUSY_STAT = 0x00048002 [ 4644.880959] radeon :01:00.0: R_008680_CP_STAT = 0x80268643 [ 4644.880961] radeon :01:00.0: GRBM_SOFT_RESET=0x7F6B [ 4644.881014] radeon :01:00.0: GRBM_STATUS = 0x3828 [ 4644.881016] radeon :01:00.0: GRBM_STATUS_SE0 = 0x0007 [ 4644.881019] radeon :01:00.0: GRBM_STATUS_SE1 = 0x0007 [ 4644.881021] radeon :01:00.0: SRBM_STATUS = 0x20C0 [ 4644.881023] radeon :01:00.0: R_008674_CP_STALLED_STAT1 = 0x [ 4644.881025] radeon :01:00.0: R_008678_CP_STALLED_STAT2 = 0x [ 4644.881027] radeon :01:00.0: R_00867C_CP_BUSY_STAT = 0x [ 4644.881030] radeon :01:00.0: R_008680_CP_STAT = 0x [ 4644.898500] radeon :01:00.0: GPU reset succeeded, trying to resume R600_SB=0 is stable while with R600_SB=1 I can somewhat reliably reproduce these lockups. Is there someting I can do to debug this? Or are problems like this already expected? In any case please keep up this work! Even though it might eventually be replaced by some llvm backend, this could make your users very happy in the mean time. Regards, Stefan ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] r600g: status of my work on the shader optimization
On Friday 15 February 2013 02:21:56 Vadim Girlin wrote: > I pushed the fix for one problem that I noticed with FlightGear to that > branch, please test if it helps with your lockups as well. I have a commit d6327a397ed54b3b2e40d521a762a380c25cba48 but the lockups still occur: [ 125.082988] radeon :01:00.0: GPU lockup CP stall for more than 1msec [ 125.082993] radeon :01:00.0: GPU lockup (waiting for 0x21e8 last fence id 0x21e1) [ 125.084038] radeon :01:00.0: Saved 215 dwords of commands on ring 0. [ 125.084042] radeon :01:00.0: GPU softreset: 0x0003 [ 125.099151] radeon :01:00.0: GRBM_STATUS = 0xE7730828 [ 125.099153] radeon :01:00.0: GRBM_STATUS_SE0 = 0xFC01 [ 125.099154] radeon :01:00.0: GRBM_STATUS_SE1 = 0x0007 [ 125.099155] radeon :01:00.0: SRBM_STATUS = 0x20C0 [ 125.099156] radeon :01:00.0: R_008674_CP_STALLED_STAT1 = 0x [ 125.099157] radeon :01:00.0: R_008678_CP_STALLED_STAT2 = 0x00010800 [ 125.099158] radeon :01:00.0: R_00867C_CP_BUSY_STAT = 0x00028006 [ 125.099159] radeon :01:00.0: R_008680_CP_STAT = 0x80038647 [ 125.099160] radeon :01:00.0: GRBM_SOFT_RESET=0x7F6B [ 125.099211] radeon :01:00.0: GRBM_STATUS = 0x3828 [ 125.099212] radeon :01:00.0: GRBM_STATUS_SE0 = 0x0007 [ 125.099213] radeon :01:00.0: GRBM_STATUS_SE1 = 0x0007 [ 125.099214] radeon :01:00.0: SRBM_STATUS = 0x20C0 [ 125.099215] radeon :01:00.0: R_008674_CP_STALLED_STAT1 = 0x [ 125.099216] radeon :01:00.0: R_008678_CP_STALLED_STAT2 = 0x [ 125.099217] radeon :01:00.0: R_00867C_CP_BUSY_STAT = 0x [ 125.099219] radeon :01:00.0: R_008680_CP_STAT = 0x [ 125.116712] radeon :01:00.0: GPU reset succeeded, trying to resume > If that fix doesn't help, you might want to test with the debug build of > mesa (configure with "--enable-debug" option) if you haven't tried it > yet - many issues then may be detected by the asserts in the code. Unfortunately no change. I do not see any messages or abort()s. Just the lockups. Regards, Stefan ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] r600g: status of my work on the shader optimization
Good news! I gave the r600-sb branch a good testing at commit 265ae41b1f1d086d35d274c7378c43cddb8215c8 and so far I've not had a single lockup in about 1 1/2 hours of flight time! The downside is that this is with R600_HYPERZ=0. But with HYPERZ enabled, I get lockups on master as well, so it would seem your branch is in pretty good shape. Testing done on a Radeon HD 5670 using kernel 3.8 Regards, Stefan ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
