Re: [Mesa-dev] [PATCH] mesa gallium: use compute shaders for vaapi blit

On Wed, Apr 3, 2019 at 9:06 AM Ilia Mirkin wrote: > On Wed, Apr 3, 2019 at 8:38 AM Marek Olšák wrote: > > > > On Tue, Apr 2, 2019 at 2:14 PM Eric Anholt wrote: > >> > >> Ilia Mirkin writes: > >> > >> > Shouldn't this sort of decision be

Re: [Mesa-dev] [PATCH] mesa gallium: use compute shaders for vaapi blit

On Tue, Apr 2, 2019 at 2:14 PM Eric Anholt wrote: > Ilia Mirkin writes: > > > Shouldn't this sort of decision be left up to the driver? If the > > driver would like to use CS for blits, fine, but why not let it blit > > in the most optimal way possible and force it to use a compute shader? > >

Re: [Mesa-dev] [PATCH v2 8/8] egl: add EGL_platform_device support

Do you have a branch with patch 7/8 and 8/8? I'm interested in EGL_EXT_platform_device on radeonsi. Thanks, Marek On Wed, Oct 3, 2018 at 4:36 AM Mathias Fröhlich wrote: > Hi Emil, > > Ok, thanks for picking that up. > > On Tuesday, 2 October 2018 12:23:30 CEST Emil Velikov wrote: > > On Thu,

Re: [Mesa-dev] [PATCH v2 4/8] gallium/u_vbuf: export u_vbuf_get_minmax_index

Reviewed-by: Marek Olšák Marek On Fri, Mar 22, 2019 at 10:30 AM Qiang Yu wrote: > This helper function can be used by driver which > always need min/max index. > > Signed-off-by: Qiang Yu > --- > src/gallium/auxiliary/util/u_vbuf.c | 7 +++ > src/gallium/auxili

Re: [Mesa-dev] [PATCH 4/9] gallium: add a cap to force compute minmax indices

On Fri, Mar 15, 2019 at 9:30 PM Qiang Yu wrote: > From: Erico Nunes > > pipe_draw_info has min_index and max_index fields that can be useful in > indexed drawing, however gallium may decide to not compute them in some > cases to avoid impacting performance if the driver won't need them. >

Re: [Mesa-dev] [PATCH 1/2] radeonsi: always use compute rings for clover on CI and newer (v2)

On Tue, Apr 2, 2019 at 4:57 AM Michel Dänzer wrote: > On 2019-04-02 12:39 a.m., Marek Olšák wrote: > > On Mon, Apr 1, 2019 at 1:28 PM Jan Vesely > wrote: > >> On Mon, 2019-04-01 at 12:30 -0400, Marek Olšák wrote: > >>> Does the attached patch fix the copy-buff

Re: [Mesa-dev] [PATCH 1/2] radeonsi: always use compute rings for clover on CI and newer (v2)

On Mon, Apr 1, 2019 at 1:28 PM Jan Vesely wrote: > On Mon, 2019-04-01 at 12:30 -0400, Marek Olšák wrote: > > Does the attached patch fix the copy-buffer test? > > it does thanks. > Won't the compute only context still need some synchronization? > Is there anything else to gu

[Mesa-dev] [PATCH] radeonsi: don't use PFP_SYNC_ME with compute-only contexts

From: Marek Olšák Fixes: a1378639ab1 "radeonsi: always use compute rings for clover on CI and newer (v2)" --- src/gallium/drivers/radeonsi/si_cp_dma.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/gallium/drivers/radeonsi/si_cp_dma.c b/src/gallium/driver

Re: [Mesa-dev] [PATCH] RFC: Workaround for pthread_setaffinity_np() seccomp filtering

Fröhlich > wrote: > > > > On Friday, 1 March 2019 12:15:08 CET Eero Tamminen wrote: > > > Hi, > > > > > > On 1.3.2019 11.12, Michel Dänzer wrote: > > > > On 2019-02-28 8:41 p.m., Marek Olšák wrote: > > > >>> On Thu, Feb 28, 2

Re: [Mesa-dev] [PATCH] RFC: Workaround for pthread_setaffinity_np() seccomp filtering

FYI, starting with AMD Ryzen, multithreaded apps and libs pretty much have to change thread affinity to get good performance out of multithreading. Marek On Thu, Feb 28, 2019, 11:41 AM Marek Olšák wrote: > On Thu, Feb 28, 2019 at 11:13 AM Marc-André Lureau < > marcandre.lur...@gmail.c

Re: [Mesa-dev] [PATCH 1/2] radeonsi: always use compute rings for clover on CI and newer (v2)

On Thu, Feb 28, 2019 at 4:44 AM Jan Vesely wrote: > On Tue, 2019-02-26 at 18:34 -0500, Marek Olšák wrote: > > I ran a simple test verifying that compute is working properly on the > > compute ring. > > I guess this was not on raven? With his patch I no loner see gfx > ti

[Mesa-dev] [PATCH 0/5] RadeonSI: Displayable DCC for Ravens

Hi, This series enables DCC for scanout on Ravens. It requires kernel driver version >= 3.31.0 and my xf86-video-amdgpu patch. There is one issue to resolve: Steam crashes in addrlib/ComputeDccAddrFromCoord. Please review, Thanks, Marek ___ mesa-dev

[Mesa-dev] [PATCH 3/5] radeonsi: add support for displayable DCC for 1 RB chips

From: Marek Olšák This is the simpler codepath - just disable RB and pipe alignment for DCC. --- src/amd/common/ac_gpu_info.c | 2 + src/amd/common/ac_gpu_info.h | 3 + src/amd/common/ac_surface.c| 25 +++- src/amd/common/ac_surface.h

[Mesa-dev] [PATCH 4/5] radeonsi: add support for displayable DCC for multi-RB chips

From: Marek Olšák A compute shader is used to reorder DCC data from aligned to unaligned. --- src/amd/common/ac_gpu_info.c | 1 + src/amd/common/ac_gpu_info.h | 3 + src/amd/common/ac_surface.c | 125 -- src/amd/common

[Mesa-dev] [PATCH 5/5] radeonsi: enable displayable DCC on Ravens

From: Marek Olšák --- src/amd/common/ac_gpu_info.c | 8 src/amd/vulkan/winsys/amdgpu/radv_amdgpu_winsys.c | 4 2 files changed, 12 insertions(+) diff --git a/src/amd/common/ac_gpu_info.c b/src/amd/common/ac_gpu_info.c index d890172227c..c53335bbb7d 100644

[Mesa-dev] [PATCH 2/5] radeonsi: add ability to bind images as image buffers

From: Marek Olšák so that we can bind DCC (texture) as an image buffer. --- src/gallium/drivers/radeonsi/si_descriptors.c | 9 ++--- src/gallium/drivers/radeonsi/si_pipe.h| 2 ++ 2 files changed, 8 insertions(+), 3 deletions(-) diff --git a/src/gallium/drivers/radeonsi

[Mesa-dev] [PATCH 1/5] radeonsi/gfx9: add support for PIPE_ALIGNED=0

From: Marek Olšák Needed by displayable DCC. We need to flush L2 after rendering if PIPE_ALIGNED=0 and DCC is enabled. --- src/gallium/drivers/radeonsi/si_blit.c| 7 --- .../drivers/radeonsi/si_compute_blit.c| 9 +++-- src/gallium/drivers/radeonsi/si_pipe.h

Re: [Mesa-dev] [PATCH] RFC: Workaround for pthread_setaffinity_np() seccomp filtering

eau wrote: > > > On Thu, Feb 28, 2019 at 1:17 AM Marek Olšák wrote: > > >> I'd rather have something more robust than an env var, like catching > SIGSYS. > > > > SIGSYS is info for the invoking parent, not the (Mesa) process doing the > > syscall. > &g

Re: [Mesa-dev] [PATCH] radeonsi/nir: move si_lower_nir() call into compiler thread

Reviewed-by: Marek Olšák Marek On Tue, Feb 26, 2019 at 10:30 PM Timothy Arceri wrote: > This helps improve compile times. For example the shader-db dolphin > shader shaders/dolphin/ubershaders/120.shader_test goes from > ~1.69 -> ~1.57 seconds on my machine with this change.

Re: [Mesa-dev] [PATCH] gallium: Implement APPLE_object_purgeable (iris, freedeno, vc4)

On Tue, Feb 26, 2019 at 4:39 PM Chris Wilson wrote: > A few of the GEM drivers provide matching ioctls to allow control of > their bo caches. Hook these up to APPLE_object_purgeable to allow > clients to discard video memory under pressure where they are able to > fallback to restoring content

Re: [Mesa-dev] [PATCH] glsl: fix shader cache for packed param list

Reviewed-by: Marek Olšák Marek On Wed, Feb 27, 2019 at 2:26 AM Timothy Arceri wrote: > Some types of params such as some builtins are always padded. We > need to keep track of this so we can restore the list correctly. > > Here we also remove a couple of cache entries that are

Re: [Mesa-dev] [PATCH] RFC: Workaround for pthread_setaffinity_np() seccomp filtering

I'd rather have something more robust than an env var, like catching SIGSYS. Marek On Wed, Feb 27, 2019 at 6:13 PM wrote: > From: Marc-André Lureau > > Since commit d877451b48a59ab0f9a4210fc736f51da5851c9a ("util/u_queue: > add UTIL_QUEUE_INIT_SET_FULL_THREAD_AFFINITY"), mesa calls >

[Mesa-dev] [PATCH] radeonsi: fix assertion failure by using the correct type

From: Marek Olšák src/gallium/drivers/radeonsi/si_state_viewport.c:196: si_emit_guardband: Assertion `vp_as_scissor.maxx <= max_viewport_size[vp_as_scissor.quant_mode] && vp_as_scissor.maxy <= max_viewport_size[vp_as_scissor.quant_mode]' failed. The comparison was unsigned, so

[Mesa-dev] [PATCH 2/3] omx: clean up enc_LoadImage_common

From: Marek Olšák - add *pipe - add documentation --- .../state_trackers/omx/vid_enc_common.c | 53 +-- 1 file changed, 37 insertions(+), 16 deletions(-) diff --git a/src/gallium/state_trackers/omx/vid_enc_common.c b/src/gallium/state_trackers/omx/vid_enc_common.c index

[Mesa-dev] [PATCH 3/3] omx: add a compute path in enc_LoadImage_common

From: Marek Olšák --- .../state_trackers/omx/bellagio/vid_enc.c | 3 + .../state_trackers/omx/tizonia/h264eprc.c | 3 + .../state_trackers/omx/vid_enc_common.c | 219 +++--- .../state_trackers/omx/vid_enc_common.h | 8 +- 4 files changed, 196 insertions

[Mesa-dev] [PATCH 1/3] gallium: add pipe_grid_info::last_block

From: Marek Olšák The OpenMAX state tracker will use this. RadeonSI is adapted to use pipe_grid_info::last_block instead of its internal state. --- src/gallium/auxiliary/util/u_screen.c | 3 +++ src/gallium/docs/source/screen.rst| 2 ++ src/gallium/drivers/radeonsi

Re: [Mesa-dev] [PATCH 1/2] radeonsi: always use compute rings for clover on CI and newer (v2)

t; > Jan > > On Tue, Feb 26, 2019 at 3:00 PM Marek Olšák wrote: > >> I'll just push it. >> >> Marek >> >> On Mon, Feb 25, 2019 at 9:37 PM Dieter Nützel >> wrote: >> >>> Hello Marek, >>> >>> this series need a rebase (i

Re: [Mesa-dev] [PATCH 5/6] radeonsi: compile clear and copy buffer compute shaders on demand

ble. > > Thanks, > Tim > > On 12/1/19 10:53 am, Marek Olšák wrote: > > From: Marek Olšák > > > > same as all other shaders > > --- > > src/gallium/drivers/radeonsi/si_compute_blit.c | 14 ++ > > src/gallium/drivers/radeonsi/si_pipe.

Re: [Mesa-dev] [PATCH 1/2] radeonsi: always use compute rings for clover on CI and newer (v2)

I'll just push it. Marek On Mon, Feb 25, 2019 at 9:37 PM Dieter Nützel wrote: > Hello Marek, > > this series need a rebase (if you have some time). > > Dieter > > Am 12.02.2019 19:12, schrieb Marek Olšák: > > From: Marek Olšák > > > > initialize al

Re: [Mesa-dev] [PATCH 0/4] RadeonSI: Upload constants to VRAM via SDMA

We need to extend the CS ioctl to allow submitting 2 command buffers at the same time. Marek On Mon, Feb 25, 2019, 10:06 PM Dieter Nützel wrote: > Hello Marek, > > you wrote with your series sent: > > [-] > Trivial benchmarks such as glxgears can expect 20% decrease > in performance due to the

Re: [Mesa-dev] [PATCH] mesa: fix display list corner case assertion

Reviewed-by: Marek Olšák Marek On Mon, Feb 25, 2019 at 5:03 PM Brian Paul wrote: > This fixes a failed assertion in glDeleteLists() for the following > case: > > list = glGenLists(1); > glDeleteLists(list, 1); > > when those are the first display list commands issued

Re: [Mesa-dev] [PATCH] util: Don't block SIGSYS for new threads

Reviewed-by: Marek Olšák Marek On Sat, Feb 23, 2019 at 2:05 AM Drew Davenport wrote: > SIGSYS is needed for programs using seccomp for sandboxing. > --- > src/util/u_thread.h | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/src/util/u_thread.h b/s

Re: [Mesa-dev] [PATCH v2] radeonsi: fix query buffer allocation

Reviewed-by: Marek Olšák Marek On Sun, Feb 24, 2019 at 6:56 PM Timothy Arceri wrote: > Fix the logic for buffer full check on alloc. > > This patch just takes the fix Nicolai attached to the bug report > and updates it to work on master. > > Fixes: e0f0d3675d4

Re: [Mesa-dev] [PATCH] st/mesa: Reduce array updates due to current changes.

Reviewed-by: Marek Olšák Marek On Sun, Feb 24, 2019 at 1:46 AM wrote: > From: Mathias Fröhlich > > Hi Brian, > > Following a small optimization in the gallium state tracker to > avoid flagging ST_NEW_VERTEX_ARRAYS a bit more often: > > please review! > > be

[Mesa-dev] [PATCH 7/7] radeonsi: implement ARB/KHR_parallel_shader_compile callbacks

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_pipe.c | 31 ++ 1 file changed, 31 insertions(+) diff --git a/src/gallium/drivers/radeonsi/si_pipe.c b/src/gallium/drivers/radeonsi/si_pipe.c index b965d9d64d4..7dbd4cb2c40 100644 --- a/src/gallium/drivers/radeonsi

[Mesa-dev] [PATCH 5/7] util/queue: hold a lock when reading num_threads in util_queue_finish

From: Marek Olšák Reviewed-by: Ian Romanick --- src/util/u_queue.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/util/u_queue.c b/src/util/u_queue.c index cfd2a08e3c8..5e0c1095569 100644 --- a/src/util/u_queue.c +++ b/src/util/u_queue.c @@ -585,29 +585,29

[Mesa-dev] [PATCH 1/7] mesa: implement ARB/KHR_parallel_shader_compile

From: Marek Olšák Tested by piglit. --- docs/features.txt | 2 +- docs/relnotes/19.0.0.html | 2 ++ src/mapi/glapi/gen/gl_API.xml | 15 ++- src/mesa/main/dd.h | 7 +++ src/mesa/main/extensions_table.h

[Mesa-dev] [PATCH 4/7] util/queue: add ability to kill a subset of threads

From: Marek Olšák for ARB_parallel_shader_compile --- src/util/u_queue.c | 52 ++ src/util/u_queue.h | 5 ++--- 2 files changed, 36 insertions(+), 21 deletions(-) diff --git a/src/util/u_queue.c b/src/util/u_queue.c index 48c5c79552d..cfd2a08e3c8

[Mesa-dev] [PATCH 3/7] util/queue: move thread creation into a separate function

From: Marek Olšák Reviewed-by: Ian Romanick --- src/util/u_queue.c | 56 ++ 1 file changed, 32 insertions(+), 24 deletions(-) diff --git a/src/util/u_queue.c b/src/util/u_queue.c index 3812c824b6d..48c5c79552d 100644 --- a/src/util/u_queue.c +++ b

[Mesa-dev] [PATCH 6/7] util/queue: add util_queue_adjust_num_threads

From: Marek Olšák for ARB_parallel_shader_compile Reviewed-by: Ian Romanick --- src/util/u_queue.c | 50 -- src/util/u_queue.h | 8 2 files changed, 52 insertions(+), 6 deletions(-) diff --git a/src/util/u_queue.c b/src/util/u_queue.c

[Mesa-dev] [PATCH 2/7] gallium: implement ARB/KHR_parallel_shader_compile

From: Marek Olšák --- src/gallium/include/pipe/p_screen.h| 13 ++ src/mesa/state_tracker/st_cb_program.c | 59 +- 2 files changed, 71 insertions(+), 1 deletion(-) diff --git a/src/gallium/include/pipe/p_screen.h b/src/gallium/include/pipe/p_screen.h index

Re: [Mesa-dev] [PATCH] radeonsi: fix query buffer allocation

Nicolai also has a change in si_query_buffer_reset. Is there a reason the change is not here? Thanks, Marek On Thu, Feb 21, 2019 at 10:09 PM Timothy Arceri wrote: > Fix the logic for buffer full check on alloc. > > This patch just takes the fix Nicolai attached to the bug report > and updates

Re: [Mesa-dev] [PATCH] tgsi: don't set tgsi_info::uses_bindless_images for constbufs and hw atomics

stuff -- probably needs a bit of subtlety there. > > -ilia > > On Thu, Feb 21, 2019 at 12:03 AM Marek Olšák wrote: > > > > From: Marek Olšák > > > > This might have decreased performance for radeonsi/tgsi, because most > > most shaders claimed they used bind

[Mesa-dev] [PATCH] tgsi: don't set tgsi_info::uses_bindless_images for constbufs and hw atomics

From: Marek Olšák This might have decreased performance for radeonsi/tgsi, because most most shaders claimed they used bindless. Cc: 18.3 19.0 --- src/gallium/auxiliary/tgsi/tgsi_scan.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/gallium/auxiliary/tgsi

Re: [Mesa-dev] A few NIR compile time optimisations

On Wed, Feb 20, 2019 at 2:31 AM Connor Abbott wrote: > > > On Wed, Feb 20, 2019 at 4:29 AM Marek Olšák wrote: > >> On Tue, Feb 19, 2019 at 7:57 PM Rob Clark wrote: >> >>> On Tue, Feb 19, 2019 at 6:49 PM Marek Olšák wrote: >>> > >>>

Re: [Mesa-dev] A few NIR compile time optimisations

On Tue, Feb 19, 2019 at 7:57 PM Rob Clark wrote: > On Tue, Feb 19, 2019 at 6:49 PM Marek Olšák wrote: > > > > st_link_shader takes 55% of CPU time with NIR, and 9% with TGSI. > > > > nir_validate_shader 49% > > > > nir_validate_shader is overused. It do

Re: [Mesa-dev] [PATCH 4/4] radeonsi: use SDMA for uploading data through const_uploader

box->width + (box->x % SI_MAP_BUFFER_ALIGNMENT), sctx->screen->info.tcc_cache_line_size, , (struct pipe_resource**), Marek On Mon, Feb 11, 2019 at 4:38 AM Nicolai Hähnle wrote: > On 07.02.19 02:22, Marek Olšák wrot

Re: [Mesa-dev] [PATCH 0/4] RadeonSI: Upload constants to VRAM via SDMA

Yeah, u_threaded_context is broken. Marek On Thu, Feb 14, 2019 at 8:06 PM Dieter Nützel wrote: > Am 12.02.2019 05:10, schrieb Dieter Nützel: > > Am 12.02.2019 03:22, schrieb Dieter Nützel: > >> Am 12.02.2019 00:40, schrieb Dieter Nützel: > >>> Sorry that I step in so late, but the whole family

Re: [Mesa-dev] A few NIR compile time optimisations

st_link_shader takes 55% of CPU time with NIR, and 9% with TGSI. nir_validate_shader 49% nir_validate_shader is overused. It doesn't make sense even in debug builds. Marek On Wed, Feb 13, 2019 at 4:26 AM Connor Abbott wrote: > Reviewed-by: Connor Abbott > > I'm a bit surprised it's that

Re: [Mesa-dev] [PATCH] radeonsi/nir: set shader_buffers_declared properly

Reviewed-by: Marek Olšák Marek On Mon, Feb 11, 2019 at 10:46 PM Timothy Arceri wrote: > --- > src/gallium/drivers/radeonsi/si_shader_nir.c | 32 ++-- > 1 file changed, 22 insertions(+), 10 deletions(-) > > diff --git a/src/gallium/drivers/radeonsi/si_shade

Re: [Mesa-dev] [PATCH v2 2/2] radeonsi/nir: set colors_read properly

For the series: Reviewed-by: Marek Olšák Marek On Mon, Feb 11, 2019 at 7:15 PM Timothy Arceri wrote: > shader-db results for VEGA64: > > Totals from affected shaders: > SGPRS: 1976 -> 1976 (0.00 %) > VGPRS: 1240 -> 1144 (-7.74 %) > Spilled SGPRs: 145 -> 145 (0.00

Re: [Mesa-dev] [Mesa-stable] [PATCH] drirc: Add sddm-greeter to adaptive_sync blacklist.

Pushed, thanks! Marek On Mon, Feb 18, 2019 at 10:01 PM Mario Kleiner wrote: > This is the sddm login screen. > > Fixes: a9c36dbf9c56 ("drirc: Initial blacklist for adaptive sync") > Signed-off-by: Mario Kleiner > Cc: 19.0 > --- > src/util/00-mesa-defaults.conf | 3 +++ > 1 file changed, 3

[Mesa-dev] [PATCH 2/3] radeonsi: always enable NIR for Civilization 6 to fix corruption

From: Marek Olšák Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104602 --- src/util/00-mesa-defaults.conf | 3 +++ 1 file changed, 3 insertions(+) diff --git a/src/util/00-mesa-defaults.conf b/src/util/00-mesa-defaults.conf index cb0e6e659e2..81f23c97941 100644 --- a/src/util/00-mesa

[Mesa-dev] [PATCH 3/3] driconf: add Civ6Sub executable for Civilization 6

From: Marek Olšák I'm getting Civ6Sub instead of Civ6. --- src/util/00-mesa-defaults.conf | 6 ++ 1 file changed, 6 insertions(+) diff --git a/src/util/00-mesa-defaults.conf b/src/util/00-mesa-defaults.conf index 81f23c97941..8abc50c9f26 100644 --- a/src/util/00-mesa-defaults.conf +++ b

[Mesa-dev] [PATCH 1/3] radeonsi: add driconf option radeonsi_enable_nir

From: Marek Olšák --- src/gallium/drivers/radeonsi/driinfo_radeonsi.h | 1 + src/gallium/drivers/radeonsi/si_pipe.c | 3 ++- src/util/xmlpool/t_options.h| 5 + 3 files changed, 8 insertions(+), 1 deletion(-) diff --git a/src/gallium/drivers/radeonsi

Re: [Mesa-dev] [PATCH v6 5/5] gallium/auxiliary/vl: Add video compositor compute shader render

On Thu, Feb 14, 2019, 5:13 PM James Zhu > On 2019-02-14 4:44 p.m., Marek Olšák wrote: > > This still creates gfx shaders. If we switch MM to compute IBs, calling > any gfx function will crash. > > Right now only fs_video_buffer/fs_weave_rgb/fs_rgba have compute shader > repla

Re: [Mesa-dev] [PATCH v6 5/5] gallium/auxiliary/vl: Add video compositor compute shader render

This still creates gfx shaders. If we switch MM to compute IBs, calling any gfx function will crash. Marek On Tue, Feb 12, 2019, 2:53 PM Zhu, James Add compute shader initilization, assign and cleanup in vl_compositor API. > Set video compositor compute shader render as default when pipe

Re: [Mesa-dev] [PATCH 00/26] RadeonSI: Primitive culling with async compute

On Thu, Feb 14, 2019 at 1:43 PM Dieter Nützel wrote: > For the whole series (the updated branch merged in) > > Tested-by: Dieter Nützel > > on Polaris 20 > > FreeCAD, Blender, UH, UV, US, some VTK apps > No surprising speed up but e.g. NO slowdown. > > tb stands even for > [Mesa-dev] [PATCH

[Mesa-dev] [PATCH 0/4] RadeonSI: Follow-up for the primitive culling series

Hi, This fixes Sea Islands and makes it faster. Please review. Thanks, Marek ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev

[Mesa-dev] [PATCH 3/4] radeonsi: fix cache coherency for primitive culling on CI

From: Marek Olšák --- .../radeonsi/si_compute_prim_discard.c| 20 ++- 1 file changed, 11 insertions(+), 9 deletions(-) diff --git a/src/gallium/drivers/radeonsi/si_compute_prim_discard.c b/src/gallium/drivers/radeonsi/si_compute_prim_discard.c index 1b36bd6ca53

[Mesa-dev] [PATCH 2/4] winsys/amdgpu: clean up and remove nonsensical assertion

From: Marek Olšák The assertion considers max_dw from the current IB in the chain, but big_ib_buffer is a buffer for the next IB, which can be smaller. --- src/gallium/winsys/amdgpu/drm/amdgpu_cs.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/src/gallium/winsys/amdgpu

[Mesa-dev] [PATCH 1/4] winsys/amdgpu: enable chaining for compute IBs

From: Marek Olšák --- src/gallium/winsys/amdgpu/drm/amdgpu_cs.c | 12 ++-- 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/src/gallium/winsys/amdgpu/drm/amdgpu_cs.c b/src/gallium/winsys/amdgpu/drm/amdgpu_cs.c index 912307e7d11..eb2944766fc 100644 --- a/src/gallium/winsys

[Mesa-dev] [PATCH 4/4] radeonsi: emulate REWIND using INDIRECT_BUFFER for primitive culling on CI

From: Marek Olšák This increases "Paraview - Many Spheres" performance from 37.91 to 43 fps on Hawaii. --- src/gallium/drivers/r300/r300_blit.c | 2 +- src/gallium/drivers/r300/r300_render.c| 2 +- src/gallium/drivers/r600/r600_hw_context.c| 2 +- src/galli

Re: [Mesa-dev] [PATCH 00/26] RadeonSI: Primitive culling with async compute

I have some fixes for Sea Islands that improve Radeon 290X performance to 43 fps, moving it just below Radeon VII in the picture. Marek On Wed, Feb 13, 2019 at 12:16 AM Marek Olšák wrote: > Hi, > > This patch series uses async compute to do primitive culling before > the v

Re: [Mesa-dev] [PATCH 00/26] RadeonSI: Primitive culling with async compute

> [38980.844642] [drm:amdgpu_ctx_mgr_fini [amdgpu]] *ERROR* ctx > 1bee3a35 is still alive > > Newer 'amd-staging-drm-next' needed? #0bf64b0a9f78 currently > > If I only had some big triangle apps...;-) > > Dieter > > Am 13.02.2019 17:36, schrieb Marek Olšák: >

Re: [Mesa-dev] [PATCH 00/26] RadeonSI: Primitive culling with async compute

4642] [drm:amdgpu_ctx_mgr_fini [amdgpu]] *ERROR* ctx > 1bee3a35 is still alive > > Newer 'amd-staging-drm-next' needed? #0bf64b0a9f78 currently > > If I only had some big triangle apps...;-) > > Dieter > > Am 13.02.2019 17:36, schrieb Marek Olšák: > > Dieter,

Re: [Mesa-dev] [PATCH 00/26] RadeonSI: Primitive culling with async compute

On Wed, Feb 13, 2019 at 11:51 AM Axel Davy wrote: > On 13/02/2019 17:42, Marek Olšák wrote: > > On Wed, Feb 13, 2019 at 2:28 AM Axel Davy wrote: > >> On 13/02/2019 06:15, Marek Olšák wrote: >> > I decided to enable this optimization on all Pro graphics cards. >&g

Re: [Mesa-dev] [PATCH 00/26] RadeonSI: Primitive culling with async compute

On Wed, Feb 13, 2019 at 2:28 AM Axel Davy wrote: > On 13/02/2019 06:15, Marek Olšák wrote: > > I decided to enable this optimization on all Pro graphics cards. > > The reason is that I haven't had time to benchmark games. > > This decision may be changed based on co

Re: [Mesa-dev] [PATCH 00/26] RadeonSI: Primitive culling with async compute

d AMD64 > > Please have a look at my post @Phoronix: > > https://www.phoronix.com/forums/forum/phoronix/latest-phoronix-articles/1079916-radeonsi-picks-up-primitive-culling-with-async-compute-for-performance-wins?p=1079984#post1079984 > > Thanks, > Dieter > > Am 13.02

[Mesa-dev] [PATCH 26/26] radeonsi: cull primitives with async compute for large draw calls

From: Marek Olšák --- src/gallium/drivers/radeonsi/Makefile.sources |1 + src/gallium/drivers/radeonsi/meson.build |1 + .../radeonsi/si_compute_prim_discard.c| 1537 + src/gallium/drivers/radeonsi/si_debug.c | 32 +- src/gallium/drivers/radeonsi

[Mesa-dev] [PATCH 08/26] winsys/amdgpu: add a parallel compute IB coupled with a gfx IB

From: Marek Olšák --- src/amd/common/ac_gpu_info.c | 6 + src/amd/common/ac_gpu_info.h | 2 + src/gallium/drivers/r600/r600_pipe_common.c | 4 +- src/gallium/drivers/radeon/radeon_winsys.h| 36 ++- src/gallium/drivers/radeonsi/si_fence.c

[Mesa-dev] [PATCH 24/26] radeonsi: add si_vs_prolog_bits::unpack_instance_id_from_vertex_id:1

From: Marek Olšák The prim discard compute shader bakes InstanceID into the output index buffer. --- src/gallium/drivers/radeonsi/si_shader.c | 25 ++-- src/gallium/drivers/radeonsi/si_shader.h | 1 + 2 files changed, 24 insertions(+), 2 deletions(-) diff --git a/src

[Mesa-dev] [PATCH 20/26] radeonsi: allow si_shader_select_with_key to return an optimized shader or fail

From: Marek Olšák If a prim discard compute shader hasn't finished compilation, we don't want to any shader. --- src/gallium/drivers/radeonsi/si_state.h | 7 .../drivers/radeonsi/si_state_shaders.c | 38 +-- 2 files changed, 33 insertions(+), 12 deletions

[Mesa-dev] [PATCH 18/26] radeonsi: add threadgroups_per_cu param into si_get_compute_resource_limits

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_compute.c | 9 ++--- src/gallium/drivers/radeonsi/si_pipe.h| 3 ++- 2 files changed, 8 insertions(+), 4 deletions(-) diff --git a/src/gallium/drivers/radeonsi/si_compute.c b/src/gallium/drivers/radeonsi/si_compute.c index

[Mesa-dev] [PATCH 23/26] radeonsi: add helper si_get_minimum_num_gfx_cs_dwords

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_gfx_cs.c | 8 +--- src/gallium/drivers/radeonsi/si_pipe.h | 11 +++ 2 files changed, 12 insertions(+), 7 deletions(-) diff --git a/src/gallium/drivers/radeonsi/si_gfx_cs.c b/src/gallium/drivers/radeonsi/si_gfx_cs.c index

[Mesa-dev] [PATCH 22/26] radeonsi: add a cs parameter into si_cp_copy_data

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_compute.c | 2 +- src/gallium/drivers/radeonsi/si_cp_dma.c | 9 - src/gallium/drivers/radeonsi/si_perfcounter.c | 2 +- src/gallium/drivers/radeonsi/si_pipe.h| 2 +- src/gallium/drivers/radeonsi/si_state_draw.c | 2

[Mesa-dev] [PATCH 25/26] ac: add LLVM code for triangle culling

From: Marek Olšák --- src/amd/Makefile.sources | 2 + src/amd/common/ac_llvm_cull.c | 275 ++ src/amd/common/ac_llvm_cull.h | 59 src/amd/common/meson.build| 2 + 4 files changed, 338 insertions(+) create mode 100644 src/amd/common

[Mesa-dev] [PATCH 07/26] winsys/amdgpu: reorder chunks, make BO_HANDLES first, IB and FENCE last

From: Marek Olšák --- src/gallium/winsys/amdgpu/drm/amdgpu_cs.c | 36 +++ 1 file changed, 17 insertions(+), 19 deletions(-) diff --git a/src/gallium/winsys/amdgpu/drm/amdgpu_cs.c b/src/gallium/winsys/amdgpu/drm/amdgpu_cs.c index 319741f7d0b..4a588d52930 100644 --- a/src

[Mesa-dev] [PATCH 12/26] radeonsi: return the last part's return value from @wrapper

From: Marek Olšák The primitive discard compute shader will get the position output this way. --- src/gallium/drivers/radeonsi/si_shader.c | 29 +--- 1 file changed, 26 insertions(+), 3 deletions(-) diff --git a/src/gallium/drivers/radeonsi/si_shader.c b/src/gallium

[Mesa-dev] [PATCH 17/26] radeonsi: use pipe_draw_info::instance_count indirectly

From: Marek Olšák It will be modified by compute shader culling. --- src/gallium/drivers/radeonsi/si_state_draw.c | 36 1 file changed, 22 insertions(+), 14 deletions(-) diff --git a/src/gallium/drivers/radeonsi/si_state_draw.c b/src/gallium/drivers/radeonsi

[Mesa-dev] [PATCH 19/26] radeonsi: add a cs parameter into si_cp_release_mem

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_fence.c | 5 ++--- src/gallium/drivers/radeonsi/si_perfcounter.c | 2 +- src/gallium/drivers/radeonsi/si_pipe.h| 2 +- src/gallium/drivers/radeonsi/si_query.c | 6 +++--- src/gallium/drivers/radeonsi/si_state_draw.c | 4

[Mesa-dev] [PATCH 21/26] radeonsi: make some functions non-static

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_state.c | 20 ++- src/gallium/drivers/radeonsi/si_state.h | 15 ++ .../drivers/radeonsi/si_state_shaders.c | 8 3 files changed, 25 insertions(+), 18 deletions(-) diff --git a/src

[Mesa-dev] [PATCH 16/26] radeonsi: make functions for creating LLVM functions non-static

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_shader.c | 30 +-- .../drivers/radeonsi/si_shader_internal.h | 25 2 files changed, 32 insertions(+), 23 deletions(-) diff --git a/src/gallium/drivers/radeonsi/si_shader.c b/src/gallium/drivers

[Mesa-dev] [PATCH 14/26] radeonsi: make si_initialize_compute reusable

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_compute.c | 14 +++--- src/gallium/drivers/radeonsi/si_pipe.h| 1 + 2 files changed, 8 insertions(+), 7 deletions(-) diff --git a/src/gallium/drivers/radeonsi/si_compute.c b/src/gallium/drivers/radeonsi/si_compute.c index

[Mesa-dev] [PATCH 15/26] radeonsi: move si_*_descriptors_idx functions into si_state.h

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_descriptors.c | 14 -- src/gallium/drivers/radeonsi/si_state.h | 14 ++ 2 files changed, 14 insertions(+), 14 deletions(-) diff --git a/src/gallium/drivers/radeonsi/si_descriptors.c b/src/gallium/drivers

[Mesa-dev] [PATCH 01/26] tgsi/scan: add uses_drawid

From: Marek Olšák --- src/gallium/auxiliary/tgsi/tgsi_scan.c | 3 +++ src/gallium/auxiliary/tgsi/tgsi_scan.h | 1 + src/gallium/drivers/radeonsi/si_shader_nir.c | 3 +++ 3 files changed, 7 insertions(+) diff --git a/src/gallium/auxiliary/tgsi/tgsi_scan.c b/src/gallium/auxiliary

[Mesa-dev] [PATCH 04/26] ac: add ac_get_i1_sgpr_mask

From: Marek Olšák --- src/amd/common/ac_llvm_build.c | 16 src/amd/common/ac_llvm_build.h | 2 ++ 2 files changed, 18 insertions(+) diff --git a/src/amd/common/ac_llvm_build.c b/src/amd/common/ac_llvm_build.c index 9395bd1bbda..7adca596943 100644 --- a/src/amd/common

[Mesa-dev] [PATCH 10/26] radeonsi: add si_cp_copy_data

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_compute.c | 21 -- src/gallium/drivers/radeonsi/si_cp_dma.c | 28 +++ src/gallium/drivers/radeonsi/si_perfcounter.c | 13 ++--- src/gallium/drivers/radeonsi/si_pipe.h| 3 ++ src/gallium

[Mesa-dev] [PATCH 09/26] winsys/amdpgu: retry after ENOMEM to work around a GDS memory management bug

From: Marek Olšák If GDS/GWS/OA resources are being used by IBs that are currently busy, the kernel driver returns -ENOMEM instead of waiting until those resources are idle. --- src/gallium/winsys/amdgpu/drm/amdgpu_cs.c | 17 + 1 file changed, 13 insertions(+), 4 deletions

[Mesa-dev] [PATCH 13/26] radeonsi: extract COMPUTE_RESOURCE_LIMITS code into a helper

From: Marek Olšák --- src/gallium/drivers/radeonsi/si_compute.c | 32 ++- src/gallium/drivers/radeonsi/si_pipe.h| 3 +++ 2 files changed, 23 insertions(+), 12 deletions(-) diff --git a/src/gallium/drivers/radeonsi/si_compute.c b/src/gallium/drivers/radeonsi

[Mesa-dev] [PATCH 06/26] winsys/amdgpu: make IBs writable and expose their address

From: Marek Olšák --- src/gallium/drivers/radeon/radeon_winsys.h | 1 + src/gallium/winsys/amdgpu/drm/amdgpu_cs.c | 5 - 2 files changed, 5 insertions(+), 1 deletion(-) diff --git a/src/gallium/drivers/radeon/radeon_winsys.h b/src/gallium/drivers/radeon/radeon_winsys.h index 822d2e88ee2

[Mesa-dev] [PATCH 11/26] radeonsi: use pipe_draw_info::prim and primitive_restart indirectly

From: Marek Olšák so that the fields can be changed by the driver. --- src/gallium/drivers/radeonsi/si_state_draw.c | 73 +++- 1 file changed, 41 insertions(+), 32 deletions(-) diff --git a/src/gallium/drivers/radeonsi/si_state_draw.c b/src/gallium/drivers/radeonsi

[Mesa-dev] [PATCH 05/26] ac: add REWIND and GDS registers to register headers

From: Marek Olšák --- src/amd/common/sid.h | 16 1 file changed, 16 insertions(+) diff --git a/src/amd/common/sid.h b/src/amd/common/sid.h index 5c8eee0124d..3c0b7001d2b 100644 --- a/src/amd/common/sid.h +++ b/src/amd/common/sid.h @@ -206,6 +206,7 @@ #define PKT3_RELEASE_MEM

[Mesa-dev] [PATCH 02/26] ac: add radeon_info::marketing_name, replacing the winsys callback

From: Marek Olšák --- src/amd/common/ac_gpu_info.c | 2 ++ src/amd/common/ac_gpu_info.h | 1 + src/gallium/drivers/r600/r600_pipe_common.c | 13 + src/gallium/drivers/radeon/radeon_winsys.h| 2 -- src/gallium/drivers/radeonsi/si_get.c

[Mesa-dev] [PATCH 00/26] RadeonSI: Primitive culling with async compute

Hi, This patch series uses async compute to do primitive culling before the vertex shader. It significantly improves performance for applications that use a lot of geometry that is invisible because primitives don't intersect sample points or there are a lot of back faces, etc. It passes

[Mesa-dev] [PATCH 03/26] ac: add radeon_info::is_pro_graphics

From: Marek Olšák --- src/amd/common/ac_gpu_info.c | 3 +++ src/amd/common/ac_gpu_info.h | 1 + 2 files changed, 4 insertions(+) diff --git a/src/amd/common/ac_gpu_info.c b/src/amd/common/ac_gpu_info.c index 4e3190015e7..6971e4f0a8e 100644 --- a/src/amd/common/ac_gpu_info.c +++ b/src/amd

Re: [Mesa-dev] [PATCH] radeonsi: Fix guardband computation for large render targets

Pushed, thanks! Marek On Tue, Feb 12, 2019 at 2:55 PM Oscar Blumberg wrote: > Stop using 12.12 quantization for viewports that are not contained in > the lower 4k corner of the render target as the hardware needs to keep > both absolute and relative coordinates representable. > --- >

[Mesa-dev] [PATCH 2/2] gallium/u_tests: use a compute-only context to test GCN compute ring

From: Marek Olšák --- src/gallium/auxiliary/util/u_tests.c | 8 +++- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/src/gallium/auxiliary/util/u_tests.c b/src/gallium/auxiliary/util/u_tests.c index 365d4fa8f17..7b87337bb98 100644 --- a/src/gallium/auxiliary/util/u_tests.c

[Mesa-dev] [PATCH 1/2] radeonsi: always use compute rings for clover on CI and newer (v2)

From: Marek Olšák initialize all non-compute context functions to NULL. v2: fix SI --- src/gallium/drivers/radeonsi/si_blit.c| 14 ++- src/gallium/drivers/radeonsi/si_clear.c | 7 +- src/gallium/drivers/radeonsi/si_compute.c | 15 +-- src/gallium/drivers/radeonsi

Re: [Mesa-dev] [PATCH 1/2] radeonsi: use compute for clear_render_target when possible

It's pushed already. Marek On Mon, Feb 11, 2019, 11:41 PM Dieter Nützel Maybe rebase? > > Dieter > > Am 24.01.2019 00:28, schrieb Marek Olšák: > > From: Sonny Jiang > > > > Signed-off-by: Sonny Jiang > > Signed-off-by: Marek Olšák > > --- &g

<    1   2   3   4   5   6   7   8   9   10   >