[Mesa-dev] [PATCH 0/7] Reduce worst-case memory usage in NIR

Work on soft fp64 support has uncovered some pessimal memory usage inside NIR. Code that uses soft fp64 hits two problems: huge numbers of basic blocks and huge numbers of values. Several places in the compiler use straight forward, naive data structures that consume massive amounts of memory in

[Mesa-dev] [PATCH 3/7] util/slab: Add function to flush allocations from a child pool

From: Ian Romanick Ralloc has a feature that all allocations from a temporary memory context can be whisked away in a single call without fear of leaks. As the slab allocator is designed for use in multhreaded scenarios with a child pool per CPU, it lacks this feature. However, many users will

[Mesa-dev] [PATCH 2/7] nir/phi_builder: Use hash table to store [value, block] -> def mapping

From: Ian Romanick Changes in peak memory usage according to Valgrind massif: mean soft fp64 using uint64: 5,499,881,802 => 1,343,998,123 gfxbench5 aztec ruins high 11:62,415,414 =>62,415,414 deus ex mankind divided 148: 62,317,965 =>62,317,965 deus ex mankind divided 2890:

[Mesa-dev] [PATCH 6/7] nir: Add function to release memory used for liveness tracking

From: Ian Romanick Signed-off-by: Ian Romanick --- src/compiler/nir/nir.h | 2 ++ src/compiler/nir/nir_liveness.c | 17 + 2 files changed, 19 insertions(+) diff --git a/src/compiler/nir/nir.h b/src/compiler/nir/nir.h index 74c700026ad..ac20864c6ca 100644 --- a/src/com

[Mesa-dev] [PATCH 5/7] nir: Fix holes in nir_instr

From: Ian Romanick Found using pahole. Changes in peak memory usage according to Valgrind massif: mean soft fp64 using uint64: 1,343,998,123 => 1,342,766,051 gfxbench5 aztec ruins high 11:62,415,414 =>62,369,974 deus ex mankind divided 148: 62,317,965 =>62,290,717 deus ex man

Re: [Mesa-dev] [PATCH 03/22] nir/spirv: initial handling of OpenCL.std extension opcodes

On Tue, Nov 13, 2018 at 9:48 AM Karol Herbst wrote: > Not complete, mostly just adding things as I encounter them in CTS. But > not getting far enough yet to hit most of the OpenCL.std instructions. > > Anyway, this is better than nothing and covers the most common builtins. > > Signed-off-by: Ka

Re: [Mesa-dev] [PATCH] virgl: don't mark buffers as unclean after a write

On Tue, Nov 20, 2018 at 3:34 PM Ilia Mirkin wrote: > > I'm unfamiliar with the surrounding code, but wanted to point out that > a single buffer may be mapped as a SSBO, UBO, TBO, etc. > transfer_map/unmap isn't really aware of how the buffer is used. You > can, however, keep track of such bindings

Re: [Mesa-dev] [PATCH 01/10] mesa: Rename gl_vertex_array_object::_Enabled -> Enabled.

Good Morning, On Tuesday, 20 November 2018 23:49:28 CET Marek Olšák wrote: > For the series: > > Reviewed-by: Marek Olšák Thanks! and pushed now! best Mathias ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/m

<    1   2