| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
This change will double cache size for branches which have a lower
LP_MAX_SHADER_VARIANTS limit (it will not do anything on master).
The reason is that nowadays shaders tend to be quite a bit larger than they
were (they were big when llvmpipe didn't have a fs loop, got much smaller with
that loop, and since then have gradually increased quite a bit though still
smaller than without the fs loop for various reasons - among them being d3d10
compliance, usage of 8-wide vectors, non-swizzled blend code). Thus effectively
less shaders would be cached (unless they were very small and the variant limit
was hit first). Also, since we're getting rid of the IR nowadays, the cached
shaders shouldn't need all that much memory actually.
|
| |
|
|
|
|
|
|
|
|
| |
When the limit was changed to be defined in terms of LP_MAX_SHADER_VARIANTS
(75f1fea14f524ef05e980d825fda3ae226ae2ffe) when it was increased, this
inadvertently lowered the limit in some branches (that have a lower
LP_MAX_SHADER_VARIANTS number) when merged. So, make sure the limit is always
at least the number it once was.
Reviewed-by: Jose Fonseca <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The previous limit of of 128*1024 was reported to cause frequent recompiles
in some apps due to shader variant thrashing on IRC in some apps leading
to noticeable lags.
Note that the LP_MAX_SHADER_VARIANTS limit (1024) was more or less impossible
to reach, since even simple fragment shaders without texturing (glxgears) used
more than twice than 128 instructions, hence the instruction limit would have
always been reached first (excluding things like trivial shaders not writing
color). Even with the new limit it is VERY likely the instruction limit is hit
first.
Should help with such lags due to recompiles (though other shader types have
their own limits, LP_MAX_SETUP_VARIANTS and DRAW_MAX_SHADER_VARIANTS, in
particular the latter seems a bit small (128)).
Reviewed-by: Brian Paul <[email protected]>
|
| |
|
|
|
|
|
| |
These should just work, required by d3d10. Too large resources will
get thrown out separately anyway.
Reviewed-by: Brian Paul <[email protected]>
|
| |
|
|
|
|
|
| |
On the mesa-users list, Burlen Loring reported a speed-up with 16 cores
and his test/app.
Reviewed-by: Roland Scheidegger <[email protected]>
|
| |
|
|
|
|
|
|
|
|
| |
This adds array (1d,2d) texture support to llvmpipe.
Though probably should do something about 1d array textures requiring gobs
of memory (this issue is not strictly limited to arrays but it is probably
worse there).
Initial code by Jakob Bornecrantz <[email protected]>
Reviewed-by: José Fonseca <[email protected]>
|
| |
|
|
|
|
|
|
| |
Before, the limit was 8K. For 32-bit RGBA that would be require 1.5 GB
of memory (w/out mipmaps). That's well beyond the LP_MAX_TEXTURE_SIZE
of 1GB.
Reviewed-by: Jose Fonseca <[email protected]>
|
| |
|
|
|
|
|
|
|
|
| |
But cap the size in bytes, to avoid depleting the whole system memory,
with humongus textures.
Tested with max-texture-size piglit test.
Reviewed-by: Roland Scheidegger <[email protected]>
Reviewed-by: Brian Paul <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Number of fragment shader variants is not very representative of the
memory used by LLVM, neither is number of shader instructions, as often
texture sampling constitutes most of the generated code.
This change adds an additional trim criteria: least recently used
fragment shader variants will be freed until the total number of LLVM IR
instruction falls below a specified threshold.
Reviewed-by: Brian Paul <[email protected]>
|
| |
|
|
|
| |
Basically no change relative to hard-coded version, but this will
be useful for other changes later.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
llvmpipe can create a large number of shader variants for a single shader
(which are quite big), and they were only ever deleted if the shader itself
was deleted. This is especially apparent in things like glean
blendFunc where a new variant is created for every different subtest, chewing
up all memory.
This change limits the numbers of fragment shader variants (for all shaders)
which are kept around to a fixed number. If that would be exceeded a fixed
portion of the cached variants is deleted (since without tracking the used
variants this involves flushing we don't want to delete only one).
Always the least recently used variants (from all shaders together) are
deleted.
For now this is all per-context.
Both the number of how many variants are cached (1024) as well as how many
will be deleted at once (1/4 of the cache size) are just rough guesses and
subject to further optimization.
|
| |
|
|
|
|
|
|
| |
When the size of the scene (binned data plus referenced resources/textures)
exceeds LP_MAX_SCENE_SIZE flush/render the scene. This could be improved
in various ways but is a good start.
Fixes piglit streaming-texture-leak test.
|
| | |
|
| | |
|
| | |
|
|
|
Collect the various driver limits in one place since.
|
