Age | Commit message (Collapse) | Author |
|
|
|
We have two consumers of relocations. One is static state buffers, which
want the same relocation every time. The other is the batchbuffer, which gets
thrown out immediately after submit. This lets us reduce repeated computation
for static state buffers, and clean up the code by moving relocations nearer
to where the state buffer is computed.
|
|
|
|
|
|
The user-space suballocator that was used avoided relocation computations by
using the general and surface state base registers and allocating those types
of buffers out of pools built on top of single buffer objects. It also
avoided calls into the buffer manager for these small state allocations, since
only one buffer object was being used.
However, the buffer allocation cost appears to be low, and with relocation
caching, computing relocations for buffers is essentially free. Additionally,
implementing the suballocator required a don't-fence-subdata flag to disable
waiting on buffer maps so that writing new data didn't block on rendering using
old data, and careful handling when mapping to update old data (which we need
to do for unavoidable relocations with FBOs). More importantly, when the
suballocator filled, it had no replacement algorithm and just threw out all
of the contents and forced them to be recomputed, which is a significant cost.
This is the first step, which just changes the buffer type, but doesn't yet
improve the hash table to not result in full recompute on overflow. Because
the buffers are all allocated out of the general buffer allocator, we can
no longer use the general/surface state bases to avoid relocations, and they
are set to 0 instead.
|
|
Putting the bufmgr in the screen is not thread-safe since the emit_reloc
changes. It also led to a significant performance hit from pthread usage
for the attempted thread-safety (up to 12% of a cpu spent on refcounting
protection in single-threaded 965). The motivation had been to allow
multi-context bufmgr sharing in classic mode, but it wasn't worth the cost.
|
|
This is currently believed to work but be a significant performance loss.
Performance recovery should be soon to follow.
The dri_bo_fake_disable_backing_store() call was added to allow backing store
disable like bufmgr_fake.c did, which is a significant performance win (though
it's missing the no-fence-subdata part).
This commit is a squash merge of the 965-ttm branch, which had some history
I wanted to avoid pulling due to noisiness and brokenness at many points
for git-bisecting.
|
|
Conflicts:
src/mesa/drivers/dri/i965/brw_sf.h
src/mesa/drivers/dri/i965/intel_context.c
|
|
In the process, fix some alignment issues:
- Scratch space allocation was aligned into units of 1KB, while the allocation
wanted units of bytes, so we never allocated enough space for scratch.
- GRF register count was programmed as ALIGN(val - 1, 16) / 16 instead of
ALIGN(val, 16) / 16 - 1, which overcounted for val != 16n+1.
|
|
This reverts commit b2f1aa2389473ed09170713301b042661d70a48e.
Somehow I ended up with my branch's save-this-while-I-work-on-master commit
actually on master.
|
|
|
|
|
|
most of the sample working with some small modification
|
|
vertex/fragment programs provided as const.
bmSetFenceLock should return bmSetFence value.
|
|
Signed-off-by: Keith Packard <keithp@neko.keithp.com>
|
|
a multiple of alignment to avoid accumulating unusable free blocks.
|
|
we render. Currenly requires that some state be re-examined after
every LOCK_HARDWARE().
|
|
This driver comes from Tungsten Graphics, with a few further modifications by
Intel.
|