Image and sampler table capacity came only from the regular descriptor
limits. MoltenVK reports Metal's argument table limits there (256 sampled
images and 16 samplers per stage), below the typed heaps the tables are
reserved on top of, so both capacities were 0 and every table draw or
dispatch was refused as if the device lacked descriptor indexing. Its
update-after-bind limits are 1000000 sampled images and 500000 samplers.
- descriptorBindingSampledImageUpdateAfterBind is queried and enabled, and
the update-after-bind limits are read when either it or the storage
buffer feature is present.
- A table whose regular capacity is 0 takes its capacity from the
update-after-bind limits. Its binding is update-after-bind and its set
layout and pool are update-after-bind ones, in ShaderResources and in
DescribeSetLayouts alike. A table that fits the regular limits is
unchanged.
- An update-after-bind set layout and its allocation are checked against
the update-after-bind sampled image and sampler limits when the device
has descriptorBindingSampledImageUpdateAfterBind. Vulkan counts the
descriptors of such sets against those limits in place of the regular
ones, and requires them to be at least the regular ones only when the
feature is supported, so without it, and for storage images, the
regular limits are kept as before. Storage buffers keep their own
update-after-bind flag.
- Compute pipeline capture and replay accept update-after-bind sampled
image and sampler bindings.
Outputs whose input carries no PT_GNU_STACK get one appended with
PF_R|PF_W and alignment 16 so the kernel does not fall back to an
executable stack marking.
RequiredProgramHeaderCount and WriteLayout now derive from one shared
header list, so the reserved slot and the written header cannot
disagree at any PhNum boundary, when a free input slot should be
reused, or when the RELRO coverage guard reduces the kept count.
Space exhaustion beyond the 16-bit header limit fails loudly instead
of silently skipping the header.
The program header table moves when the layout needs more slots than
the input provides, so the stub lookup must follow the output e_phoff
instead of assuming the table stays at file offset 64.
Every OpGroupNonUniformBallot now goes through SpirvModule::AddSubgroupBallot.
A repeated ballot of the same predicate in the same block becomes an
OpCopyObject of the first ballot. Within one block the active invocations do
not change, so the result is identical. The cache is cleared at every label,
function boundary, function call and demote, and helper functions keep their
own cache.
When SPIRV-Tools is enabled, block merging and specialization can still put
two ballots of one predicate into the same block after emission, so
ValidateAndOptimizeSpirv collapses repeated block ballots the same way on
every module it returns, optimized or not.
NVIDIA 580's SPIR-V compiler segfaulted (libnvidia-gpucomp+0xe3facc) on GTA V
compute shaders that ballot one predicate many times in a block. One such
shader had 163 repeated ballots across three blocks. With them collapsed, both
crashing captures (compute-fe506552cec99bfc and compute-c02e5e3f35010fa5)
compile.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
RDNA 2 (RX 6700 XT, native image_store) and Apple M2 store the level
nearest to the exact product x * scale, not the nearest even level of
its f32 rounding. The two differ when fl32(x * scale) lands exactly on
a .5: the fma residual of the product now breaks that tie by its sign.
The residual is decorated NoContraction, without it the Intel UHD 770
does not evaluate the fma exactly. The test expects the level nearest
to the product computed in double, exact for these inputs.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01XBXM4pgkMHYEtqN1EQni2P
sceVideoOutOpen with bus type 0x20, the PS VR2 output PPSA05666 opens as
SCE_VIDEO_OUT_BUS_TYPE_HMD2_MAIN after sceHmd2ReprojectionEnableVrMode, threw
VIDEO_OUT_ERROR_INVALID_VALUE and aborted the title. The console returns that
code to a process that is not a game, with the headset connected and worn or
not, for users 255 and 0 and for a null or zeroed open parameter, so the bus
is answered that way before the other argument checks and the title carries
on with the error as its handle, as it does on the console without a headset.
The debt entry records the sequence the title follows on that handle
(RegisterBuffers2 with category 16 and 4000x2040 descriptors that point at
Hmd2 per-eye textures, ConfigureOutput and IsOutputSupported with 0x2000F, a
GPU wait on buffer 0), which a VR backend will have to model, and the test
covers the bus and the HMD output modes 0x20001 and 0x2000C to 0x20010, which
the console and this implementation answer with VIDEO_OUT_ERROR_UNKNOWN_OUTPUT_MODE
on the TV bus.
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Filled rectangles (VGT_PRIMITIVE_TYPE 7 and 17) use the same single-sample MSAA_ENABLE path as triangles; PPSA03541 stops on a rect-list draw otherwise.
Co-authored-by: Cursor <[email protected]>
Use representable viewport endpoints only when the draw has no depth attachment and its pixel shader does not read PositionZ. Preserve XY coverage, clip-space mode, depth clamping, and validation of non-finite inputs.
Balatro (PPSA21402) reaches rendering on Apple M1 with MoltenVK after this change instead of stopping at VK_EXT_depth_range_unrestricted. Depth-dependent draws retain their original behavior.
agc_driver_flip_stores_small_results waits 2 s for the flip and the
stored texels. That window also holds the driver's start and the
compute program's compilation at dispatch, so on a loaded host the
flip arrives late and the test reports that it did not reach the GPU.
Wait 10 s, as the other execution tests do.
Dean