Kenneth Graunke 837c441acb intel/nir: Don't needlessly split u2f16 for nir_type_uint32 ... Commit f695a9fed2 moved the 64-bit float <-> 16-bit float conversion splitting into a core NIR pass, so the code remaining here is only needed for 64-bit integer types. Presumably in an attempt to remove the float handling, it replaced simple bit_size == 64 checks with this expression: (full_type & (nir_type_int64 | nir_type_uint64)) I believe that the intended expression was: (full_type == nir_type_int64 || full_type == nir_type_uint64) Unfortunately, the former is incorrect. Any integer or unsigned NIR type would trigger the former expression. For example: nir_type_uint32 & (nir_type_int64 | nir_type_uint64) => nir_type_uint This meant that we were splitting e.g. u2f16 on 32-bit unsigned types into u2f32 and f2f16, when we can easily natively handle that case. To fix this, we go back to simple bit_size == 64 checks. This pass is already run after nir_lower_fp16_casts which will split the float case, so we will never see it here. fossil-db on Alchemist shows a -1.14% reduction in affected shaders for google-meet-clvk shaders. In another ChromeOS workload, it improves performance by around 8% on Meteorlake. Thanks to Sushma Venkatesh Reddy for finding this performance issue! Fixes: f695a9fed2 ("intel/compiler: use nir_lower_fp16_casts") Reviewed-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Ivan Briano <ivan.briano@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30091>		2024-07-11 02:37:05 -07:00
..
blorp	blorp: Ensure MSAA fast clear in correct modes (xe2)	2024-07-02 19:03:19 +00:00
ci	ci: simplify setting .no-auto-retry now that it isn't bundled with unrelated rules:	2024-07-07 19:31:44 +00:00
common	intel/common: fix building error in intel_common.c	2024-07-02 23:35:26 +00:00
compiler	intel/nir: Don't needlessly split u2f16 for nir_type_uint32	2024-07-11 02:37:05 -07:00
decoder	intel/decoder: Add intel_print_group_custom_spacing()	2024-04-24 17:07:50 +00:00
dev	intel/dev: Replace intel_device_info::apply_hwconfig by a gfx version check	2024-07-03 22:17:37 +00:00
ds	intel/ds: remove duplicate arguments	2024-07-03 21:10:13 +00:00
genxml	build: pass licensing information in SPDX form	2024-06-29 12:42:49 -07:00
isl	isl: Add some formats not covered in CMF table (xe2)	2024-07-02 19:03:19 +00:00
nullhw-layer	build: pass licensing information in SPDX form	2024-06-29 12:42:49 -07:00
perf	intel/perf/xe: Fix free pointer location in xe_add_config()	2024-07-05 00:25:03 -07:00
shaders	meson: use glslang --depfile argument when possible	2024-05-20 17:34:17 +00:00
tools	build: pass licensing information in SPDX form	2024-06-29 12:42:49 -07:00
vulkan	anv: fix u_trace on < Gfx12.0	2024-07-03 21:10:13 +00:00
vulkan_hasvk	hasvk: pass anv_address to predicate helper	2024-07-03 21:10:13 +00:00
meson.build	build: pass licensing information in SPDX form	2024-06-29 12:42:49 -07:00