nir/lower_mem_access_bit_sizes: increase chunk limit

Not sure about creating u64vec16 loads, but creating unaligned loads is
possible with opt_if_rewrite_uniform_uses.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37953>
This commit is contained in:
Rhys Perry 2025-10-15 15:31:34 +01:00 committed by Marge Bot
parent e89b22280f
commit 64ec757688

View file

@ -181,9 +181,9 @@ lower_mem_load(nir_builder *b, nir_intrinsic_instr *intrin,
return false;
/* Otherwise, we have to break it into chunks. We could end up with as
* many as 32 chunks if we're loading a u64vec16 as individual dwords.
* many as 128 chunks if we're loading a u64vec16 as individual bytes.
*/
nir_def *chunks[32];
nir_def *chunks[128];
unsigned num_chunks = 0;
unsigned chunk_start = 0;
while (chunk_start < bytes_read) {