nv: use regcount from meta (#10503)

This commit is contained in:
nimlgen
2025-05-24 14:14:33 +03:00
committed by GitHub
parent 82b444796d
commit c472ab636c

View File

@@ -220,13 +220,13 @@ class NVProgram(HCQProgram):
self.constbufs: dict[int, tuple[int, int]] = {0: (0, 0x160)} # dict[constbuf index, tuple[va_addr, size]]
for sh in sections:
if sh.name == f".nv.shared.{self.name}": self.shmem_usage = round_up(0x400 + sh.header.sh_size, 128)
if sh.name == f".text.{self.name}":
self.prog_addr, self.prog_sz, self.regs_usage = self.lib_gpu.va_addr+sh.header.sh_addr, sh.header.sh_size, max(sh.header.sh_info>>24, 16)
if sh.name == f".text.{self.name}": self.prog_addr, self.prog_sz = self.lib_gpu.va_addr+sh.header.sh_addr, sh.header.sh_size
elif m:=re.match(r'\.nv\.constant(\d+)', sh.name): self.constbufs[int(m.group(1))] = (self.lib_gpu.va_addr+sh.header.sh_addr, sh.header.sh_size)
elif sh.name.startswith(".nv.info"):
for typ, param, data in self._parse_elf_info(sh):
if sh.name == f".nv.info.{name}" and param == 0xa: cbuf0_size = struct.unpack_from("IH", data)[1] # EIATTR_PARAM_CBANK
elif sh.name == ".nv.info" and param == 0x12: self.lcmem_usage = struct.unpack_from("II", data)[1] + 0x240 # EIATTR_MIN_STACK_SIZE
elif sh.name == ".nv.info" and param == 0x2f: self.regs_usage = struct.unpack_from("II", data)[1] # EIATTR_REGCOUNT
# Ensure device has enough local memory to run the program
self.dev._ensure_has_local_memory(self.lcmem_usage)