github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-01-27 07:48:07 -05:00
Code Issues Packages Projects Releases Wiki Activity

fix: correct gfx10 ctl stack size (#9384)

Browse Source
This commit is contained in:
wozeparrot
2025-03-09 00:03:20 -05:00
committed by GitHub
parent 456697d0be
commit b6fe5ab4dd
1 changed files with 2 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
tinygrad/runtime/ops_amd.py
View File

@@ -584,7 +584,8 @@ class AMDDevice(HCQCompiled):
sgrp_size_per_cu, lds_size_per_cu, hwreg_size_per_cu = 0x4000, 0x10000, 0x1000
vgpr_size_per_cu = 0x60000 if self.target in {110000, 110001, 120000, 120001} else 0x40000
wg_data_size = round_up((vgpr_size_per_cu + sgrp_size_per_cu + lds_size_per_cu + hwreg_size_per_cu) * (self.max_cu_id + 1), mmap.PAGESIZE)
ctl_stack_size = round_up(12 * (self.max_cu_id + 1) * (self.max_wave_id + 1) + 8 + 40, mmap.PAGESIZE) if self.target//10000 != 10 else 0x7000
ctl_stack_size = round_up(12 * (self.max_cu_id + 1) * (self.max_wave_id + 1) + 8 + 40, mmap.PAGESIZE)
if self.target//10000 == 10: ctl_stack_size = min(ctl_stack_size, 0x7000)
debug_memory_size = round_up((self.max_cu_id + 1) * (self.max_wave_id + 1) * 32, 64)
self.compute_queue = self.create_queue(kfd.KFD_IOC_QUEUE_TYPE_COMPUTE, 0x800000, ctx_save_restore_size=wg_data_size + ctl_stack_size,