commaai
/
tinygrad
mirror of https://github.com/commaai/tinygrad.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

feat: use more correct reg for local dims (#6048)

This commit is contained in:
wozeparrot 2024-08-12 11:15:37 -07:00 committed by GitHub
parent 529832d223
commit dc2617bffd
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
extra/mockgpu/amd/amdgpu.py
View File

@ -11,7 +11,7 @@ SUB = PACKET3_SET_SH_REG_START - BASE_ADDR
regCOMPUTE_PGM_LO = 0x1bac - SUB
regCOMPUTE_USER_DATA_0 = 0x1be0 - SUB
regCOMPUTE_START_X = 0x1ba4 - SUB
regCOMPUTE_NUM_THREAD_X = 0x1ba7 - SUB
CACHE_FLUSH_AND_INV_TS_EVENT = 0x14
@ -153,7 +153,7 @@ class PM4Executor(AMDQueue):
prg_addr = (self.gpu.regs[regCOMPUTE_PGM_LO] + (self.gpu.regs[regCOMPUTE_PGM_LO + 1] << 32)) << 8
args_addr = self.gpu.regs[regCOMPUTE_USER_DATA_0] + (self.gpu.regs[regCOMPUTE_USER_DATA_0 + 1] << 32)
lc = [self.gpu.regs[i] for i in range(regCOMPUTE_START_X+3, regCOMPUTE_START_X+6)]
lc = [self.gpu.regs[i] for i in range(regCOMPUTE_NUM_THREAD_X, regCOMPUTE_NUM_THREAD_X+3)]
prg_sz = 0
for st,sz in self.gpu.mapped_ranges: