[BugTIR] fix thread_sync occurs in letstmt #16454
Conversation
|
cc @vinx13 @spectrometerHBH please spend a little time do a review if you see this |
src/tir/transforms/storage_access.cc
Outdated
|
|
||
| void StorageAccessVisitor::VisitStmt_(const LetStmtNode* op) { | ||
| allow_append_ = true; | ||
| curr_stmt_.access.clear(); |
There was a problem hiding this comment.
shouldn't call clear here, it can only be used after finishing handling a statement
There was a problem hiding this comment.
can you check my new solution?
| A_shared_1[ax0] = A[blockIdx_x * 512 + ax0] | ||
| in_thread_A_temp_1 = T.Buffer((1,), data=in_thread_A_temp, scope="local") | ||
| in_thread_A_temp_1[0] = T.float32(0) | ||
| with T.LetStmt(in_thread_A_temp_1[0] + A_shared_1[threadIdx_x]) as A_temp: |
There was a problem hiding this comment.
in case here, before take change, StorageAccessVisitor will call VisitExpr_(const BufferLoadNode* op) directly (not called by father call like VisitStmt_(const BufferStoreNode* op) traverse child), so ICHECK(allow_append_) << op << " " << scope.to_string(); will return error.
if we take VisitStmt_(const LetStmtNode* op) and traverse child of LetStmtNode
(case here it will traverse BufferLoadNode A_shared which will increase curr_stmt_.access by 1, and then visit BufferStoreNode in_thread_A_temp , ICHECK_EQ(curr_stmt_.access.size(), 0U); will return error)
Do you have any insights on how to solve this problem? @vinx13
3 participants
See original discuss
LayerNorm Error in thread_storage_sync when read x into shared memory
I try to read x into shared memory to accelerate layernorm, script here 1
but error occurs in pass thread_sorage_sync pass.I found it is because the error in lower LetStmt ,lowered script