Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fix LARS wrong thread synchronization #36431

Merged
merged 1 commit into from
Oct 14, 2021
Merged

Fix LARS wrong thread synchronization #36431

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
41 changes: 31 additions & 10 deletions paddle/fluid/operators/optimizers/lars_momentum_op.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -165,8 +165,10 @@ __global__ void L2NormKernel(
int tid = threadIdx.x + blockDim.x * blockIdx.x;
int grid_stride = LARS_BLOCK_SIZE * gridDim.x;
const MT rescale_pow = rescale_grad * rescale_grad;
s_buffer[0] = static_cast<MT>(0);
s_buffer[1] = static_cast<MT>(0);
if (threadIdx.x == 0) {
s_buffer[0] = static_cast<MT>(0);
s_buffer[1] = static_cast<MT>(0);
}
MT p_tmp = static_cast<MT>(0);
MT g_tmp = static_cast<MT>(0);

Expand All @@ -175,8 +177,12 @@ __global__ void L2NormKernel(
p_tmp = static_cast<MT>(p_data[tid]);
g_tmp = static_cast<MT>(g_data[tid]);
}
s_buffer[0] += math::blockReduceSum<MT>(p_tmp * p_tmp, FINAL_MASK);
s_buffer[1] += math::blockReduceSum<MT>(g_tmp * g_tmp, FINAL_MASK);
MT tmp0 = math::blockReduceSum<MT>(p_tmp * p_tmp, FINAL_MASK);
MT tmp1 = math::blockReduceSum<MT>(g_tmp * g_tmp, FINAL_MASK);
if (threadIdx.x == 0) {
s_buffer[0] += tmp0;
s_buffer[1] += tmp1;
}
} else {
/* Avoid occupy too much temp buffer. Slice the whole data into 2 parts,
the front of data whose quantity is excatly multiple of grid-thread
Expand All @@ -185,8 +191,12 @@ __global__ void L2NormKernel(
p_tmp = static_cast<MT>(p_data[tid]);
g_tmp = static_cast<MT>(g_data[tid]);
tid += grid_stride;
s_buffer[0] += math::blockReduceSum<MT>(p_tmp * p_tmp, FINAL_MASK);
s_buffer[1] += math::blockReduceSum<MT>(g_tmp * g_tmp, FINAL_MASK);
MT tmp0 = math::blockReduceSum<MT>(p_tmp * p_tmp, FINAL_MASK);
MT tmp1 = math::blockReduceSum<MT>(g_tmp * g_tmp, FINAL_MASK);
if (threadIdx.x == 0) {
s_buffer[0] += tmp0;
s_buffer[1] += tmp1;
}
__syncthreads();
}
MT p_val = 0;
Expand All @@ -195,8 +205,12 @@ __global__ void L2NormKernel(
p_val = static_cast<MT>(p_data[tid]);
g_val = static_cast<MT>(g_data[tid]);
}
s_buffer[0] += math::blockReduceSum<MT>(p_val * p_val, FINAL_MASK);
s_buffer[1] += math::blockReduceSum<MT>(g_val * g_val, FINAL_MASK);
MT tmp0 = math::blockReduceSum<MT>(p_val * p_val, FINAL_MASK);
MT tmp1 = math::blockReduceSum<MT>(g_val * g_val, FINAL_MASK);
if (threadIdx.x == 0) {
s_buffer[0] += tmp0;
s_buffer[1] += tmp1;
}
}
__syncthreads();

Expand All @@ -208,8 +222,15 @@ __global__ void L2NormKernel(
cg->sync(); // Grid sync for writring partial result to gloabl memory
MT p_part_sum = threadIdx.x < gridDim.x ? p_buffer[threadIdx.x] : 0;
MT g_part_sum = threadIdx.x < gridDim.x ? g_buffer[threadIdx.x] : 0;
*p_n = Sqrt(math::blockReduceSum<MT>(p_part_sum, FINAL_MASK));
*g_n = Sqrt(rescale_pow * math::blockReduceSum<MT>(g_part_sum, FINAL_MASK));
MT tmp0 = math::blockReduceSum<MT>(p_part_sum, FINAL_MASK);
MT tmp1 = math::blockReduceSum<MT>(g_part_sum, FINAL_MASK);
if (threadIdx.x == 0) {
s_buffer[0] = tmp0;
s_buffer[1] = tmp1;
}
__syncthreads();
*p_n = Sqrt(s_buffer[0]);
*g_n = Sqrt(rescale_pow * s_buffer[1]);
#endif
}

Expand Down