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

Refactor prune identity as much as possible #8849

Merged
merged 5 commits into from
Aug 5, 2022
Merged

Refactor prune identity as much as possible #8849

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
d402b88
Prune a line of parallel cast ops
Yipeng1994 Aug 4, 2022
a8e5c16
Avoid repeated pruning
Yipeng1994 Aug 4, 2022
0631c5b
Code clean up
Yipeng1994 Aug 4, 2022
d77610a
Remove identity op
Yipeng1994 Aug 4, 2022
d7873a4
Update oneflow/core/job_rewriter/auto_parallel.cpp
Yipeng1994 Aug 5, 2022
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
35 changes: 25 additions & 10 deletions oneflow/core/job_rewriter/auto_parallel.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@ See the License for the specific language governing permissions and
limitations under the License.
*/
#include <chrono>
#include "oneflow/core/common/hash_container.h"
#include "oneflow/core/common/util.h"
#include "oneflow/core/job_rewriter/job_pass.h"
#include "oneflow/core/job/job.pb.h"
Expand Down Expand Up @@ -85,26 +86,38 @@ Maybe<void> AutoParallelPass::RemoveParallelCastOps(Job* job) const {
return op_conf.has_user_conf()
&& (op_conf.user_conf().op_type_name() == "parallel_cast"
|| op_conf.user_conf().op_type_name() == "hierarchical_parallel_cast"
|| op_conf.user_conf().op_type_name() == "hierarchical_parallel_cast_like");
|| op_conf.user_conf().op_type_name() == "hierarchical_parallel_cast_like"
|| op_conf.user_conf().op_type_name() == "identity");
};
std::vector<std::string> del_op_names;
op_graph.ForEachNode([&](const OpNode* op_node) {
HashSet<std::string> del_op_name_set;
std::function<void(const OpNode*)> Try2Delete = [&](const OpNode* op_node) {
if (del_op_name_set.find(op_node->op().op_name()) != del_op_name_set.end()) { return; }
const OperatorConf& op_conf = op_node->op().op_conf();
if (!op_conf.ctrl_in_op_name().empty()) { return; }
if (ctrl_in_op_names.find(op_conf.name()) != ctrl_in_op_names.end()) { return; }
if (!IsParallelCastOp(op_conf)) { return; }
if (op_node->in_edges().size() != 1) { return; }
user_op::UserOpConfWrapper conf_wrapper(op_conf);
const LogicalBlobId& parallel_cast_in_lbi = GenLogicalBlobId(conf_wrapper.input("in", 0));
const LogicalBlobId& parallel_cast_out_lbi = GenLogicalBlobId(conf_wrapper.output("out", 0));
const OpNode* producer = op_graph.OpNode4OpName(parallel_cast_in_lbi.op_name());

// Find the first op which won't be deleted
const OpNode* source_op = op_node;
const OpNode* producer = op_node->SoleInEdge()->src_node();
while (IsParallelCastOp(producer->op().op_conf())) {
Try2Delete(producer);
if (del_op_name_set.find(producer->op().op_name()) == del_op_name_set.end()) { break; }
source_op = producer;
producer = source_op->SoleInEdge()->src_node();
}
user_op::UserOpConfWrapper conf_wrapper_in(source_op->op().op_conf());
const LogicalBlobId& parallel_cast_in_lbi = GenLogicalBlobId(conf_wrapper_in.input("in", 0));

user_op::UserOpConfWrapper conf_wrapper_out(op_conf);
const LogicalBlobId& parallel_cast_out_lbi =
GenLogicalBlobId(conf_wrapper_out.output("out", 0));
if (op_node->parallel_desc() != producer->parallel_desc()) { return; }
const NdSbp& parallel_cast_nd_sbp = op_node->NdSbp4Lbi(parallel_cast_in_lbi);
for (const OpEdge* out_edge : op_node->out_edges()) {
const OpNode* consumer = out_edge->dst_node();
if (IsParallelCastOp(consumer->op().op_conf())) { return; }
if (consumer->parallel_desc() != op_node->parallel_desc()) { return; }
if (consumer->NdSbp4Lbi(parallel_cast_out_lbi) != parallel_cast_nd_sbp) { return; }
}
op_name2nd_sbp_signature[producer->op().op_name()] = producer->nd_sbp_signature();
for (const OpEdge* out_edge : op_node->out_edges()) {
Expand All @@ -124,8 +137,10 @@ Maybe<void> AutoParallelPass::RemoveParallelCastOps(Job* job) const {
}
}
del_op_names.emplace_back(op_conf.name());
del_op_name_set.insert(op_conf.name());
LOG(INFO) << "\tremove " << op_conf.name();
});
};
op_graph.ForEachNode(Try2Delete);
for (const auto& pair : op_name2op_conf) { job_builder.MutOpsOnlyOnce({pair.second}); }
for (const auto& pair : op_name2nd_sbp_signature) {
job_builder.AddNdSbpSignature4OpName(pair.first, pair.second);
Expand Down