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

[xpu] fused_MT xpu kernel support #51571

Merged
merged 8 commits into from
Mar 20, 2023
Merged

[xpu] fused_MT xpu kernel support #51571

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
38eecea
[xpu] fused_MT xpu kernel support
mayang002 Mar 13, 2023
bb267b9
[xpu] fused_MT xpu kernel support: typo fix
mayang002 Mar 15, 2023
292fe4a
Fix typo bug of fused_multi_transformer_xpu_kernel.cc and Merge branc…
mayang002 Mar 15, 2023
370757c
[xpu] typo fix in fused_multi_transformer_quant_pass.cc and fused_mul…
mayang002 Mar 15, 2023
0ca6edf
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
mayang002 Mar 16, 2023
3896c56
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
mayang002 Mar 16, 2023
28c5b29
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
mayang002 Mar 16, 2023
dde77c2
[xpu] add c++ test of fused_multi_transformer_xpu_quant_pass
mayang002 Mar 17, 2023
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
2 changes: 2 additions & 0 deletions cmake/external/xpu.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -142,6 +142,8 @@ if(WITH_XPU_XFT)
message(STATUS "Compile with XPU XFT!")
add_definitions(-DPADDLE_WITH_XPU_XFT)

set(XPU_XFT_INC_DIR "${XPU_INC_DIR}/xft")
include_directories(${XPU_XFT_INC_DIR})
set(XPU_XFT_LIB "${XPU_LIB_DIR}/${XPU_XFT_LIB_NAME}")
endif()

Expand Down
6 changes: 6 additions & 0 deletions paddle/fluid/framework/ir/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -235,6 +235,8 @@ if(WITH_XPU)
pass_library(link_xpu_op_max_pass inference DIR xpu DEPS ${XPU_PASS_DEPS})
pass_library(delete_isolated_node_pass inference DIR xpu DEPS
${XPU_PASS_DEPS})
pass_library(fused_multi_transformer_xpu_quant_pass inference DIR xpu DEPS
${XPU_PASS_DEPS})
endif()

cc_library(
Expand Down Expand Up @@ -493,4 +495,8 @@ if(WITH_XPU)
test_delete_isolated_node_pass
SRCS xpu/delete_isolated_node_pass_test.cc
DEPS delete_isolated_node_pass)
cc_test(
test_fused_multi_transformer_xpu_quant_pass
SRCS xpu/fused_multi_transformer_xpu_quant_pass_tester.cc
DEPS fused_multi_transformer_xpu_quant_pass)
endif()
8 changes: 4 additions & 4 deletions paddle/fluid/framework/ir/fuse_multi_transformer_layer_pass_tester.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,7 +75,7 @@ TEST(FuseMultiTransformerLayerPass, encoder_fp) {
1,
{2, -1, 16, 1024, 64},
0);
auto* out = layers.fused_multi_transformer(x,
auto outs = layers.fused_multi_transformer(x,
cache_kv,
src_mask,
qkv_w,
Expand All @@ -93,7 +93,7 @@ TEST(FuseMultiTransformerLayerPass, encoder_fp) {
0.1,
1e-12);

x = out;
x = outs[0];
}
std::unique_ptr<ir::Graph> graph(new ir::Graph(layers.main_program()));
graph->Set("__param_scope__", CreateParamScope());
Expand Down Expand Up @@ -126,7 +126,7 @@ TEST(FuseMultiTransformerLayerPass, decoder_fp) {
for (int i = 0; i < num_layers; ++i) {
auto* shape_out = layers.shape(src_mask);
auto* time_stamp = layers.slice(shape_out, {0}, {3}, {4});
auto* out = layers.fused_multi_transformer(x,
auto outs = layers.fused_multi_transformer(x,
cache_kv,
src_mask,
qkv_w,
Expand All @@ -145,7 +145,7 @@ TEST(FuseMultiTransformerLayerPass, decoder_fp) {
1e-12,
time_stamp);

x = out;
x = outs[0];
}
std::unique_ptr<ir::Graph> graph(new ir::Graph(layers.main_program()));
auto param_scope = CreateParamScope();
Expand Down
9 changes: 9 additions & 0 deletions paddle/fluid/framework/ir/node.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -151,6 +151,15 @@ class Node {
var_desc_->SetName(new_name);
}

void RenameOp(const std::string& new_name) {
PADDLE_ENFORCE_EQ(
type_ == Type::kOperation && op_desc_,
true,
platform::errors::InvalidArgument("Node must be type of variable."));
name_ = new_name;
op_desc_->SetType(new_name);
}

int DescOrder() const { return desc_order_; }

int GetVarNodeBlockId() const {
Expand Down
1 change: 1 addition & 0 deletions paddle/fluid/framework/ir/pass.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,7 @@ static const std::vector<std::string> support_subgraph_passes = {
"fuse_multi_transformer_layer_pass",
"delete_quant_dequant_linear_op_pass",
"delete_weight_dequant_linear_op_pass",
"fused_multi_transformer_xpu_quant_pass",
"fc_xpu_fuse_pass",
"delete_op_device_pass"};

Expand Down
58 changes: 31 additions & 27 deletions paddle/fluid/framework/ir/pass_tester_helper.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -571,33 +571,35 @@ struct Layers {
return out;
}

VarDesc* fused_multi_transformer(VarDesc* x,
VarDesc* cache_kv,
VarDesc* src_mask,
VarDesc* qkv_w,
VarDesc* qkv_bias,
VarDesc* out_linear_w,
VarDesc* out_linear_bias,
VarDesc* ffn1_w,
VarDesc* ffn1_bias,
VarDesc* ffn2_w,
VarDesc* ffn2_bias,
VarDesc* ln_scale,
VarDesc* ln_bias,
VarDesc* ffn_ln_scale,
VarDesc* ffn_ln_bias,
float epsilon,
float dropout_rate,
VarDesc* time_stamp = nullptr,
VarDesc* qkv_out_scale = nullptr,
VarDesc* out_linear_out_scale = nullptr,
VarDesc* ffn1_out_scale = nullptr,
VarDesc* ffn2_out_scale = nullptr,
std::vector<float> qkv_in_scale = {},
std::vector<float> out_linear_in_scale = {},
std::vector<float> ffn1_in_scale = {},
std::vector<float> ffn2_in_scale = {}) {
std::vector<VarDesc*> fused_multi_transformer(
VarDesc* x,
VarDesc* cache_kv,
VarDesc* src_mask,
VarDesc* qkv_w,
VarDesc* qkv_bias,
VarDesc* out_linear_w,
VarDesc* out_linear_bias,
VarDesc* ffn1_w,
VarDesc* ffn1_bias,
VarDesc* ffn2_w,
VarDesc* ffn2_bias,
VarDesc* ln_scale,
VarDesc* ln_bias,
VarDesc* ffn_ln_scale,
VarDesc* ffn_ln_bias,
float epsilon,
float dropout_rate,
VarDesc* time_stamp = nullptr,
VarDesc* qkv_out_scale = nullptr,
VarDesc* out_linear_out_scale = nullptr,
VarDesc* ffn1_out_scale = nullptr,
VarDesc* ffn2_out_scale = nullptr,
std::vector<float> qkv_in_scale = {},
std::vector<float> out_linear_in_scale = {},
std::vector<float> ffn1_in_scale = {},
std::vector<float> ffn2_in_scale = {}) {
VarDesc* out = lod_tensor(unique_name());
VarDesc* cache_kv_out = lod_tensor(unique_name());
OpDesc* op = program_.MutableBlock(0)->AppendOp();
std::string op_type = qkv_out_scale ? "fused_multi_transformer_int8"
: "fused_multi_transformer";
Expand All @@ -623,6 +625,7 @@ struct Layers {
op->SetAttr("dropout_rate", dropout_rate);
op->SetAttr("epsilon", epsilon);
op->SetOutput("Out", {out->Name()});
op->SetOutput("CacheKVOut", {cache_kv_out->Name()});

if (time_stamp) {
op->SetInput("TimeStep", {time_stamp->Name()});
Expand All @@ -638,7 +641,8 @@ struct Layers {
op->SetAttr("ffn1_in_scale", ffn1_in_scale);
op->SetAttr("ffn2_in_scale", ffn2_in_scale);
}
return out;
std::vector<VarDesc*> outs = {out, cache_kv_out};
return outs;
}

VarDesc* dequantize_linear(VarDesc* x,
Expand Down
Loading