De-duplicate GPU parameters. (#4454)

* Only define `gpu_id` and `n_gpus` in `LearnerTrainParam`
* Pass LearnerTrainParam through XGBoost vid factory method.
* Disable all GPU usage when GPU related parameters are not specified (fixes XGBoost choosing GPU over aggressively).
* Test learner train param io.
* Fix gpu pickling.
This commit is contained in:
Jiaming Yuan
2019-05-29 11:55:57 +08:00
committed by GitHub
parent a3fedbeaa8
commit c589eff941
69 changed files with 927 additions and 562 deletions

View File

@@ -1,21 +1,14 @@
/*!
* Copyright 2018 XGBoost contributors
* Copyright 2018-2019 XGBoost contributors
*/
#include <xgboost/metric.h>
#include <map>
#include "../helpers.h"
using Arg = std::pair<std::string, std::string>;
#if defined(__CUDACC__)
#define N_GPU() Arg{"n_gpus", "1"}
#else
#define N_GPU() Arg{"n_gpus", "0"}
#endif
TEST(Metric, DeclareUnifiedTest(RMSE)) {
xgboost::Metric * metric = xgboost::Metric::Create("rmse");
metric->Configure({N_GPU()});
auto lparam = xgboost::CreateEmptyGenericParam(0, NGPUS);
xgboost::Metric * metric = xgboost::Metric::Create("rmse", &lparam);
metric->Configure({});
ASSERT_STREQ(metric->Name(), "rmse");
EXPECT_NEAR(GetMetricEval(metric, {0, 1}, {0, 1}), 0, 1e-10);
EXPECT_NEAR(GetMetricEval(metric,
@@ -27,8 +20,9 @@ TEST(Metric, DeclareUnifiedTest(RMSE)) {
}
TEST(Metric, DeclareUnifiedTest(MAE)) {
xgboost::Metric * metric = xgboost::Metric::Create("mae");
metric->Configure({N_GPU()});
auto lparam = xgboost::CreateEmptyGenericParam(0, NGPUS);
xgboost::Metric * metric = xgboost::Metric::Create("mae", &lparam);
metric->Configure({});
ASSERT_STREQ(metric->Name(), "mae");
EXPECT_NEAR(GetMetricEval(metric, {0, 1}, {0, 1}), 0, 1e-10);
EXPECT_NEAR(GetMetricEval(metric,
@@ -39,8 +33,9 @@ TEST(Metric, DeclareUnifiedTest(MAE)) {
}
TEST(Metric, DeclareUnifiedTest(LogLoss)) {
xgboost::Metric * metric = xgboost::Metric::Create("logloss");
metric->Configure({N_GPU()});
auto lparam = xgboost::CreateEmptyGenericParam(0, NGPUS);
xgboost::Metric * metric = xgboost::Metric::Create("logloss", &lparam);
metric->Configure({});
ASSERT_STREQ(metric->Name(), "logloss");
EXPECT_NEAR(GetMetricEval(metric, {0, 1}, {0, 1}), 0, 1e-10);
EXPECT_NEAR(GetMetricEval(metric,
@@ -51,8 +46,9 @@ TEST(Metric, DeclareUnifiedTest(LogLoss)) {
}
TEST(Metric, DeclareUnifiedTest(Error)) {
xgboost::Metric * metric = xgboost::Metric::Create("error");
metric->Configure({N_GPU()});
auto lparam = xgboost::CreateEmptyGenericParam(0, NGPUS);
xgboost::Metric * metric = xgboost::Metric::Create("error", &lparam);
metric->Configure({});
ASSERT_STREQ(metric->Name(), "error");
EXPECT_NEAR(GetMetricEval(metric, {0, 1}, {0, 1}), 0, 1e-10);
EXPECT_NEAR(GetMetricEval(metric,
@@ -60,17 +56,17 @@ TEST(Metric, DeclareUnifiedTest(Error)) {
{ 0, 0, 1, 1}),
0.5f, 0.001f);
EXPECT_ANY_THROW(xgboost::Metric::Create("error@abc"));
EXPECT_ANY_THROW(xgboost::Metric::Create("error@abc", &lparam));
delete metric;
metric = xgboost::Metric::Create("error@0.5f");
metric->Configure({N_GPU()});
metric = xgboost::Metric::Create("error@0.5f", &lparam);
metric->Configure({});
EXPECT_STREQ(metric->Name(), "error");
delete metric;
metric = xgboost::Metric::Create("error@0.1");
metric->Configure({N_GPU()});
metric = xgboost::Metric::Create("error@0.1", &lparam);
metric->Configure({});
ASSERT_STREQ(metric->Name(), "error@0.1");
EXPECT_STREQ(metric->Name(), "error@0.1");
EXPECT_NEAR(GetMetricEval(metric, {0, 1}, {0, 1}), 0, 1e-10);
@@ -82,8 +78,9 @@ TEST(Metric, DeclareUnifiedTest(Error)) {
}
TEST(Metric, DeclareUnifiedTest(PoissionNegLogLik)) {
xgboost::Metric * metric = xgboost::Metric::Create("poisson-nloglik");
metric->Configure({N_GPU()});
auto lparam = xgboost::CreateEmptyGenericParam(0, NGPUS);
xgboost::Metric * metric = xgboost::Metric::Create("poisson-nloglik", &lparam);
metric->Configure({});
ASSERT_STREQ(metric->Name(), "poisson-nloglik");
EXPECT_NEAR(GetMetricEval(metric, {0, 1}, {0, 1}), 0.5f, 1e-10);
EXPECT_NEAR(GetMetricEval(metric,
@@ -96,8 +93,9 @@ TEST(Metric, DeclareUnifiedTest(PoissionNegLogLik)) {
#if defined(XGBOOST_USE_NCCL) && defined(__CUDACC__)
TEST(Metric, MGPU_RMSE) {
{
xgboost::Metric * metric = xgboost::Metric::Create("rmse");
metric->Configure({Arg{"n_gpus", "-1"}});
auto lparam = xgboost::CreateEmptyGenericParam(0, -1);
xgboost::Metric * metric = xgboost::Metric::Create("rmse", &lparam);
metric->Configure({});
ASSERT_STREQ(metric->Name(), "rmse");
EXPECT_NEAR(GetMetricEval(metric, {0}, {0}), 0, 1e-10);
EXPECT_NEAR(GetMetricEval(metric,
@@ -108,8 +106,8 @@ TEST(Metric, MGPU_RMSE) {
}
{
xgboost::Metric * metric = xgboost::Metric::Create("rmse");
metric->Configure({Arg{"n_gpus", "-1"}, Arg{"gpu_id", "1"}});
auto lparam = xgboost::CreateEmptyGenericParam(1, -1);
xgboost::Metric * metric = xgboost::Metric::Create("rmse", &lparam);
ASSERT_STREQ(metric->Name(), "rmse");
EXPECT_NEAR(GetMetricEval(metric, {0, 1}, {0, 1}), 0, 1e-10);
EXPECT_NEAR(GetMetricEval(metric,