diff --git a/presto/docker/config/template/etc_coordinator/config_native.properties b/presto/docker/config/template/etc_coordinator/config_native.properties
index b08eca85..ccb7cac8 100644
--- a/presto/docker/config/template/etc_coordinator/config_native.properties
+++ b/presto/docker/config/template/etc_coordinator/config_native.properties
@@ -25,6 +25,10 @@ node-scheduler.max-pending-splits-per-task=2000
 node-scheduler.max-splits-per-node=2000
 
 # Optimizer flags
+# New option in 0.282 to control inference of NOT NULL columns in joins (undocumented).
+#optimizer.joins-not-null-inference-strategy=USE_FUNCTION_METADATA
+# Use a default filter factor to estimate the selectivity of filters in queries.
+#optimizer.default-filter-factor-enabled=true
 # Use known constraints to simplify plan and filters.
 optimizer.exploit-constraints=true
 # Rewrite large IN lists as joins for performance in some cases.
diff --git a/presto/scripts/generate_presto_config.sh b/presto/scripts/generate_presto_config.sh
index 35249a0a..48016eb2 100755
--- a/presto/scripts/generate_presto_config.sh
+++ b/presto/scripts/generate_presto_config.sh
@@ -83,4 +83,13 @@ if ../../pbench/pbench genconfig -p params.json -t template generated 2>&1 | gre
     echo_error "ERROR in pbench genconfig.  Configs were not generated successfully"
 fi
 
+# now perform other variant-specific modifications to the generated configs
+if [[ "${VARIANT_TYPE}" == "gpu" ]]; then
+  # for GPU variant, uncomment these optimizer settings
+  # optimizer.joins-not-null-inference-strategy=USE_FUNCTION_METADATA
+  # optimizer.default-filter-factor-enabled=true
+  COORD_CONFIG="generated/etc_coordinator/config_native.properties"
+  sed -i 's/\#optimizer/optimizer/g' ${COORD_CONFIG}
+fi
+
 echo_success "Configs were generated successfully"