diff --git a/CMakeLists.txt b/CMakeLists.txt
index 7a74288..6f45d04 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -15,9 +15,10 @@ if(TATAMI_STATS_FETCH_EXTERN)
     add_subdirectory(extern)
 else()
     find_package(tatami_tatami 3.0.0 CONFIG REQUIRED)
+    find_package(ltla_subpar 0.3.0 CONFIG REQUIRED)
 endif()
 
-target_link_libraries(tatami_stats INTERFACE tatami::tatami)
+target_link_libraries(tatami_stats INTERFACE tatami::tatami ltla::subpar)
 
 include(GNUInstallDirs)
 target_include_directories(tatami_stats INTERFACE 
diff --git a/cmake/Config.cmake.in b/cmake/Config.cmake.in
index 1fd297e..cb5c884 100644
--- a/cmake/Config.cmake.in
+++ b/cmake/Config.cmake.in
@@ -2,5 +2,6 @@
 
 include(CMakeFindDependencyMacro)
 find_dependency(tatami_tatami 3.0.0 CONFIG REQUIRED)
+find_dependency(ltla_subpar 0.3.0 CONFIG REQUIRED)
 
 include("${CMAKE_CURRENT_LIST_DIR}/tatami_tatami_statsTargets.cmake")
diff --git a/extern/CMakeLists.txt b/extern/CMakeLists.txt
index a937b67..c172b5b 100644
--- a/extern/CMakeLists.txt
+++ b/extern/CMakeLists.txt
@@ -3,8 +3,14 @@ include(FetchContent)
 FetchContent_Declare(
   tatami
   GIT_REPOSITORY https://github.com/tatami-inc/tatami
-  GIT_TAG master # ^v3.0.0
+  GIT_TAG master # ^3.0.0
 )
 
-FetchContent_MakeAvailable(tatami)
+FetchContent_Declare(
+  subpar
+  GIT_REPOSITORY https://github.com/LTLA/subpar
+  GIT_TAG master # ^0.3.0
+)
 
+FetchContent_MakeAvailable(tatami)
+FetchContent_MakeAvailable(subpar)
diff --git a/include/tatami_stats/counts.hpp b/include/tatami_stats/counts.hpp
index 4a72751..34e687f 100644
--- a/include/tatami_stats/counts.hpp
+++ b/include/tatami_stats/counts.hpp
@@ -86,6 +86,7 @@ void apply(bool row, const tatami::Matrix<Value_, Index_>* p, Output_* output, i
         }
 
     } else {
+        num_threads = subpar::sanitize_num_workers(num_threads, otherdim); // provides some protection against silly num_threads iputs.
         std::vector<Output_*> threaded_output_ptrs(num_threads, output);
         std::vector<std::vector<Output_> > threaded_output(num_threads - 1);
         for (int t = 1; t < num_threads; ++t) {