Docs: more details about concurrency, parallelism, threads

lib/utility.js

@@ -103,7 +103,11 @@ cache(true);
 
 /**
  * Gets or, when a concurrency is provided, sets
- * the number of threads _libvips'_ should create to process each image.
+ * the maximum number of threads _libvips_ should use to process _each image_.
+ * These are from a thread pool managed by glib,
+ * which helps avoid the overhead of creating new threads.
+ *
+ * This method always returns the current concurrency.
  *
  * The default value is the number of CPU cores,
  * except when using glibc-based Linux without jemalloc,
@@ -111,10 +115,19 @@ cache(true);
  *
  * A value of `0` will reset this to the number of CPU cores.
  *
- * The maximum number of images that can be processed in parallel
- * is limited by libuv's `UV_THREADPOOL_SIZE` environment variable.
+ * Some image format libraries spawn additional threads,
+ * e.g. libaom manages its own 4 threads when encoding AVIF images,
+ * and these are independent of the value set here.
  *
- * This method always returns the current concurrency.
+ * The maximum number of images that sharp can process in parallel
+ * is controlled by libuv's `UV_THREADPOOL_SIZE` environment variable,
+ * which defaults to 4.
+ *
+ * https://nodejs.org/api/cli.html#uv_threadpool_sizesize
+ *
+ * For example, by default, a machine with 8 CPU cores will process
+ * 4 images in parallel and use up to 8 threads per image,
+ * so there will be up to 32 concurrent threads.
  *
  * @example
  * const threads = sharp.concurrency(); // 4