#### Debug builds

When [building Tensorflow](https://www.tensorflow.org/install/source), passing
`--config=dbg` to Bazel will build with debugging information and without
optimizations, allowing you to use GDB or other debuggers to debug C++ code. For
example, you can build the pip package with debugging information by running:

```bash

Anyway, in any of the cases above, FastAPI will still work asynchronously and be extremely fast.

But by following the steps above, it will be able to do some performance optimizations.

## Technical Details { #technical-details }

Modern versions of Python have support for **"asynchronous code"** using something called **"coroutines"**, with **`async` and `await`** syntax.

   *
   * TODO(cpovirk): Write a test that asserts that AbstractFuture has no clinit?
   */

  /**
   * Tag interface marking trusted subclasses. This enables some optimizations. The implementation
   * of this interface must also be an AbstractFuture and must not override or expose for overriding
   * any of the public methods of ListenableFuture.
   */

from OkHttp 4.x:

**OkHttp is now packaged as separate JVM and Android artifacts.** This allows us to offer
platform-specific features and optimizations. If your build system handles [Gradle module metadata],
this change should be automatic.

**MockWebServer has a new coordinate and package name.** We didn’t like that our old artifact

// --------------------------------------------------------------------------
TF_SessionOptions* TF_NewSessionOptions() {
  TF_SessionOptions* out = new TF_SessionOptions;
  // Disable optimizations for static graph to allow calls to Session::Extend.
  out->options.config.mutable_experimental()
      ->set_disable_optimize_for_static_graph(true);
  return out;
}

common:linux --experimental_guard_against_concurrent_changes

# Configure short or long logs
common:short_logs --output_filter=DONT_MATCH_ANYTHING
common:verbose_logs --output_filter=

# Instruction set optimizations
# TODO(gunan): Create a feature in toolchains for avx/avx2 to
#   avoid having to define linux/win separately.
common:avx_linux --copt=-mavx
common:avx_linux --host_copt=-mavx
common:avx_linux --copt="-mf16c"

:= 1; i < len(table); i++ { table[i].montgomeryMul(table[i-1], table[0], m) } out.resetFor(m) out.limbs[0] = 1 out.montgomeryRepresenta(m) tmp := NewNat().ExpandFor(m) for _, b := range e { for _, j := range []int{4, 0} { // Square four times. Optimization note: this can be implemented // more efficiently than with generic Montgomery multiplication. out.montgomeryMul(out, out, m) out.montgomeryMul(out, out, m) out.montgomeryMul(out, out, m) out.montgomeryMul(out, out, m) // Select x^k in constant...

    }
    return (ListenableFuture<? extends T>[]) collection.toArray(new ListenableFuture<?>[0]);
  }

  // This can't be a TrustedFuture, because TrustedFuture has clever optimizations that
  // mean cancel won't be called if this Future is passed into setFuture, and then
  // cancelled.
  private static final class InCompletionOrderFuture<T extends @Nullable Object>
      extends AbstractFuture<T> {

     * system defaults will be used.
     *
     * Most applications should not call this method, and instead use the system defaults. Those
     * classes include special optimizations that can be lost if the implementations are decorated.
     *
     * If necessary, you can create and configure the defaults yourself with the following code:
     *
     * ```java

:= 1; i < len(table); i++ { table[i].montgomeryMul(table[i-1], table[0], m) } out.resetFor(m) out.limbs[0] = 1 out.montgomeryRepresenta(m) tmp := NewNat().ExpandFor(m) for _, b := range e { for _, j := range []int{4, 0} { // Square four times. Optimization note: this can be implemented // more efficiently than with generic Montgomery multiplication. out.montgomeryMul(out, out, m) out.montgomeryMul(out, out, m) out.montgomeryMul(out, out, m) out.montgomeryMul(out, out, m) // Select x^k in constant...