// Compute the filter.
    long filter = 0;
    int size = chars.cardinality();
    boolean containsZero = chars.get(0);
    // Compute the hash table.
    char[] table = new char[chooseTableSize(size)];
    int mask = table.length - 1;
    for (int c = chars.nextSetBit(0); c != -1; c = chars.nextSetBit(c + 1)) {
      // Compute the filter at the same time.
      filter |= 1L << c;

        return (logX2Floor < 2 * logFloor + 1) ? logFloor : logFloor + 1;
    }
    throw new AssertionError();
  }

  /*
   * The maximum number of bits in a square root for which we'll precompute an explicit half power
   * of two. This can be any value, but higher values incur more class load time and linearly
   * increasing memory consumption.
   */

    // Compute the filter.
    long filter = 0;
    int size = chars.cardinality();
    boolean containsZero = chars.get(0);
    // Compute the hash table.
    char[] table = new char[chooseTableSize(size)];
    int mask = table.length - 1;
    for (int c = chars.nextSetBit(0); c != -1; c = chars.nextSetBit(c + 1)) {
      // Compute the filter at the same time.
      filter |= 1L << c;

            print(
                'ERROR: TensorFlow only supports small CUDA compute'
                ' capabilities of sm_30 and higher. Please re-specify the list'
                ' of compute capabilities excluding version %s.' % ver)
            all_valid = False
          if ver < 35:
            print('WARNING: XLA does not support CUDA compute capabilities '

internal fun <T : Any> AtomicReference<Tags>.computeIfAbsent(
  type: KClass<T>,
  compute: () -> T,
): T {
  var computed: T? = null

  while (true) {
    val tags = get()

    // If the element is already present. Return it.
    val existing = tags[type]
    if (existing != null) return existing

    if (computed == null) {
      computed = compute()
    }

namespace gradients {

// =============== Experimental C++ API for computing gradients ===============

// Sample gradient function:
//
// class AddGradientFunction : public GradientFunction {
//  public:
//   Status Compute(Context* ctx,
//                  absl::Span<AbstractTensorHandle* const> grad_inputs,
//                  absl::Span<AbstractTensorHandle*> grad_outputs) override {
//     grad_outputs[0] = grad_inputs[0];

 *          since this whole object is kept in the classloader between builds.
 *          Anything that changes must be in a val with a get() method that recomputes the value each time.
 */
object BuildEnvironment {

    /**
     * A selection of environment variables injected into the environment by the `codeql-env.sh` script.
     */
    private

        // use original key
        return parent.computeIfPresent(key, remappingFunction);
    }

    @Override
    public V compute(final K key, final BiFunction<? super K, ? super V, ? extends V> remappingFunction) {
        // use original key
        return parent.compute(key, remappingFunction);
    }

    @Override

     */
    @Nullable
    <T> T get(@Nonnull Key<T> key);

    /**
     * Retrieve of compute the data associated with the specified key.
     *
     * @param key the key for which to retrieve the session data, must not be {@code null}
     * @param supplier the supplier will compute the new value
     * @return the session data associated with the key
     */
    @Nullable

// Copyright 2009 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.

//go:build ignore

// Compute Fibonacci numbers with two goroutines
// that pass integers back and forth.  No actual
// concurrency, just threads and synchronization
// and foreign code on multiple pthreads.

package main

import (
	big "."
	"runtime"
)