# Set lld as the linker.
common:cuda_clang --host_linkopt="-fuse-ld=lld"
common:cuda_clang --host_linkopt="-lm"
common:cuda_clang --linkopt="-fuse-ld=lld"
common:cuda_clang --linkopt="-lm"

# Set up compilation CUDA version and paths and use the CUDA Clang toolchain.
common:cuda_clang_official --config=cuda_clang
common:cuda_clang_official --config=cuda_version

# Build with nvcc for CUDA and clang for host

So, the function below it will be executed once for each combination of arguments. And then the values returned by each of those combinations of arguments will be used again and again whenever the function is called with exactly the same combination of arguments.

For example, if you have a function:

```Python
@lru_cache

    public static final int SMB2_LEASE_HANDLE_CACHING = 0x02;  // H - Handle caching  
    public static final int SMB2_LEASE_WRITE_CACHING  = 0x04;  // W - Write caching
    
    // Common combinations
    public static final int SMB2_LEASE_READ_HANDLE    = 0x03;  // RH
    public static final int SMB2_LEASE_READ_WRITE     = 0x05;  // RW
    public static final int SMB2_LEASE_FULL           = 0x07;  // RWH
}
```

            assertFalse((noAccess & ACE.FILE_READ_DATA) != 0, "No access should not include FILE_READ_DATA");
        }

        @Test
        @DisplayName("Should validate inheritance flag combinations")
        void shouldValidateInheritanceFlagCombinations() {
            // Object and container inheritance
            int bothInherit = ACE.FLAGS_OBJECT_INHERIT | ACE.FLAGS_CONTAINER_INHERIT;

      }
      statsCounter.recordEviction();
      return removal;
    }
  }

  /**
   * Any updates to LocalCache.Strength used in CacheBuilder need to be matched in this class for
   * compilation purposes.
   */
  enum Strength {
    /*
     * TODO(kevinb): If we strongly reference the value and aren't loading, we needn't wrap the
     * value. This could save ~8 bytes per entry.
     */

		settings = ParseTagSetting(str, ",")
	)

	// optimize match english letters and midline
	// The following code is basically called in for.
	// In order to avoid the performance problems caused by repeated compilation of regular expressions,
	// it only needs to be done once outside, so optimization is done here.
	if idx != -1 && regEnLetterAndMidline.MatchString(str[0:idx]) {
		name = str[0:idx]
	} else {

        assertArrayEquals(testMac, mockQueryAddress.macAddress);
    }

    @Test
    void readRDataWireFormat_shouldParseNodeFlags() throws Exception {
        // Prepare test data with various flag combinations
        byte[] src = new byte[100];
        int srcIndex = 0;

        Field rDataLengthField = NameServicePacket.class.getDeclaredField("rDataLength");
        rDataLengthField.setAccessible(true);

            assertEquals(expectedSize, response.getSize());
        }

        @Test
        @DisplayName("Should handle response with different fileId and fileName combinations")
        void testDifferentFileIdAndFileNameCombinations() throws SMBProtocolDecodingException {
            // Test with empty fileName

            "2, 7", // SMB2_0_INFO_FILESYSTEM, FS_FULL_SIZE_INFO
            "3, 0" // SMB2_0_INFO_SECURITY, any class
    })
    @DisplayName("Test createInformation with various valid info type and class combinations")
    void testCreateInformationValidCombinations(byte infoType, byte infoClass) throws Exception {
        // Use reflection to test private method

And you could have a single dependency that requires several other dependencies with `yield`, etc.

You can have any combinations of dependencies that you want.

**FastAPI** will make sure everything is run in the correct order.

/// note | Technical Details