listOf(
                    "clean",
                    "performance:${testProject}PerformanceAdHocTest",
                    tests.map { """--tests "$it"""" }.joinToString(" "),
                    """--warmups 2 --runs 2 --checks none""",
                    "-PtestJavaVersion=${os.perfTestJavaVersion.major}",
                    "-PtestJavaVendor=${os.perfTestJavaVendor.name.lowercase()}",

    for (int i = 0; i < 8; i++) {
      limiter.acquire(); // // #1
    }
    stopwatch.sleepMillis(4500); // #2: back to cold state (warmup period + repay last acquire)
    for (int i = 0; i < 3; i++) { // only three steps, we're somewhere in the warmup period
      limiter.acquire(); // #3
    }

    limiter.setRate(4.0); // double the rate!

    for (int i = 0; i < 8; i++) {
      limiter.acquire(); // // #1
    }
    stopwatch.sleepMillis(4500); // #2: back to cold state (warmup period + repay last acquire)
    for (int i = 0; i < 3; i++) { // only three steps, we're somewhere in the warmup period
      limiter.acquire(); // #3
    }

    limiter.setRate(4.0); // double the rate!

    stage: Stage,
) : OsAwareBaseGradleBuildType(os = Os.LINUX, stage = stage, init = {
        id(buildTypeId(model))
        name = "Compile All"
        description = "Compiles all production/test source code and warms up the build cache"

        features {
            publishBuildStatusToGithub(model)
        }

        applyDefaults(
            model,
            this,

*   Added warmup capabilities to `tf.keras.optimizers.schedules.CosineDecay` learning rate scheduler. You can now specify an initial and target learning rate, and our scheduler will perform a linear interpolation between the two after which it will begin a decay phase....