* connection is typically faster than establishing a new one.
 *
 * When a single logical call requires multiple streams due to redirects or authorization
 * challenges, we prefer to use the same physical connection for all streams in the sequence. There
 * are potential performance and behavior consequences to this preference. To support this feature,

type StaticPolicyOptions struct {
	// flag to enable extra allocation restrictions to avoid
	// different containers to possibly end up on the same core.
	// we consider "core" and "physical CPU" synonim here, leaning
	// towards the terminoloy k8s hints. We acknowledge this is confusing.
	//
	// looking at https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/,

		//    reshaping it to be narrower and longer. As long as the maximum
		//    seats setting will prevent dispatching too many requests at once
		//    to prevent overloading kube-apiserver (and/or etcd or the VM or
		//    a physical machine it is running on), we believe the relaxed
		//    version should be good enough to achieve the P&F goals.
		//
		// TODO: Confirm that the current cap of maximumSeats allow us to
		//   achieve the above.

    }

    private void appendOsStats(final StringBuilder buf) {
        buf.append("\"os\":{");
        final OsProbe osProbe = OsProbe.getInstance();
        buf.append("\"memory\":{");
        buf.append("\"physical\":{");
        append(buf, "free", () -> osProbe.getFreePhysicalMemorySize()).append(',');
        append(buf, "total", () -> osProbe.getTotalPhysicalMemorySize());
        buf.append("},");

        <td>Bytes in idle (unused) spans.</td>
    </tr>

    <tr>
        <td>HeapReleased</td>
        <td id="HeapReleased">{{.HeapReleased}} bytes</td>
        <td>Physical memory returned to the OS.</td>
    </tr>

    <tr>
        <td>HeapObjects</td>
        <td id="HeapObjects">{{.HeapObjects}} objects</td>
        <td>Number of allocated heap objects.</td>
    </tr>

)

type IfMsghdr struct {
	Msglen  uint16
	Version uint8
	Type    uint8
	Index   uint16
	Flags   int32
	Addrs   int32
	Data    IfData
}

type IfData struct {
	Type       uint8
	Physical   uint8
	Addrlen    uint8
	Hdrlen     uint8
	Recvquota  uint8
	Xmitquota  uint8
	Mtu        uint64
	Metric     uint64
	Link_state uint64
	Baudrate   uint64
	Ipackets   uint64
	Ierrors    uint64

[[share_results_between_ci_builds]]
== Share results between CI builds

The build cache can do more than go back-and-forth in time: it can also bridge physical distance between computers, allowing results generated on one machine to be re-used by another.

## Server Machine and Server Program

There's a small detail about names to keep in mind. 💡

The word "**server**" is commonly used to refer to both the remote/cloud computer (the physical or virtual machine) and also the program that is running on that machine (e.g. Uvicorn).

Just keep in mind that when you read "server" in general, it could refer to one of those two things.

ways to detect malicious models/graphs/checkpoints, so the recommended way to
mitigate the risk in this scenario is to sandbox the model execution.

### Hardware attacks

Physical GPUs or TPUs can also be the target of attacks. [Published
research](https://scholar.google.com/scholar?q=gpu+side+channel) shows that it
might be possible to use side channel attacks on the GPU to leak data from other

	// to the Windows docs) in addition to the FlushViewOfFile call
	// above, " ... to flush all the dirty pages plus the metadata for
	// the file and ensure that they are physically written to disk".
	// Windows DOC links:
	//
	// https://docs.microsoft.com/en-us/windows/win32/api/memoryapi/nf-memoryapi-flushviewoffile