- name: Enable KVM group perms
        # https://github.blog/changelog/2023-02-23-hardware-accelerated-android-virtualization-on-actions-windows-and-linux-larger-hosted-runners/
        run: |
          echo 'KERNEL=="kvm", GROUP="kvm", MODE="0666", OPTIONS+="static_node=kvm"' | sudo tee /etc/udev/rules.d/99-kvm4all.rules
          sudo udevadm control --reload-rules
          sudo udevadm trigger --name-match=kvm

# Only include debug info for files under tensorflow/, excluding kernels, to
# reduce the size of the debug info in the binary. This is because if the debug
# sections in the ELF binary are too large, errors can occur. See
# https://github.com/tensorflow/tensorflow/issues/48919.
# Users can still include debug info for a specific kernel, e.g. with:
#     --config=dbg --per_file_copt=+tensorflow/core/kernels/identity_op.*@-g

Go 1.22 changed how the runtime interacts with transparent huge pages on Linux.
In particular, a common default Linux kernel configuration can result in
significant memory overheads, and Go 1.22 no longer works around this default.
To work around this issue without adjusting kernel settings, transparent huge
pages can be disabled for Go memory with the
[`disablethp` setting](/pkg/runtime#hdr-Environment_Variables).

Maschinen im Cluster verfügbare Menge berücksichtigt werden.

Wenn Ihre Anwendung **einfach** ist, wird dies wahrscheinlich **kein Problem darstellen** und Sie müssen möglicherweise keine festen Speichergrenzen angeben. Wenn Sie jedoch **viel Speicher verbrauchen** (z. B. bei **Modellen für maschinelles Lernen**), sollten Sie überprüfen, wie viel Speicher Sie verbrauchen, und die **Anzahl der Container** anpassen, die in **jeder Maschine** ausgeführt werden (und möglicherweise weitere Maschinen...

# kernel registration - REGISTER_OP, REGISTER_KERNEL_BUILDER, etc.
#
# This setting is currently experimental. The 'v2' implementation does _not_
# correspond to a particular, finalized design; rather, it relates to
# developing one.
#
# The current aim of the 'v2' implementation is to allow 'unused' ops and
# kernels to be discarded by the linker (to the benefit of binary size).

Care should be taken when using vector instructions to ensure that they are available at
runtime.
To use vector instructions a machine must have both the vector facility (bit 129 in the
facility list) and kernel support.
Without kernel support a vector instruction will have no effect (it will be equivalent
to a <code>NOP</code> instruction).
</p>

<p>
Addressing modes:
</p>

<ul>

<li>

- Removed a warning related to Linux user namespaces and kernel version. Previously, if the feature gate `UserNamespacesSupport` was enabled, the kubelet warned when detecting a Linux kernel version earlier than 6.3.0. While user namespace support generally requires kernel 6.3 or newer, it can also work on older kernels. ([#130243](https://github.com/kubernetes/kubernetes/pull/130243), [@rata](https://github.com/rata))

- Removed a warning around Linux user namespaces and kernel version. If the feature gate `UserNamespacesSupport` was enabled, the kubelet previously warned when detecting a Linux kernel version earlier than 6.3.0. User namespace support on Linux typically does still need kernel 6.3 or newer, but it can work in older kernels too. ([#131784](https://github.com/kubernetes/kubernetes/pull/131784), [@rata](https://github.com/rata)) [SIG Node]...

  TF_DeleteBuffer(op_list_buf);
}

class DummyKernel : public tensorflow::OpKernel {
 public:
  explicit DummyKernel(tensorflow::OpKernelConstruction* context)
      : OpKernel(context) {}
  void Compute(tensorflow::OpKernelContext* context) override {}
};

// Test we can query kernels
REGISTER_OP("TestOpWithSingleKernel")
    .Input("a: float")
    .Input("b: float")

Contêineres Linux rodam usando o mesmo kernel Linux do hospedeiro (máquina, máquina virtual, servidor na nuvem, etc). Isso simplesmente significa que eles são muito leves (comparados com máquinas virtuais emulando um sistema operacional completo).