# `compute_XY` enables PTX embedding in addition to SASS. PTX
# is forward compatible beyond the current compute capability major
# release while SASS is only forward compatible inside the current
# major release. Example: sm_80 kernels can run on sm_89 GPUs but
# not on sm_90 GPUs. compute_80 kernels though can also run on sm_90 GPUs.
build:cuda_clang --repo_env=HERMETIC_CUDA_COMPUTE_CAPABILITIES="sm_60,sm_70,sm_80,sm_89,compute_90"

	// what we have sent to the client during PutObjectPart. The reason is
	// that ETags are encrypted. Hence, the client will send a list of complete
	// part ETags of which may not match the ETag of any part. For example
	//   ETag (client):          30902184f4e62dd8f98f0aaff810c626
	//   ETag (server-internal): 20000f00ce5dc16e3f3b124f586ae1d88e9caa1c598415c2759bbb50e84a59f630902184f4e62dd8f98f0aaff810c626
	//

                         * actually shutdown the connection before other threads
                         * have committed themselves (e.g. InterruptTest example).
                         */
                        try {
                            trans.disconnect(true);
                        } catch (Exception e) {

// For ARM64, V0 through V31 with arrangement may appear.
//
// For 386/AMD64 register list specifies 4VNNIW-style multi-source operand.
// For range of 4 elements, Intel manual uses "+3" notation, for example:
//
//	VP4DPWSSDS zmm1{k1}{z}, zmm2+3, m128
//
// Given asm line:
//
//	VP4DPWSSDS Z5, [Z10-Z13], (AX)
//
// zmm2 is Z10, and Z13 is the only valid value for it (Z10+3).

2. Review and delete any PodSecurityPolicy objects you do not want all pod-creating users to be able to use (NOTE: Privileged users that were making use of those policies will also lose access to those policies). For example:
  * `kubectl delete podsecuritypolicies/my-privileged-policy`
3. After upgrading to 1.5.5, re-create the exported PodSecurityPolicy objects:
  * `kubectl create -f psp.yaml`

## Downloads for v1.5.5

		// as a List() call is an important performance improvement.
		//
		// Spark based s3a committers are a  big enough use-case to have this optimization.
		//
		// A sample code to see the improvements is as follows, this sample code is
		// simply a read on JSON from MinIO and write it back as "parquet".
		//
		// import org.apache.spark.sql.SparkSession

				suite.SetUpAccMgmtPlugin(c)

				suite.TestAccMgmtPlugin(c)
			},
		)
	}
}

// TestAccMgmtPlugin - this test assumes that the access-management-plugin is
// the same as the example in `docs/iam/access-manager-plugin.go` -
// specifically, it denies only `s3:Put*` operations on non-root accounts.
func (s *TestSuiteIAM) TestAccMgmtPlugin(c *check) {

  - Add best-effort eviction for DaemonSet pods while scaling down non-empty nodes
  - Add build support for ARM64
  
  ### AliCloud
  - Add missing daemonsets and replicasets to ALI example cluster role
  
  ### Apache CloudStack
  - Add support for Apache CloudStack
  
  ### AWS
  - Regenerate list of EC2 instances
  - Fix pricing endpoint in AWS China Region
  
  ### Azure

  
  Action required for custom scheduler plugin developers:
  Plugins **have to** implement a QueueingHint for Pod/Update event if the rejection from them could be resolved by updating unscheduled Pods themselves.
  Example: suppose you develop a custom plugin that denies Pods that have a `schedulable=false` label. 

 - Insert a debug container in clusters that don’t have ephemeral containers enabled 
 - Modify a crashing container for easier debugging by changing its image, for example to busybox, or its command, for example, to `sleep 1d` so you have time to `kubectl exec`.
 - Inspect configuration files on a node's host filesystem