Pull cuda image from nvcr #20622
Merged
Pull cuda image from nvcr #20622
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
In the event we are rate-limited or restricted from pulling from DockerHub, this should work around the problem by pulling from Nvidia's site. Issue runtimes/automation/122 Signed-off-by: Adam Brousseau <adam.brousseau@ca.ibm.com>
|
jenkins compile plinux jdk17 |
|
Rebuilt on a machine that didn't already have the cuda image built locally I am able to pull the container by hand from internal without creds as well so I think this is good to merge once 714 passes. |
pshipton
approved these changes
Nov 18, 2024
AdamBrousseau
added a commit
to AdamBrousseau/ci-jenkins-pipelines
that referenced
this pull request
Nov 19, 2024
Sometimes pulls fail because DockerHub is rate limited. We are also facing pull restrictions from DockerHub for Semeru. Related eclipse-openj9/openj9#20622 Signed-off-by: Adam Brousseau <adam.brousseau@ca.ibm.com>
h3110n3rv3
added a commit
to h3110n3rv3/openj9
that referenced
this pull request
Nov 19, 2024
commit 44766de Merge: 975e741 fe89ec5 Author: Marius Pirvu <mpirvu@ca.ibm.com> Date: Tue Nov 19 09:18:48 2024 -0500 Merge pull request eclipse-openj9#20614 from JamesKingdon/trampolineResetFix Reset trampoline pointers to top of space commit 975e741 Merge: 9618d90 9f566fa Author: Peter Shipton <Peter_Shipton@ca.ibm.com> Date: Tue Nov 19 08:34:48 2024 -0500 Merge pull request eclipse-openj9#20631 from tajila/jfr3 Disable JFR JCL APIs commit 9618d90 Merge: 8f240e6 36dc396 Author: Dmitri Pivkine <Dmitri_Pivkine@ca.ibm.com> Date: Mon Nov 18 21:45:37 2024 -0500 Merge pull request eclipse-openj9#20630 from amicic/VLHGC_barrier_32bit Don't use Offheap APIs for 32 bit commit 9f566fa Author: tajila <atobia@ca.ibm.com> Date: Mon Nov 18 20:05:47 2024 -0500 Disable JFR JCL APIs Disable JFR JCL APIs until JFR natives are implemented. Related to: eclipse-openj9#20607 Signed-off-by: tajila <atobia@ca.ibm.com> commit 8f240e6 Merge: 7534b4e 6662a08 Author: Keith W. Campbell <keithc@ca.ibm.com> Date: Mon Nov 18 20:05:46 2024 -0500 Merge pull request eclipse-openj9#20621 from tajila/jfr2 Fix conditon in JFR buffer test commit 6662a08 Author: tajila <atobia@ca.ibm.com> Date: Mon Nov 18 15:49:59 2024 -0500 Fix conditon in JFR buffer test Signed-off-by: tajila <atobia@ca.ibm.com> commit 36dc396 Author: Aleksandar Micic <Aleksandar_Micic@ca.ibm.com> Date: Mon Nov 18 19:11:09 2024 -0500 Don't use Offheap APIs for 32 bit A couple of Offheap APIs are guarded with 64 bit compile flag. A more accurate fix would be to guard them with Offheap specific build flag (and do not compile the whole dir for 32bit), but it would require more complex changes. Signed-off-by: Aleksandar Micic <Aleksandar_Micic@ca.ibm.com> commit 7534b4e Merge: 3d719d4 276f38f Author: Graham Chapman <graham_chapman@ca.ibm.com> Date: Mon Nov 18 18:14:00 2024 -0500 Merge pull request eclipse-openj9#20620 from tajila/jfr3 Release jfrSampler lock before acquiring VMaccess commit 3d719d4 Merge: 247d6ec c0ff1fc Author: Peter Shipton <Peter_Shipton@ca.ibm.com> Date: Mon Nov 18 17:17:42 2024 -0500 Merge pull request eclipse-openj9#20622 from AdamBrousseau/cuda_docker_non_dockerhub Pull cuda image from nvcr commit 247d6ec Merge: 068838e fee4aeb Author: Aleksandar Micic <amicic@ca.ibm.com> Date: Mon Nov 18 16:55:50 2024 -0500 Merge pull request eclipse-openj9#20514 from LinHu2016/coding-std-update2 Clean up MethodTypesIterator class commit 068838e Merge: d98d498 3dc8eb4 Author: Aleksandar Micic <amicic@ca.ibm.com> Date: Mon Nov 18 16:46:15 2024 -0500 Merge pull request eclipse-openj9#20564 from LinHu2016/off-heap-incremental4 New vm/gc API j9gc_objaccess_indexableDataDisplacement commit c0ff1fc Author: Adam Brousseau <adam.brousseau@ca.ibm.com> Date: Mon Nov 18 16:03:35 2024 -0500 Pull cuda image from nvcr In the event we are rate-limited or restricted from pulling from DockerHub, this should work around the problem by pulling from Nvidia's site. Issue runtimes/automation/122 Signed-off-by: Adam Brousseau <adam.brousseau@ca.ibm.com> commit 3dc8eb4 Author: lhu <linhu@ca.ibm.com> Date: Fri Nov 8 11:50:32 2024 -0500 New vm/gc API j9gc_objaccess_indexableDataDisplacement new API j9gc_objaccess_indexableDataDisplacement used by the JIT. Returns the displacement for the data of moved array object. For adjacent Array and 0 size array, displacement = dst - src For Off-heap Array, displacement = 0 should only be called for off-heap eanbled case. Signed-off-by: lhu <linhu@ca.ibm.com> commit d98d498 Merge: 77f153d ee35ed5 Author: Graham Chapman <graham_chapman@ca.ibm.com> Date: Mon Nov 18 14:52:20 2024 -0500 Merge pull request eclipse-openj9#20611 from tajila/jfr2 Prevent writes to free'd JFR buffer commit 77f153d Merge: 038df36 326a800 Author: Marius Pirvu <mpirvu@ca.ibm.com> Date: Mon Nov 18 14:09:15 2024 -0500 Merge pull request eclipse-openj9#20592 from luke-li-2003/ServerCacheReflectClass Cache Client's Reflect Class Pointers at JITServer commit 276f38f Author: tajila <atobia@ca.ibm.com> Date: Mon Nov 18 13:59:52 2024 -0500 Release jfrSampler lock before acquiring VMaccess Signed-off-by: tajila <atobia@ca.ibm.com> commit ee35ed5 Author: tajila <atobia@ca.ibm.com> Date: Sun Nov 17 16:06:08 2024 -0500 Prevent writes to free'd JFR buffer Also, check that threadObject is not NULL before dereference. Signed-off-by: tajila <atobia@ca.ibm.com> commit 038df36 Merge: 224374d 0573107 Author: Aleksandar Micic <amicic@ca.ibm.com> Date: Mon Nov 18 11:12:43 2024 -0500 Merge pull request eclipse-openj9#20596 from LinHu2016/off-heap-incremental6 Update verbose GC for off-heap commit 224374d Merge: 6289c28 d56338f Author: Babneet Singh <babsingh@users.noreply.github.com> Date: Mon Nov 18 10:20:28 2024 -0500 Merge pull request eclipse-openj9#20606 from ThanHenderson/omr-unnamed Fix OMR_GLUE unnamed spelling mistake commit 6289c28 Merge: efaa4f6 1af5097 Author: Tobi <atobia@ca.ibm.com> Date: Mon Nov 18 09:56:54 2024 -0500 Merge pull request eclipse-openj9#20605 from pshipton/cracbean Move CRaCMXBeanImpl and export jdk.crac.management commit 1af5097 Author: Peter Shipton <Peter_Shipton@ca.ibm.com> Date: Fri Nov 15 14:28:45 2024 -0500 Move CRaCMXBeanImpl and export jdk.crac.management Issue eclipse-openj9#20587 Signed-off-by: Peter Shipton <Peter_Shipton@ca.ibm.com> commit 326a800 Author: Luke Li <luke.li@ibm.com> Date: Thu Nov 14 16:13:54 2024 -0500 Remove J9Class Occurrences in Vector API Substitute uses of J9Class in the vector API with TR_OpaqueClassBlock with frontend references. The Vector API, being a part of the optimizer, should never use the inside of a J9Class Signed-off-by: Luke Li <luke.li@ibm.com> commit a97a4ba Author: Luke Li <luke.li@ibm.com> Date: Mon Oct 28 09:31:53 2024 -0400 Remove Server VM Access Request in VectorAPIExpansion.cpp The server should make a specific request to the client in getJ9ClassFromClassNode() and getVectorSizeFromVectorSpecies(), instead of requesting a pointer from the KnownObjectTable as that requires VM Access and will cause a fatal assertion fail at J9KnownObjectTable.cpp:207. Signed-off-by: Luke Li <luke.li@ibm.com> commit fe89ec5 Author: James Kingdon <jkingdon@ca.ibm.com> Date: Fri Nov 15 15:24:35 2024 -0500 Reset trampoline pointers to top of space After flushing the code cache under FSD there is a call to resetTrampolines. The _trampolineAllocationMark and _trampolineReservationMark need to be reset to their starting positions but were incorrectly set to the end of the trampoline space. As a result all active segments at the time of the cache flush are marked as full (assuming a platform that uses trampolines). This can be seen by taking a JIT verbose log with option `codecache` which will show CODECACHE: CodeCache 00007E759000C660 marked as full in reserveSpaceForTrampoline commit d56338f Author: Nathan Henderson <nathan.t.henderson@outlook.com> Date: Fri Nov 15 11:32:46 2024 -0800 Fix OMR_GLUE unnamed spelling mistake Signed-off-by: Nathan Henderson <nathan.henderson@ibm.com> commit 0573107 Author: lhu <linhu@ca.ibm.com> Date: Tue Nov 12 12:09:29 2024 -0500 Update verbose GC for off-heap -new outputOffHeapInfo in CopyForward and Mark for balancedGC <offheap candidates="xxx" cleared="xxx" />" - output offheap-objects in MemoryInfo <offheap-objects objects="xxx" bytes="xxx" /> Signed-off-by: lhu <linhu@ca.ibm.com> commit fee4aeb Author: lhu <linhu@ca.ibm.com> Date: Tue Nov 5 10:51:09 2024 -0500 Clean up MethodTypesIterator class Use consistent data types and code formatting Signed-off-by: lhu <linhu@ca.ibm.com>
karianna
pushed a commit
to adoptium/ci-jenkins-pipelines
that referenced
this pull request
Nov 21, 2024
* Switch to ubi8 base cuda image The ubuntu image does not support ppc64le arch. We are now building (Semeru) on ppc64le in the cent7 container and require the cuda libs. Related ibmruntimes#228 Signed-off-by: Adam Brousseau <adam.brousseau@ca.ibm.com> * Pull cuda image from nvcr Nvidia Sometimes pulls fail because DockerHub is rate limited. We are also facing pull restrictions from DockerHub for Semeru. Related eclipse-openj9/openj9#20622 Signed-off-by: Adam Brousseau <adam.brousseau@ca.ibm.com> --------- Signed-off-by: Adam Brousseau <adam.brousseau@ca.ibm.com>
AdamBrousseau
added a commit
to AdamBrousseau/ci-jenkins-pipelines
that referenced
this pull request
Dec 12, 2024
In the event we are rate-limited or restricted from pulling from DockerHub, this should work around the problem by pulling from Nvidia's site. Issue runtimes/automation/122 Related eclipse-openj9/openj9#20622 Signed-off-by: Adam Brousseau <adam.brousseau@ca.ibm.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
In the event we are rate-limited or restricted from pulling from DockerHub, this should work around the problem by pulling from Nvidia's site.
Issue runtimes/automation/122