May 2023 Newsletter

We are pleased to announce the release of HEAVY.AI v7.0. This release brings CPU/GPU parallelism through a new resource manager, machine learning and statistical methods available in HeavyDB, table joins in Immerse , and several other improvements. Full release notes are available here.

Immerse:

• Users can now join tables via user interface, and persist this joined table as a data source in an Immerse dashboard (BETA).
• New Text chart, enabled with feature_flag `ui/enable_new_text_chart`, provides improved styling and customization options.

• 7.0 marks the beta release of HeavyML, a new set of capabilities to execute accelerated machine learning workflows directly from SQL.
• Named model creation is supported via a new CREATE MODEL statement (see the full release notes and documentation for more details)
• Row-wise inference (GPU-accelerated for GPU queries) can be performed via a new ML_PREDICT row-wise operator. This can be used in Immerse custom SQL and persisted into dashboards, allowing end-users to consume models without needing to know how to create or administer them.

• New Executor Resource Manager (ERM)  allows for CPU queries to run fully in parallel, and one or more CPU queries to run in parallel while a GPU query is executing (parallel GPU query kernel execution is not supported yet).
• Support execution of CPU queries where the input datasets do not fit into the CPU buffer pool by executing on a fragment-by-fragment basis, paging from storage.
• ERM is defaulted ON, but can be disabled with configuration `enable-executor-resource-mgr=0`

• The 7.0 release requires
  • NVIDIA 520.61.05 (CUDA 11.8)  or greater drivers installed on your system before upgrading. See NVIDIA’s documentation for driver installation/upgrade guidance.
  • Ubuntu 20.04 or greater. Ubuntu 18.04 is no longer recommended.
• A data set which has been upgraded to v7.0 cannot be rolled back to v6.x. It is critical that a full and complete backup of your $HEAVYAI_PATH folder is prepared prior to upgrade.
   

• Tarball: heavyai-ee-7.0.0-20230501-be4f51b048-Linux-x86_64-render.tar.gz
• Docker: heavyai-ee-7.0.0-20230501-be4f51b048-Linux-x86_64-render-docker.tar.gz

• Tarball: heavyai-ee-7.0.0-20230501-be4f51b048-Linux-x86_64-cpu.tar.gz
• Docker: heavyai-ee-7.0.0-20230501-be4f51b048-Linux-x86_64-cpu-docker.tar.gz

• Docker String Dictionary Server: heavyai-ee-7.0.0-20230501-be4f51b048-Linux-x86_64-sds-docker.tar.gz
• Linux ODBC: HEAVYAIODBC-7.0.0-20230501-be4f51b048-Linux.tar.gz
• Windows ODBC: HEAVYAIODBC-7.0.0-20230501-be4f51b048-Windows.zip

Knowledge Base Highlights at HEAVY.AI

• Connecting to HeavyDB using PowerBI-Desktop
• Adding GPU Monitoring Using HeavyConnect and System Tables

Thought Leadership at HEAVY.AI

At Mobile World Congress in Barcelona.  Dr. Mike Flaxman, HEAVY.AI Head of Product, facilitated an engaging session: How are Digital Twins Driving Network Transformation?

We invite you to watch this session and learn more about digital twins and how they can help communication service, energy and infrastructure providers model and drive faster, more efficient transformation.

Resources at HEAVY.AI

Please visit our Support Portal for access to our Knowledge Base, Community and your support tickets.