Partition and Isolate Approaches for Consolidating HPC and

  • Slides: 30
Download presentation
Partition and Isolate: Approaches for Consolidating HPC and Commodity Workloads Jack Lange Assistant Professor

Partition and Isolate: Approaches for Consolidating HPC and Commodity Workloads Jack Lange Assistant Professor University of Pittsburgh

Summary • Commodity and HPC systems have been converging – Commodity off the shelf

Summary • Commodity and HPC systems have been converging – Commodity off the shelf components – Linux based HPC systems – Cloud computing • Problem: Real HPC applications need HPC environments – Tightly coupled, massively parallel, and synchronized – Current services must provide dedicated HPC systems • Can we co-host HPC applications on commodity systems? • Dual Stack Approach – Provision the underlying software stack along with application – Commodity stack should handle commodity applications – HPC stack can provide HPC environment

User Space Partitioning Memory Socket 1 Socket 2 Cores 1 2 5 6 3

User Space Partitioning Memory Socket 1 Socket 2 Cores 1 2 5 6 3 4 7 8 Commodity Partition Memory HPC Partition • Current systems do support this, but… • Interference still exists inside the system software – Inherent feature of commodity systems

HPC vs. Commodity Systems • Commodity systems have fundamentally different focus than HPC systems

HPC vs. Commodity Systems • Commodity systems have fundamentally different focus than HPC systems – Amdahl’s vs. Gustafson’s laws – Commodity: Optimized for common case • HPC: Common case is not good enough – At large (tightly coupled) scales, percentiles lose meaning – Collective operations must wait for slowest node – 1% of nodes can make 99% suffer – HPC systems must optimize outliers (worst case)

High Performance Computing (HPC) • Large scale simulations to solve Big Problems 5

High Performance Computing (HPC) • Large scale simulations to solve Big Problems 5

Dual Stack Approach • Partition – Segment the underlying hardware resources – Assign them

Dual Stack Approach • Partition – Segment the underlying hardware resources – Assign them to exclusively to specific workloads • Isolate – Prevent interference from other workloads – Hardware: partitions must be course grained – Software: eliminate shared state • Implementation – Independent system software running on isolated resources

HPC in the cloud • Clouds are starting to look like supercomputers… – Are

HPC in the cloud • Clouds are starting to look like supercomputers… – Are we seeing a convergence? • Not yet – – Noise issues Poor isolation Resource contention Lack of control over topology • Very bad for tightly coupled parallel apps – Require specialized environments that solve these problems • Approaching convergence – Vision: Dynamically partition cloud resources into HPC and commodity zones – This talk: partitioning compute nodes with performance isolation

Commodity VMMs • Virtualization is considered an “enterprise” technology – Designed for commodity environments

Commodity VMMs • Virtualization is considered an “enterprise” technology – Designed for commodity environments – Fundamentally different, but not wrong! • Example: KVM architecture issues – Userspace handlers – Fairly complex memory management – Locking and periodic optimizations – Presence of system noise

Palacios VMM • OS-independent embeddable virtual machine monitor – Established compatibility with Linux, Kitten,

Palacios VMM • OS-independent embeddable virtual machine monitor – Established compatibility with Linux, Kitten, and Minix • Specifically targets HPC applications and environments – Consistent performance with very low variance • Deployable on supercomputers, clusters (Infiniband/Ethernet), and servers – 0 -3% overhead at large scales (thousands of nodes) • VEE 2011, IPDPS 2010, ROSS 2011 Open source and freely available http: //www. v 3 vee. org/palacios 9

Palacios/Linux • Palacios/Linux provides lightweight and high performance virtualized environments – Internally manages dedicated

Palacios/Linux • Palacios/Linux provides lightweight and high performance virtualized environments – Internally manages dedicated resources • Memory and CPU scheduling – Does not bother with “enterprise features” • Page sharing/merging, swapping, overcommitting resources • Palacios enables scalable HPC performance on commodity platforms

VMM Comparison • Primary difference: Consistency – Requirement for tightly coupled performance at large

VMM Comparison • Primary difference: Consistency – Requirement for tightly coupled performance at large scale • Example: KVM nested paging architecture – Maintains free page caches to optimize performance • Requires cache management – Shares page tables to optimize memory usage • Requires synchronization VMM % of exits Mean Std Dev # NPFS KVM 52% 8804 5232 3, 265, 156 Palacios 50% 10876 2685 1, 872, 017

Dual Stack Architecture • Partitioning at the OS level Commodity Application(s) HPC Application Commodity

Dual Stack Architecture • Partitioning at the OS level Commodity Application(s) HPC Application Commodity Linux HPC Linux KVM Linux Kernel Palacios VMM Linux Module Interface Palacios Resource Managers Hardware • Enable cloud to host both commodity and HPC apps – Each zone optimized for the target applications

Evaluation • Goal: Measure VM isolation properties • Partitioned a single node into HPC

Evaluation • Goal: Measure VM isolation properties • Partitioned a single node into HPC and commodity zones – Commodity Zone: Parallel Kernel compilation – HPC Zone: Set of standard HPC benchmarks – System: • Dual 6 -core AMD Opteron with NUMA topology • Linux guest environments (HPC and commodity) • Important: Local node only – Does not promise good performance at scale – But, poor performance will magnify at large scales

Results Commodity VMMs degrade with contention Palacios delivers consistent performance Mini. FE: Unstructured implicit

Results Commodity VMMs degrade with contention Palacios delivers consistent performance Mini. FE: Unstructured implicit finite element solver Mantevo Project -- https: //software. sandia. gov/mantevo/index. html

Discussion • A dual stack approach can provide HPC environments on commodity clouds –

Discussion • A dual stack approach can provide HPC environments on commodity clouds – HPC and commodity workloads can dynamically share resources – HPC requirements can be met without fully dedicated resources • Networking is still an open issue – Need mechanisms for isolation and partitioning – Need high performance networking architectures • 1 Gbit is not good enough • 10 Gbit is good, Infiniband is better – Need control over placement and topologies

Multi-stack Operating Systems • Future Exascale Systems are moving towards in situ organization •

Multi-stack Operating Systems • Future Exascale Systems are moving towards in situ organization • Applications traditionally have utilized their own platforms • Visualization, storage, analysis, etc • Everything must now collapse onto a single platform

What this means for the OS • At Petascale we could optimize each environment

What this means for the OS • At Petascale we could optimize each environment separately – Each had their own OS and hardware • At Exascale workloads will be co-located – Can a single OS handle all workloads effectively? • Probably Not – Each has different resource requirements and behaviors – Exascale will need to support multiple OS environments on the same hardware

Current Supercomputer OS architectures UNIX LINUX 18 Source: http: //en. wikipedia. org/wiki/Fil e: Operating_systems_used_on

Current Supercomputer OS architectures UNIX LINUX 18 Source: http: //en. wikipedia. org/wiki/Fil e: Operating_systems_used_on

Will Linux continue to dominate? • An open question at this point – Exascale

Will Linux continue to dominate? • An open question at this point – Exascale systems may mark a radical departure from traditional architectures • Kitten: Open-source Lightweight Kernel from Sandia – Minimal compute node OS • Provides mostly Linux-compatible user environment – Supports unmodified compiler toolchains and ELF executables • But doesn’t include the enterprise Linux features – Simple memory management, application managed I/O, etc 19

Dual Stack Architecture HPC Application Commodity Application(s) Linux Kernel Kitten Palacios VMM Linux Module

Dual Stack Architecture HPC Application Commodity Application(s) Linux Kernel Kitten Palacios VMM Linux Module Interface Palacios Resource Managers Hardware • Provide LWK environment on a commodity system – Each zone optimized for the target applications

Stream Palacios/Kitten provides higher memory throughput than Linux – 400 MB/s (4. 74%) Palacios/Kitten

Stream Palacios/Kitten provides higher memory throughput than Linux – 400 MB/s (4. 74%) Palacios/Kitten provides more consistent memory performance than Linux – 340 MB/s lower standard deviation 21

Selfish Detour Linux Virtual Kitten Palacios/Kitten can reduce noise from Linux • Eliminates Periodic

Selfish Detour Linux Virtual Kitten Palacios/Kitten can reduce noise from Linux • Eliminates Periodic timer interrupts 22

Better than native? • Results are preliminary – Followed best practices for configuring Linux

Better than native? • Results are preliminary – Followed best practices for configuring Linux – Didn’t try to optimize VMM performance • Virtualization can improve system performance – when the system is running a commodity OS B. Kocoloski, J. Lange, Better than Native: Using Virtualization to Improve Compute Node Performance, (ROSS 2012)

Beyond Virtualization • Virtualization imposes overhead – Power: requires transistors – Performance: small, but

Beyond Virtualization • Virtualization imposes overhead – Power: requires transistors – Performance: small, but present – Interference: Still some dependencies on host OS • Might not be available on exascale hardware • Can we provide native partitioning? – We think so – Linux provides the ability to dynamically remove resources (CPUs, memory, etc) – These can be taken over by a second OS

Dual Stack Architecture Commodity Application(s) Linux HPC Application Palacio VMM Kitten Hardware • Provide

Dual Stack Architecture Commodity Application(s) Linux HPC Application Palacio VMM Kitten Hardware • Provide LWK environment on a commodity system – Each zone optimized for the target applications

Approach • OS partition created via offlined resources – CPUs, memory, PCI devices •

Approach • OS partition created via offlined resources – CPUs, memory, PCI devices • Secondary OS “booted” on offline resources • Issues: – OS initialization • Boot process • Resource discovery – Coordination and communication – Security and safety

Dual Stack Memory • Maybe we don’t need to provide an entirely separate OS

Dual Stack Memory • Maybe we don’t need to provide an entirely separate OS – Instead selectively manage some resources • Dual stack memory – Provide a separate memory management layer to Linux • Features – Selectively manage heap per application – Provide applications with direct control over memory layout – Transparently back memory using large pages – Without overhead added by Linux

Dual Stack Architecture Commodity Application(s) HPC Application Linux Memory Management Hardware • Provide LWK

Dual Stack Architecture Commodity Application(s) HPC Application Linux Memory Management Hardware • Provide LWK memory manager on a commodity OS

Conclusion • Commodity systems are not designed to support HPC workloads – Different requirements

Conclusion • Commodity systems are not designed to support HPC workloads – Different requirements and behaviors than commodity applications • A multi stack approach can provide HPC environments in commodity systems – HPC requirements can be met without separate physical systems – HPC and commodity workloads can dynamically share resources – Isolated system software environments

Thank you Jack Lange Assistant Professor University of Pittsburgh jacklange@cs. pitt. edu http: //www.

Thank you Jack Lange Assistant Professor University of Pittsburgh jacklange@cs. pitt. edu http: //www. cs. pitt. edu/~jacklange