Berkeley Cluster Projects David E Culler cullercs berkeley
Berkeley Cluster Projects David E. Culler culler@cs. berkeley. edu http: //now. cs. berkeley. edu/ 11/23, 1998 NOW 1
Goals • Make a fundamental change in how we design and construct large-scale systems – market reality: » 50%/year performance growth => cannot allow 1 -2 year engineering lag – technological opportunity: » single-chip “Killer Switch” => fast, scalable communication • Highly integrated building-wide, campus-wide systems • Explore novel system design concepts in this new “cluster” paradigm NOW 2
100 node Ultra/Myrinet NOW 3
Fast Communication Challenge Killer Platform ns ms µs ° ° ° Comm. . Software Comm. Software Network Interface Hardware Killer Switch • Fast processors and fast networks • The time is spent in crossing between them NOW 4
Opening: Intelligent Network Interfaces • Dedicated Processing power and storage embedded in the Network Interface M M $ P M P $ $ $ P P P 160 MB/s Myricom NIC • An I/O card today • Tomorrow on chip? M Mryicom Net P M I/O bus (S-Bus) 50 MB/s $ P Sun Ultra 170 NOW 5
NOW System Architecture Parallel Apps Large Seq. Apps Sockets, Split-C, MPI, HPF, v. SM Global Layer UNIX Resource Management Network RAM Distributed Files Process Migration UNIX Workstation Comm. SW UNIX Workstation Comm. SW Net Inter. HW Fast Commercial Switch (Myrinet) NOW 6
Communication Performance Direct Network Access Latency 1/BW • Log. P: Latency, Overhead, and Bandwidth • Active Messages: lean layer supporting programming models NOW 7
World-Record Disk-to-Disk Sort • Sustain 500 MB/s disk bandwidth and 1, 000 MB/s network bandwidth NOW 8
Massive Cheap Storage • Basic unit: 2 PCs double-ending four SCSI chains Currently serving Fine Art at http: //www. thinker. org/imagebase/ NOW 9
Cluster of SMPs (CLUMPS) • Four Sun E 5000 s – 8 processors – 3 Myricom NICs • Multiprocessor, Multi. NIC, Multi-Protocol NOW 10
Information Servers • Basic Storage Unit: – Ultra 2, 300 GB raid, 800 GB tape stacker, ATM – scalable backup/restore • Dedicated Info Servers – web, – security, – mail, … • VLANs project into dept. NOW 11
Millennium Computational Community Business SIMS BMRC Chemistry C. S. E. E. Biology Gigabit Ethernet Astro NERSC M. E. Physics N. E. IEOR C. E. MSME Transport Economy Math NOW 12
Millennium PC Clumps • Inexpensive, easy to manage Cluster • Replicated in many departments • Prototype for very large PC cluster NOW 13
Proactive Infrastructure Information appliances Stationary desktops Scalable Servers NOW 14
- Slides: 14