Beowulf cluster computing with Windows Book

	Series Foreword
	Foreword
	Preface
1	Introduction	1
1.1	Definitions and Taxonomy	1
1.2	Opportunities and Advantages	3
1.3	A Short History	6
1.4	Elements of a Cluster	8
1.5	Description of the Book	10
I	Enabling Technologies
2	An Overview of Cluster Computing	15
2.1	A Taxonomy of Parallel Computing	16
2.2	Hardware System Structure	19
2.3	Node Software	25
2.4	Resource Management	25
2.5	Distributed Programming	27
2.6	Conclusions	29
3	Node Hardware	31
3.1	Overview of a Beowulf Node	32
3.2	Processors	38
3.3	Motherboard	41
3.4	Memory	43
3.5	BIOS	46
3.6	Secondary Storage	47
3.7	PCI Bus	49
3.8	Example of a Beowulf Node	50
3.9	Boxes, Shelves, Piles, and Racks	50
3.10	Node Assembly	52
4	Windows 2000	61
4.1	Introduction to Windows 2000	61
4.2	Operating System Model	61
4.3	System Architecture	72
4.4	Win32 API	91
4.5	Tools to Explore Windows 2000	91
5	Network Hardware	95
5.1	Interconnect Technologies	95
5.2	A Detailed Look at Ethernet	100
5.3	Network Practicalities: Interconnect Choice	106
6	Windows 2000 Networking	113
6.1	Windows 2000 Networking Components	113
6.2	Networking APIs	114
6.3	Network-Resource Name Resolution	132
6.4	Protocol Drivers	138
6.5	NDIS Drivers	141
6.6	Conclusion	148
7	Setting Up Clusters: Installation and Configuration	149
7.1	Resource Considerations	149
7.2	Basic Windows 2000 Cluster Installation Procedure	154
7.3	Off and Running	155
8	How Fast Is My Beowulf?	157
8.1	Metrics	157
8.2	Ping-Pong Test	160
8.3	The LINPACK Benchmark	160
8.4	The NAS Parallel Benchmark Suite	162
II	Parallel Programming
9	Parallel Programming with MPI	167
9.1	Hello World in MPI	168
9.2	Manager/Worker Example	175
9.3	Two-Dimensional Jacobi Example with One-Dimensional Decomposition	180
9.4	Collective Operations	184
9.5	Parallel Monte Carlo Computation	189
9.6	Installing MPICH under Windows 2000	189
9.7	Tools	192
9.8	MPI Implementations for Clusters	193
9.9	MPI Routine Summary	193
10	Advanced Topics in MPI Programming	199
10.1	Dynamic Process Management in MPI	199
10.2	Fault Tolerance	202
10.3	Revisiting Mesh Exchanges	204
10.4	Motivation for Communicators	211
10.5	More on Collective Operations	213
10.6	Parallel I/O	215
10.7	Remote Memory Access	221
10.8	Using C++ and Fortran 90	224
10.9	MPI, OpenMP, and Threads	226
10.10	Measuring MPI Performance	227
10.11	MPI-2 Status	229
10.12	MPI Routine Summary	229
11	Parallel Programming with PVM	235
11.1	Overview	235
11.2	Program Examples	240
11.3	Fork/Join	240
11.4	Dot Product	244
11.5	Matrix Multiply	249
11.6	One-Dimensional Heat Equation	255
11.7	Using PVM	263
11.8	PVM Console Details	268
11.9	Host File Options	270
11.10	XPVM	273
12	Fault-Tolerant and Adaptive Programs with PVM	279
12.1	Considerations for Fault Tolerance	280
12.2	Building Fault-Tolerant Parallel Applications	281
12.3	Adaptive Programs	287
III	Managing Clusters
13	Cluster Workload Management	299
13.1	Goal of Workload Management Software	299
13.2	Workload Management Activities	300
14	Condor: A Distributed Job Scheduler	307
14.1	Introduction to Condor	307
14.2	Using Condor	312
14.3	Condor Architecture	326
14.4	Installing Condor under Windows 2000	330
14.5	Configuring Condor	332
14.6	Administration Tools	337
14.7	Cluster Setup Scenarios	339
14.8	Conclusion	343
15	Maui Scheduler: A Multifunction Cluster Scheduler	345
15.1	Overview	345
15.2	Installation and Initial Configuration	346
15.3	Advanced Configuration	347
15.4	Steering Workload and Improving Quality of Information	359
15.5	Troubleshooting	361
15.6	Conclusions	361
16	PBS: Portable Batch System	363
16.1	History of PBS	363
16.2	Using PBS	367
16.3	Installing PBS	372
16.4	Configuring PBS	373
16.5	Managing PBS	379
16.6	Troubleshooting	381
17	MPI Software Technology, Inc., Cluster CoNTroller	385
17.1	Overview of Cluster CoNTroller	385
17.2	Installation Configuration	386
17.3	Administrative Features	388
17.4	User Commands	389
17.5	Future Enhancements	397
18	Cornell Theory Center	399
18.1	CTC Users	399
18.2	Systems Overview	400
18.3	Dedicated Clusters	402
18.4	Other Cluster Resources	404
18.5	Providing a Consistent User Interface	405
18.6	Helping Users Get Started	409
18.7	Batch and Interactive Job Scheduling	409
18.8	CTC Success Story	409
19	Conclusions	411
19.1	Future Directions for Hardware Components	411
19.2	Future Directions for Software Components	413
19.3	Final Thoughts	416
	A: Glossary of Terms	419
	B: Annotated Reading List	431
	C: Annotated URLs	433
	References	437
	Index	439