[ Bottom of Page | Previous Page | Contents | Index | Library Home | Legal | Search ]

Performance Management Guide

Index

Numerics
A B C D E F G H I J K L M N O P Q R S T U V W X Y

Numerics

3dmon

A

about this book
accessing POWER-based timer registers
accessing POWER timer registers
accessing the processor timer
adapter statistics
adapter transmit and receive queue tuning
administration
- CPU-efficient
  - user ID
affinity
- processor
agent
- performance recording
AIXTHREAD_COND_DEBUG
AIXTHREAD_ENRUSG
AIXTHREAD_GUARDPAGES
AIXTHREAD_MINKTHREADS
AIXTHREAD_MNRATIO
AIXTHREAD_MUTEX_DEBUG
AIXTHREAD_RWLOCK_DEBUG
AIXTHREAD_SCOPE
AIXTHREAD_SLPRATIO
AIXTHREAD_STK=n
allocation and reclamation of paging space slots
alstat tool
Amdahl's law
analyzing network performance
analyzing performance with the trace facility
analyzing programs for CPU use
analyzing the PDT report
application
- parallelizing
application tuning
assessing memory requirements
asynchronous connections
- tuning
asynchronous I/O tunable parameters
ATM performance tuning
attributes
- file
- logical volume
  - changing
- tunable
  - streams
azizo

B

benchmarking
- performance
benchmarks
- industry-standard
binding
- processor
biod Count
biod daemon

C

C and C++ coding style
cache
- effective use
- fast write
- limiting enhanced JFS
- limiting persistent file
cache and TLBs
cache coherency
caches
calls
- sync/fsync
classes of workload
code-optimization techniques
coherency
- cache
combehind
commands
- bindprocessor
  - considerations
- cautions
  - time and timex
- CPU
  - acctcom
  - iostat
  - ps
  - sar
  - time
  - vmstat
- disk
  - filemon
  - filespace
  - lslv (1), (2)
  - sar
  - vmstat (1), (2), (3)
- disk I/O
  - iostat
- examples
- fdpr
- filemon
  - global reports
- ftp
- gprof
- id
- iostat
- ipfilter
- ipreport
- lockstat
- memory
  - interpreting rmss results
  - ps
  - rmss
  - rmss guidelines
  - schedtune
  - svmon
  - vmstat
  - vmtune
- mkpasswd
- monitoring and tuning
- netpmon
- netstat (1), (2)
- nfsstat
- no
- performance analysis
- performance reporting
- performance tuning
- ping
- pprof
- prof
- sar
- schedtune
- schedtune -s
- tprof
- traceroute
- vmstat
- vmtune
compiler execution time
compiler optimization techniques
compiling for floating-point performance (-qfloat)
compiling for specific hardware platforms (1), (2)
compiling with optimization
components
- workload
  - identifying
compression
computer
- how fast
configuration
- expanding
considerations
- programming
contention
- memory and bus
CPU
- determining speed
- monitoring and tuning
- monitoring use
CPU-intensive programs
- identifying
CPU-limited programs
CPU options
CPU time ratio
critical resources
- identifying
- minimizing requirements
cross invalidate
currently dispatched threads
customizing PDT

D

daemons
- cron
data serialization
deferred allocation algorithm
demux layer
designing and implementing efficient programs
determining CPU speed
device related slowdown
direct I/O
- performance
  - example
  - reads
  - writes
- tuning
directory organization
disk adapter outstanding-requests limit
disk and disk adapter tunable parameters
disk drive queue depth
disk I/O
- assessing disk performance
- assessing overall
- asynchronous
  - tuning
- detailed analysis
- miscellaneous tuning parameters
- monitoring
- monitoring and tuning
- pre-tuning baseline
- raw
- summary
- wait time reporting
disk I/O pacing
disk mirroring
- striped
disk striping
- designing
- tuning logical volume
dispatchable threads
dynamics
- program execution
- system

E

F

facility
- trace
FDDI performance tuning
file syncs
- tuning
file system
- reorganizing
file systems
- monitoring and tuning
- types
filemon reports
files
- attributes
  - changing for performance
- compression
- fragment size
- mapped
finding memory-leaking programs
fixed-disk storage management
fixed disks
fork () retry interval parameter
- tuning
fragmentation
- disk
  - assessing
free list
ftp

G

H

I J

K

kernel
- tunable parameters
kernel threads
- CPU usage
  - measuring
kernel tunable parameters

L

LAN slowdown
late allocation algorithm
LDR_CNTRL
levels of optimization
libraries
link order
- specifying
linking
- dynamic
- static
locale
- national language support
- setting
locks
- complex
- granularity
- overhead
- simple
- types of
- waiting for
log logical volume
- creating
- reorganizing
logical volume
- allocation
- attributes
  - changing for performance
- designing
- mirror write consistency
- mirrored
- relocation
- reorganizing
- scheduling policy
- striping size
- tuning
  - I/O
- write verify

M

MALLOCBUCKETS
MALLOCMULTIHEAP
management
- real-memory
managing tunable SP parameters
managing workload
mapped files
maxbuf
maxclient
maxmbuf
maxperm
maxreqs
maxservers
maxuproc
measuring CPU usage of kernel threads
measuring CPU use
memory
- AIX Memory Affinity Support
- assessing requirements
- computational versus file
- determining use
- extended shared
- monitoring and tuning
- requirements
  - calculating minimum
- using shared
- who is using
memory and bus contention
memory-limited programs
memory load control algorithm
memory load control facility
- VMM
memory load control tuning
- the e parameter
- the h parameter
- the m parameter
- the p parameter
- the w parameter
memory mapped files
memory use determination
methods
- chosing page space allocation
minfree and maxfree settings
minimizing critical-resource requirements
minperm
minperm and maxperm settings
minservers
mirroring
- disk
  - striped
mode switching
monitoring
- performance
monitoring and tuning commands and subroutines
monitoring and tuning communications I/O use
monitoring and tuning CPU use
monitoring and tuning disk I/O use
monitoring and tuning file systems
monitoring and tuning java
monitoring and tuning memory use
monitoring and tuning NFS use
monitoring disk I/O
monitoring Java
mountd
msgmax
msgmnb
msgmni
msgmnm
MTU
multiprocessing
- introduction to
- types of
multiuser

N

name resolution tuning
national language support (NLS)
- locale versus speed
ncargs
netstat
network
- tunable parameters
network file system (NFS)
network tunable parameters
- maxmbuf
- MTU
- option
- rfc1323
NFS
- see network file system (NFS) (1), (2)
NFS option tunable parameters
nfsd
nfsd Count
nice (1), (2)
NLS
- see national language support (NLS)
NODISCLAIM
non-shared
npswarn and npskill settings
NSORDER
numclust

O

objectives
- performance
- setting
offline processing
optimizing processors for FORTRAN and C
options
- thread
- useful CPU
outputs
- correlating svmon and ps
- correlating svmon and vmstat
overhead
- locking
- reducing memory scanning

P

pacing
- disk I/O
  - example
page replacement
page space allocation
- deferred
- early
- late
page space allocation methods
paging-space
- assessing I/O
- tuning
paging space and virtual memory
paging space slots
- allocation and reclamation
paging spaces
- placement and sizes
parameters
- tunable
PDT
- see performance diagnostic tool (PDT) (1), (2), (3)
performance
- device specific slowdown
- disk
  - assessing (1), (2)
  - CPU report
  - drive report
  - tty report
- disk mirroring
- disk or memory determination
- implementing
- initial diagnosis
- installation guidelines
- issues
  - SMP
- LAN slowdown
- limiting resource identification
- network
  - analyzing
- objectives
- planning for
- problem determination
- program
  - limiting factor determination
- recording
- slowdown
  - specific program
  - unpredictable
- subroutines
- time of day
- tuning
  - TCP and UDP
- user specific slowdown
performance analysis commands
performance benchmarking
performance concepts (1), (2)
performance diagnostic tool (PDT)
- customizing
- enabling
- installing
- measuring the baseline
- report analyzing
- report message
- scope of analysis
- structure
- uninstalling
- using
- using the
performance enhancements
- JFS and enhanced JFS
performance inhibitors
performance monitoring
- continuous
  - case for
performance problem
- description
- reporting
performance problems
- reporting
performance-related installation guidelines
performance-related subroutines
performance reporting commands
performance requirements
- documenting
performance toolbox (PTX)
- using the
performance tuning
- related information
performance-tuning
- introducing
performance tuning commands
physical volume
- considerations
- maximum number
- position
- range
ping
pinned storage
- misuse of
pipeline and registers
placement
- assessing file
- assessing physical data
planning and implementing for performance
platforms
- compiling for specific
portmap
POWER-based-architecture-unique timer access
prebound subroutine libraries
preprocessors and compilers
- effective use of
previously captured data
- display
priority
- process and thread
problems
- performance
  - reporting
process
- priority
processes and threads
processing
- offline
processor affinity and binding
processor scheduler
- performance overview
processor time slice
- scheduler
processor timer
- accessing
profile directed feedback (PDF)
profiling
program execution dynamics
programming considerations
programs
- CPU-intensive
  - identifying
- efficient
- executable
  - restructuring
- fdpr
- finding memory-leaking
- memory-limited
- rebindable executable
- tprof
- xmpert
ps command
PSALLOC
PTX
- see performance toolbox (PTX)
ptxrlog
ptxsplit
ptxtab

Q

queue limits
- disk device
- scsi adapter

R

RAID
- see redundant array of independent disks (RAID)
RAM
RAM disk
real memory
real-memory management
rebindable executable programs
recording
- performance
redundant array of independent disks (RAID)
- levels
- summary
registers and pipeline (1), (2)
related publications
renice
repaging
reporting performance problems
reports
- filemon
requirements
- performance
  - documenting
- workload
  - resource
resource allocation
- reflecting priorities
resource management overview
resources
- applying additional
- critical
  - identifying
responding to PDT report messages
response time
- SMP (1), (2)
restructuring executable programs
rfc1323
RT_GRQ
RT_MPC
rules
- simplifying
run queue
- scheduler

S

T

TCP/IP
- performance overview
TCP layer
tcp_mssdflt
tcp_nodelay
tcp_recvspace
tcp_sendspace (1), (2)
TCP socket buffer tuning
test case scenarios
test cases
text highlighting
thread (1), (2)
- priority
- scheduling policy
- support
thread option
thread support tunable parameters
thread tuning
threads
- environment variables
- kernel
  - measuring CPU usage
- SMP
  - scheduling
- tuning
threads and processes
thresholds
- VMM
throughput
- SMP
throughput scalability
- SMP
time of day slowdown
time ratio
timer
- C subroutine
- processor
  - accessing
timer access
- POWER-based-architecture-unique
timer registers
- POWER
  - assembler routines
- POWER-based
  - accessing
timing commands
TLB see translation lookaside buffer
token-ring (16 MB) performance tuning
token-ring (4 MB) performance tuning
tools
- alstat
- emstat
- SMP
topas monitor
- using the
trace channels
trace data
- formatting
- limiting
- viewing
trace events
- adding new
trace facility
- analyzing performance with
- controlling
- event IDs
- example
- implementing
- report formatting
- starting
- starting and controlling (1), (2)
- understanding
trace file
- example
- sample
  - formatting
  - obtaining
trace report
- filtering
- reading
translation lookaside buffer
trcrpt
tunable attributes
- streams
tunable parameters
- asynchronous I/O
- disk and disk adapter
- interprocess communication
- kernel
- miscellaneous
- network
- nfs option
- scheduler
- summary
- synchronous I/O
- thread support
- virtual memory manager
tuning
- adapter queue
- application
- async port
- asynchronoous connections
- ATM
- ethernet
- FDDI
- hand
- HIPPI
- IP
- mbuf pool
- name resolution
- network memory
- SOCC
- SP network
- TCP and UDP
- TCP maximum segment size
- TCP socket buffer
- thread
- token-ring (16 MB)
- token-ring (4 MB)
- UDP socket buffer
tuning Java
tuning logical volume striping
tuning mbuf pool performance
tuning memory pools
tuning paging-space thresholds
tuning TCP and UDP performance
tuning TCP maximum segment size
tuning VMM memory load control
tuning VMM page replacement
tuning with vmtune

U

UDP
- see user datagram protocol (UDP)
UDP layer
UDP socket buffer tuning
understanding the trace facility
unpredictable performance
user datagram protocol (UDP)
- performance overview
user ID
- administration for CPU efficiency
using disk I/O pacing
using performance diagnostic tool (PDT)

V

variables
- environment
vertual memory manager
- tunable parameters
virtual memory and paging space
virtual memory manager (VMM)
virtual memory manager tunable parameters
VMM
- see virtual memory manager (VMM)
vmstat command
volume group
- considerations
- mirroring

W

waiting threads
what is java
who should use this book
why java
wlmperf
workload
workload concurrency
- SMP
workload multiprocessability
- SMP
workload resource requirements
- estimating
workloads
- identifying
- SMP
workstation
write-around
write behind
- VMM

X

Y

YIELDLOOPTIME

[ Top of Page | Previous Page | Contents | Index | Library Home | Legal | Search ]