Lines Matching +full:high +full:- +full:performance

30 .Nd performance tuning under FreeBSD
31 .Sh SYSTEM SETUP - DISKLABEL, NEWFS, TUNEFS, SWAP
63 partitions are read-mostly, with very little writing, while
68 heavily write-loaded partitions will not bleed over into the mostly-read
81 .Dq Li "tunefs -n enable /filesystem" .
82 Softupdates drastically improves meta-data performance, mainly file
100 There is no loss of performance since the root
103 A number of run-time
117 file systems normally update the last-accessed time of a file or
138 atime turned on for mostly read-only partitions such as
151 the performance of a file system by splitting I/O operations across two
161 or essentially read-only partitions such as
164 You should only stripe partitions that require serious I/O performance,
170 File systems tend to store meta-data on power-of-2 boundaries
173 means you want to use a large off-center stripe size such as 1152 sectors
174 so sequential I/O does not seek both disks and so meta-data is distributed
179 run-time.
195 reservation, both total for system and per-user.
233 improve performance on systems where space for mapping pipe buffers
236 to fall back to using double-copy.
249 of internal memory management page-tracking overhead at the cost of wiring
268 the minimum in-core memory used to cache a directory is the physical page
270 We recommend turning this option off in memory-constrained environments;
271 however, when on, it will substantially improve the performance of services
274 Turning on this option will generally not reduce performance even with the
284 cache with dirty buffers when it would not benefit I/O performance.
292 disk controllers system-wide at any given time.
294 The default is self-tuned and
300 Note that setting too high a value
302 bad clustering performance.
303 Do not set this value arbitrarily high!
309 sysctl governs VFS read-ahead and is expressed as the number of blocks
310 to pre-read if the heuristics algorithm decides that the reads are
319 read-ahead adversely affects performance or where system memory is really
340 There are various other buffer-cache and VM page cache related sysctls.
361 high bandwidth over a fewer number of connections, especially if you have
369 can be used to introduce route-specific send and receive buffer size
388 in a marginal performance improvement unless both hosts support the window
393 to a value larger than 65536 in order to obtain good performance from
395 high-latency satellite links.
438 slightly delay the teardown of a connection, or slightly delay the ramp-up
439 of a slow-start TCP connection.
442 turning off delayed acks may be referring to the slow-start issue.
449 high range, selectable via the
475 may block large ranges of ports (usually low-numbered ports) and expect systems
501 thousand if you are running databases or large descriptor-heavy daemons.
502 The read-only
518 the system, and may be determined at run-time by inspecting the value of the
519 read-only
563 should you specify an arbitrarily high value for this parameter, it could
564 lead to a boot-time crash.
591 a large-scale system.
607 The defaults are fairly high and
626 will be able to better use higher-end CPU features for MMU, task switching,
628 Additionally, higher-end CPUs support
641 system is saturating the disk you typically see high CPU idle times and