This section provides an overview of changes that the SGI SynaptIQ software makes in your base Linux distribution.

The following packages were added by SGI:

SGI SynaptIQ supports the XFS file system running on Linux platforms. XFS is a scalable, high-performance, journalling file system that SGI has contributed to the Open Source Linux community. The XFS journalling technology lets the file system restart almost immediately after an interrupt, regardless of the number of files it is managing.

XFS is a full 64-bit file system, capable of handling files as large as a million terabytes and of delivering near raw-I/O performance. For information about XFS, see the following:

http://oss.sgi.com/projects/xfs

Kernel-supported asynchronous I/O (KAIO) differs from asynchronous I/O in that it is implemented with support from kernel modifications. These kernel modifications allow KAIO to perform split-phase I/O to maximize concurrency of I/O at a device. Split-phase I/O allows the initiating request (such as an aio_read) to truly queue the I/O at the device as the first phase of the I/O request. The second phase of the I/O request, performed as part of the I/O completion, propagates results of the request. The results may include the contents of the I/O buffer on a read, the number of bytes read or written, and any error status.

For more information about KAIO, see the following URL:

http://oss.sgi.com/projects/kaio

CSA is a set of C programs and shell scripts that provide methods for collecting per-process resource usage data, monitoring disk usage, and charging fees to specific login accounts. CSA takes this per-process accounting information and combines it by job identifier (jid) within system boot uptime periods.

CSA provides the following features not available with other Linux accounting packages:

For detailed information about CSA, see the following URL:

http://oss.sgi.com/projects/csa

SGI SynaptIQ supports the runon command, which runs a command on a particular CPU (this CPU affinity is called process pinning). The runon command can assign process pinning using fork(), or by process ID. Pinning is inherited across fork() and exec() system calls. To change the pinning, you can use the prctl(2) call. The runon(1) man page contains more information.

SGI SynaptIQ supports the hinv command, which displays the contents of the system hardware inventory table. This table is created each time a system is booted and contains entries describing various pieces of hardware in the system. The items in the table include main memory size, cache sizes, floating point unit, and disk drives. Without arguments, the hinv command displays a one line description of each entry in the table. The hinv(1) man page contains more information.

SGI SynaptIQ provides updated drivers, as described in this section.

Advanced Linux Sound Architecture (ALSA) drivers provide advanced multimedia. The asynchronous I/O (AIO) facility implements interfaces defined by capabilities. For more information, see “Advanced Linux Sound Architecture (ALSA) Drivers”.

The QLogic 1080/1280 and 2100 drivers have been supplied by QLogic Corporation, and include updated firmware and improved error handling.

The QLogic drivers were updated to the following new versions:

The Alteon Gigabit Ethernet driver has been modified to recognize and drive the SGI Gigabit Ethernet card. SGI SynaptIQ supports version 0.45.

Other drivers were updated to new versions as follows:

Performer for Linux is fully API-compatible with existing Performer applications running on IRIX and is a full distribution, including the core run-time libraries and file loaders, development header files, sample source code, and man pages.

Process Aggregates (PAGG) are modifications to the Linux kernel that let developers create loadable kernel modules to group processes into aggregates. PAGG provides functions that allow loadable kernel modules to register as providers of a specific type of process aggregate implementation. Additionally, this feature ensures that membership in a process aggregate, or group, is inherited across process forks.

PAGG can be used to support a formal job container on Linux. To learn more about Process Aggregates and the job container module for Linux, consult the following URL:

http://oss.sgi.com/projects/pagg
 

Comprehensive System Accounting provides enhanced job accounting, which will make use of the job container kernel module. For more information, see “Comprehensive System Accounting (CSA)”, and the following URL:

http://oss.sgi.com/projects/csa

SGI SynaptIQ contains support for the Advanced Linux Sound Architecture (ALSA) drivers for users who want to use OpenML or who use SGI Graphics Clusters for intensive audio purposes.

ALSA's mixer model provides fine control over muting and volumes of audio channels. Unlike the standard Linux audio drivers (Open Sound System or OSS), ALSA drivers offer a mute separate from volume and distinguish between input gains for recording and output feedback volumes for input sources.

The main advantages of using the ALSA drivers are as follows:

Performance Co-Pilot (PCP) provides a framework and set of services to support system-level performance monitoring and performance management. The PCP open source release provides a unifying abstraction for all of the interesting performance data in a system, and allows client applications to easily retrieve and process any subset of that data using a single API.

A client-server architecture allows multiple clients to monitor the same host and a single client to monitor multiple hosts (for example, in a graphics cluster). This enables centralized monitoring of distributed processing.

Integrated archive logging and replay allow client applications to use the same API to process real-time data from a host or historical data from an archive.

The framework supports APIs and configuration file formats that enable the scope of performance monitoring to be extended at all levels.

The open source release of PCP provides a subset of the features of SGI's Performance Co-Pilot products for IRIX (see http://www.sgi.com/software/co-pilot/ ).

The following is a brief list of PCP features supported in SGI SynaptIQ:

SGI SynaptIQ includes support for NFS version 3, with some bugfixes included to allow compatibility with IRIX systems.

Large amounts of physical memory coupled with the ability to create large (multi-gigabyte) shared memory segments provide a boost in performance to various types of workloads. SGI includes a configuration parameter to configure the Linux community standard implementation—that is, i686 UP and SMP kernels have 4GB memory support (SGI Graphics Cluster Series 12 are limited to 1.5 GB of memory support).

While the UNIX System V IPC semaphore facility does provide exceptional capability, its performance leaves much to be desired. Many UNIX vendors have released a low-overhead interapplication synchronization primitive known as “post /wait.”

SGI has included in this release a kernel level implementation of post/wait along with the library containing application APIs. The post allows for a process to “wait” for an event. This event can either be a timeout or a “post” from another process. A group of cooperating processes can use these “post” and “wait” facilities to synchronize among themselves.

In order to use post/wait, the kernel must be compiled with the CONFIG_PW configuration variable, and you may optionally set an additional configuration variable, CONFIG_PW_VMAX. These variables are described in the configuration help. For a user program to use the post/wait facilities, it must link against libdba.so.

For more information on post/wait, please refer to the postwait(3) man page.

The ability to overlap I/O and processing activities has always been important to high-performance applications. To allow this type of overlap in single-threaded applications, SGI has included a kernel-level implementation of POSIX asynchronous I/O and the associated API library.

SGI SynaptIQ works with raw devices as well as with file systems including pipes and sockets.

This facility is turned on by setting the CONFIG_AIO kernel option. User code can get access to the facility by linking with libdba.so. Further information can be found in the /lib/libdba/README file.

SGI SynaptIQ does not include kernel spinlock metering by default, but you can check the following URL for its availability and for additional information:

http://oss.sgi.com/projects/lockmeter

Spinlock metering allows developers to gather statistical information about the SMP kernel's use of spinlocks and mrlocks (multiple-reader single-writer spinlocks). This functionality is called spinlock metering or lockmetering.

If spinlock metering is available from http://oss.sgi.com/projects/lockmeter , it can be built into the kernel using the CONFIG_LOCKMETER configuration option (in the Kernel Hacking section of make xconfig). A kernel built with lockmetering will exhibit a small (roughly 1%) performance degradation relative to a kernel that is not configured for lockmetering.

Changes have been made to the Linux crash utility, which are explained briefly below. SGI SynaptIQ uses version 3.1.1, which supports lcrash on IDE drives as well as SCSI drives. General information about lcrash can be found in the /cmd/lcrash/README file. Major changes follow:

The following list describes patches that have been implemented as well as enhancements to configuration options, commands, and libraries: