Photo
shows Diskeeper Lite running on @Macarlo's NT
Using Diskeeper
|
Using Diskeeper |
I, @Macarlo, StarOffice registered user for all platforms, am just
performing an exaustive performance test with SO 5.0 PE on Windows NT
4.0. Because I need copmplete conectivity and interactivity with OS/2
Warp and Linux Red Hat - just the Operating Systems I have and use
everyday on the same machine for Internet works - I formated NT
partitions in FAT and not i n NTFS, for mount it as <-t vfat> on the
Unix mount point. The intensive activity I'm performing here it's
possible with not troubleshotings because I use Diskeeper Lite,
downloaded free from Executive Software.While I work on two FAT NT
partitions with StarOffice, both are permanently defragmented with
Diskeeper. I'm Diskeeper user since 1997.
Screenshots:
Special screenshot: Working on the Net with Windows NT>###
White Paper:
Defragmentation and Windows NT Performance Benefits
Introduction
A series of tests was conducted to determine, in a laboratory environment, the potential for performance improvement with defragmentation of the Windows NT File System (NTFS). This is the results of those tests.
Definitions of Terms
fragmentation—"the state of being fragmented." The word fragment means "a detached, isolated or incomplete part." It is derived from the Latin fragmentum, which in turn is derived from frangere, meaning "break." So fragmentation means that something is broken into parts that are detached, isolated or incomplete.
file fragmentation—computer disk files that are not whole but rather are broken into scattered parts.
free space fragmentation—the empty space on a disk is broken into scattered parts rather than being collected all in one big empty space.
Note: For the purpose of these tests, the term disk fragmentation includes both file fragmentation and fragmentation of the free space on the disk.
Background on the NTFS File System
The Windows NT File System (NTFS) was created specifically for use with the Windows NT Operating System. NTFS is a recoverable, secure, reliable and efficient file system which supports client-server systems.
NTFS keeps track of the contents of a volume in a file called the Master File Table (MFT). The MFT consists of an array of file records, with each file usually represented in the MFT by one file record. If, however, a file has a large number of file attributes, such as security information, or becomes highly fragmented, more than one file record may be needed. As more files and file records are added to a volume, the MFT may need to expand to accommodate more file records. The ability of the MFT to expand is part of the way NTFS supports large disks with large numbers of files. As the MFT expands, it may also fragment across the disk. It should be noted that once the MFT expands it will not shrink, but as files are deleted from the volume, unused file records become available for use by NTFS as new files are subsequently added.
Test Configurations
These tests were performed on two separate computer systems, representative of a midrange system and a higher-end system.
)Dual Pentium 90MHz running Windows NT Workstation Version 4.0, an Adaptec 2940 PCI adapter, and a Seagate Hawk SCSI-2, 4.0 Gigabyte drive formatted into six, 500 Megabyte NTFS partitions (the remaining 1 gigabyte of space on the disk was not used).
)Pentium Pro 200MHz running Windows NT Workstation Version 4.0, a BusLogic Flashpoint PCI adapter, and the same Seagate Hawk SCSI-2, 4.0 Gigabyte drive used with the Dual Pentium 90MHz machine, again formatted into six, 500 Megabyte NTFS partitions.
Procedures Used for Testing
In each test, the volumes used in testing were initially formatted NTFS. They were not converted from FAT to NTFS. Half the partitions were formatted under Windows NT 3.51 and half were formatted under Windows NT 4.0. All the tests were run under Windows NT 4.0.
A standardized procedure was used to produce a specific level of file fragmentation on each volume tested. The steps were performed as a batch procedure to create uniform, reproducible levels of fragmentation. The same procedure was followed on the partitions formatted under Windows NT 3.51 as was followed on the partitions formatted under Windows NT 4.0. It is interesting to note that the partitions created under Windows NT 3.51 displayed a substantially higher level of fragmentation than those created under Windows NT 4.0. The following procedure was used:
The entire volume was filled with files of uniform length, 2000 files per directory. Every other file was deleted, thus uniformly fragmenting the free space on the volume.
Having the free space uniformly fragmented at a known level is desirable for the write tests so that each time a file is written to the volume it is guaranteed the file will be written at a known level of fragmentation.
The uniform-length files vary in size depending on the desired level of average fragments per file: 4096 byte files for 6.2 average fragments per file (11.11 on Windows NT 3.51), 8192 byte files for 3.51 average fragments per file (6.41 on Windows NT 3.51), and 32768 byte files for 1.61 average fragments per file (2.20 on Windows NT 3.51).
Next, the Windows NT Workstation 4.0 CD-ROM was copied to the volume. Multiple directories were created on the volume and the CD-ROM was copied into each directory until the volume was full. This group of files was chosen for its variety of file types and sizes, and to make it possible for others to reproduce these tests on any Windows NT system.
The rest of the original uniform-length files were then deleted. At this point both the Windows NT files and the free space were uniformly fragmented.
These steps were performed to greater or lesser extents to create the varied degrees of fragmentation shown in the following test results.
Uniform fragmentation of free space was found to be very important to achieving reproducible results. If the free space was fragmented in a random fashion, characterizing file writes was inconsistent. This was because the size of the free spaces varied and random chance meant that a file system might find a contiguous free space in which to write a file. Once free space was fragmented in a consistent, known way, reproducible results were possible. The sizes of free spaces used for each volume are listed in the charts which follow.
The levels of fragmentation chosen for the tests were: no file fragmentation, a minimally fragmented volume, a moderately fragmented volume, and a heavily fragmented volume.
Two scenarios were used to characterize volumes with no file fragmentation. One scenario was that of a disk which had never had any file fragmentation and the other a volume which had been fragmented and subsequently defragmented. First a volume was fragmented to a high level of fragmentation¾6.20 average fragments per file on the Windows NT 4.0-formatted volume, and 11.11 average fragments per file on the Windows NT 3.51-formatted volume¾and defragmented by Diskeeper® Lite. Since the volume had been completely full and the Master File Table had expanded, this gave a worst-case condition of fully defragmented files but with a heavily fragmented Master File Table. This volume is listed in the following charts as average fragments per file of 1.0. The files from that volume were then copied to a freshly formatted volume. This results in a volume which has never been fragmented and has a contiguous Master File Table. This volume is listed as average fragments per file of Control.
The fragmentation analysis function of Diskeeper Lite was used to analyze and determine the level of fragmentation on each volume. Average Fragments per File is an index of the level of file fragmentation on the disk. If the average fragments per file is 1.00, the files are contiguous. If the figure is 1.10, then 10% of the files, on average, are in two pieces. 1.20 means 20%, 1.30 means 30%, etc. A figure of 2.00 means the files average two fragments each. 1.00 is the best figure attainable, indicating that all files (or nearly all files) are contiguous.
Once each volume was set up and the level of fragmentation was verified, the benchmark test program, NTBENCH.EXE, was run.
NTBENCH.EXE is a custom program written explicitly for the purpose of characterizing disk data transfer performance. To measure individual file read performance, a specified file is opened, timing is started, the entire file is read from start to end repetitively, timing is ended, and the file is closed. Determining individual file write performance is done in a similar manner: the specified file is opened, timing is started, the entire file is written from start to end repetitively, timing is ended, and the file is closed.
For entire disk read performance, timing is started, the volume is searched for a file, the file is opened, the entire file is read, the file is closed, the process is repeated until all files have been read, and timing is ended. To measure entire disk write performance, timing is started, the volume is searched for a file, the file is opened, a temporary file is created in the same directory, the entire file is copied to the temporary file, both files are closed and timing is ended. The temporary file is deleted before timing is restarted. This process is repeated until all files have been read and written, then cumulative timing is ended. Cached reads and writes were not performed to ensure the changes in performance measured by the benchmark program were the result of fragmentation.
The same procedure was repeated for each of the system configurations.
The performance tests were:
· Measurement of the degree of fragmentation of files on the volume, using the fragmentation analysis function of Diskeeper Lite.
· One pass of locating and reading every file on the volume using NTBENCH.EXE.
· One pass of locating and making a temporary copy of every file on the volume using NTBENCH.EXE (the performance analysis does not include the deletion of the temporary file).
This series of tests was run on the partitions formatted under Windows NT 3.51 as well as those formatted under Windows NT 4.0. Appendix A shows the test results on the Dual 90 MHz Pentium system, and Appendix B shows the results on the Pentium Pro 200 MHz system.
Conclusions
As can be seen in all the tests, defragmentation of files and free space significantly restores system performance. In fact, defragmenting the volume almost completely restores disk performance to pre-fragmentation levels.
Comparison of the defragmented volume to the control volume also points up the importance of regular defragmentation. Note that, in the testing above, the time to read and write files did not fully return to pre-fragmentation levels. This is due to Master File Table fragmentation. As stated earlier, the MFT, once it expands to accommodate additional file records for heavily fragmented files and other purposes, does not shrink. In that a fragmented Master File Table requires additional reads from the volume, it can be seen that regular defragmentation of a volume from the time it is initialized is very important, and will help ensure the MFT does not grow disproportionately.
Overall, these tests indicate that fragmentation is a major factor in the performance of a Windows NT system. Defragmentation of an NTFS volume, especially one which has been in use for some length of time, can be expected to substantially improve the overall responsiveness of the system.
Appendix A
Dual Pentium 90 MHz — Windows NT 4.0-Formatted Volumes
Number of Minutes to Read Every File on the Disk
Average Fragments per File Control 1.00 1.61 3.51 6.20
Disk Read (minutes) 4.04 4.65 6.85 7.01 8.06
% of ideal 100.00 115.10 169.55 173.51 199.50
Dual Pentium 90 MHz — Windows NT 3.51-Formatted Volumes
Number of Minutes to Read Every File on the Disk
Average Fragments per File Control 1.00 2.20 6.14 11.11
Disk Read (minutes) 4.25 4.68 7.65 6.97 6.42
% of ideal 100.00 110.12 180.00 164.00 151.06
Dual Pentium 90 MHz — Windows NT 4.0-Formatted Volumes
Number of Minutes to Read then Write Every File on the Disk
Average Fragments per File Control 1.00 1.61 3.51 6.20
Disk Read / Write (minutes) 13.71 17.05 20.27 20.46 22.13
% of ideal 100.00 124.36 147.85 149.23 161.42
Dual Pentium 90 MHz — Windows NT 3.51-Formatted Volumes
Number of Minutes to Read then Write Every File on the Disk
Average Fragments per File Control 1.00 2.20 6.14 11.11
Disk Read / Write (minutes) 14.57 16.02 22.66 23.80 27.16
% of ideal 100.00 109.95 155.53 163.35 186.41
Appendix B
Pentium Pro 200 MHz — Windows NT 4.0-Formatted Volumes
Number of Minutes to Read Every File on the Disk
Average Fragments per File Control 1.00 1.61 3.51 6.20
Disk Read (minutes) 2.07 2.41 4.13 4.81 5.79
% of ideal 100.00 116.43 199.52 232.37 279.71
Pentium Pro 200 MHz — Windows NT 3.51-Formatted Volumes
Number of Minutes to Read Every File on the Disk
Average Fragments per File Control 1.00 2.20 6.14 11.11
Disk Read (minutes) 2.38 2.31 4.87 4.88 4.66
% of ideal 100.00 97.06 204.62 205.04 195.80
Pentium Pro 200 MHz — Windows NT 4.0-Formatted Volumes
Number of Minutes to Read then Write Every File on the Disk
Average Fragments per File Control 1.00 1.61 3.51 6.20
Disk Read / Write (minutes) 10.43 12.45 15.49 16.41 16.96
% of ideal 100.00 119.37 148.51 157.33 162.61
Pentium Pro 200 MHz — Windows NT 3.51-Formatted Volumes
Number of Minutes to Read then Write Every File on the Disk
Average Fragments per File Control 1.00 2.20 6.14 11.11
Disk Read / Write (minutes) 11.25 12.71 17.76 18.87 20.49
% of ideal 100.00 112.98 157.87 167.73 182.13
© Copyright 1997 Executive Software International, Inc. All rights reserved
Diskeeper is a registered trademark owned by Executive Software International, Inc.
Windows NT is a registered trademark owned by Microsoft Corporation.
All other trademarks are the property of their respective owners.
CONTACT AUTHOR
701 North Brand Blvd.
Glendale, CA 91203
@Macarlo, Inc.
@Macarlo's Shareware & Web
OS/2
Java Lobby
Member
Java Site Accredited