patrickd@chinet.chi.il.us (Patrick Deupree) (05/29/90)
In article <10461@spool.cs.wisc.edu> thaler@shorty.cs.wisc.edu (Maurice Thaler) writes: >I believe that many people are ASSUMING that the performance in windows >3.0 is better without any really testing. > >In this week's InforWorld and PC-Week, there were conficting reports as >to the performance of Windows 3.0. In PC Week, they reported 10% >increase in speed in loading and a 50% increase in speed in execution on >a couple applications. In Infoworld, they reported that things slowed >down about 5-10% on everything. Due to the new virtual memory scheme implemented in Windows 3.0 there are a lot of factors that could account for speed differences (albeit, nott as major as you've noted). Since it saves "memory" on the disk drive, it could depend on the size of the smartdrive cache (if any) or the speed of the hard disk. It would depend on how much memory is in the system (e.g. 2 meg versus 3 meg can make quite a difference). And so on and so on. -- "Organized fandom is composed of a bunch of nitpickers with a thing for trivial pursuit." -Harlan Ellison Patrick Deupree -> patrickd@chinet.chi.il.us
mcdonald@aries.scs.uiuc.edu (Doug McDonald) (05/29/90)
In article <1990May28.232831.8843@chinet.chi.il.us> patrickd@chinet.chi.il.us (Patrick Deupree) writes: >>a couple applications. In Infoworld, they reported that things slowed >>down about 5-10% on everything. > > It would depend on how much memory is in the system (e.g. 2 meg >versus 3 meg can make quite a difference). And so on and so on. > It would depend on memory ONLY if you were previously running out. I never have. On the other hand, I have now made a lot of tests, and two things are clear: running in 286 mode is slower than 86 mode. Not just in Windows, but always. MAybe 5 to 10% slower. Second, most people run Windows 2.x in color mode with ***8*** color drivers. The Win 3.0 standard drivers are ***16*** color. This means that things like BitBlts and icons are slower by some amount, maybe 4/3, as there is often 4/3 as much work to do. This seems to account for much of the slowdown. I tried comparing 2.11 and 3.0 in mono mode and they got much closer in speed. Also, to the person who posted the "4" at 0x437 for getting rid of the "old version" message: THANKS. Doug McDonald
patrickd@chinet.chi.il.us (Patrick Deupree) (05/29/90)
In article <1990May29.003825.20599@ux1.cso.uiuc.edu> mcdonald@aries.scs.uiuc.edu (Doug McDonald) writes: (in response to my stating that Windows speed has to do with system memory). >It would depend on memory ONLY if you were previously running out. >I never have. With the way Windows swaps memory to disk (or disk cache) speed does have to do with memory. If I have enough in my system so that the system has to swap to disk, it will slow things down. If I have enough system memory so that most or all of my stuff can remain resident, it's quicker. -- "Organized fandom is composed of a bunch of nitpickers with a thing for trivial pursuit." -Harlan Ellison Patrick Deupree -> patrickd@chinet.chi.il.us
jeffmu@microsoft.UUCP (Jeff MUZZY) (05/29/90)
In Patrick's peformance issues... I would also have to include Hard disk fragmentation as a issue. Some people don't believe that you can get from a 5-10% performance increase from tuning a system by running smartdrv and a hard disk unfragmentation/compression program every week (Based on my compile and automated tests results.) . This will help in compiling programs too.. -- jeffmu@microsoft or uunet!microsoft!jeffmu MaBellNet: (206) 882-8080 <Insert your favorite disclaimer about opinions and companies here>
patrickd@chinet.chi.il.us (Patrick Deupree) (05/31/90)
In article <54910@microsoft.UUCP> jeffmu@microsoft.UUCP (Jeff MUZZY) writes: >I would also have to include Hard disk fragmentation as a issue. >Some people don't believe that you can get from a 5-10% performance >increase from tuning a system by running smartdrv and a hard disk >unfragmentation/compression program every week (Based on my compile >and automated tests results.) Hmm. Is it cool to run Nortons speed disk now. The last I remembered, Windows creates a swap file on the disk and running speed disk would screw up the organization of that file. Then again, that was quite a few months ago so it may have changed in the final release. -- "Organized fandom is composed of a bunch of nitpickers with a thing for trivial pursuit." -Harlan Ellison Patrick Deupree -> patrickd@chinet.chi.il.us
altman@sbstaff2.cs.sunysb.edu (Jeff Altman) (06/01/90)
In article <1990May31.145141.20600@chinet.chi.il.us> patrickd@chinet.chi.il.us (Patrick Deupree) writes: >In article <54910@microsoft.UUCP> jeffmu@microsoft.UUCP (Jeff MUZZY) writes: >>I would also have to include Hard disk fragmentation as a issue. >>Some people don't believe that you can get from a 5-10% performance >>increase from tuning a system by running smartdrv and a hard disk >>unfragmentation/compression program every week (Based on my compile >>and automated tests results.) > >Hmm. Is it cool to run Nortons speed disk now. The last I remembered, >Windows creates a swap file on the disk and running speed disk would screw >up the organization of that file. Then again, that was quite a few months >ago so it may have changed in the final release. First of all, never ever run a Speed Disk type utility from under Windows. This will surely screw things up. Second, the question of whether running a Speed Disk type utility is safe only comes up when running 386 Enhanced mode and using the Permanent Swap File. Before creating the permanent swap file, be sure to run a Speed Disk type utility to compact the drive usage. This enables you to have larger Swap Files as the whole point of a Permanent Swap file is to make it contiguous. This way Windows 3.0 performance will not be affected by the fragmentation on the rest of the drive. Third, if after creating a Permanent Swap File you wish to run a Speed Disk type utility, make sure that the utility you use allows you to mark files (hidden files) so that they will not be moved from their current locations. Norton Utilities Advanced 4.5 Speed Disk does have this option. - Jeff (jaltman@ccmail.sunysb.edu)
carlw@ka.excelan.com (Carl Wohlforth) (06/01/90)
Speed disk organized files into contiguous blocks. Windows 3.0 requires a contigous disk space for its swap file. For those of us with plenty of disk space performance can be improved by running speed disk to create a large empty area on the disk. Then follow MS instructions on creating a permanent swap area. I always reboot after running speed disk. In article <1990May31.145141.20600@chinet.chi.il.us>, patrickd@chinet.chi.il.us (Patrick Deupree) writes: > In article <54910@microsoft.UUCP> jeffmu@microsoft.UUCP (Jeff MUZZY) writes: > >I would also have to include Hard disk fragmentation as a issue. > >Some people don't believe that you can get from a 5-10% performance > >increase from tuning a system by running smartdrv and a hard disk > >... > > Hmm. Is it cool to run Nortons speed disk now. The last I remembered, > ... > > -- > "Organized fandom is composed of a bunch of nitpickers with a thing for > trivial pursuit." -Harlan Ellison > > Patrick Deupree -> patrickd@chinet.chi.il.us --- Carl Wohlforth, 2180 Fortune Dr., San Jose Ca., 95131 carlw@NOVELL.COM 408-473-8230 {ames, apple, leadsv, mtxinu, 3comvax}!excelan!carlw
rspangle@jarthur.Claremont.EDU (Randy Spangler) (06/01/90)
Before installingh Win 3.0, I defragmentized (via Norton SD 4.5) and backed up my hard drive. Then I stuck a 3MB swap file on. Since the swap file is a hidden system file as created, Norton SD DOES NOT attempt to move it - In other words, yes, it's safe to run SD on a drive with a swap file. Just check the options menu under unmovable files, I think, to see if the 386spart.par is there. -- -------------------------------------------------------------------------- | Randy Spangler | The less things change, the | | rspangle@jarthur.claremont.edu | more they remain the same | --------------------------------------------------------------------------
marshall@wind55.seri.gov (Marshall L. Buhl) (06/02/90)
rspangle@jarthur.Claremont.EDU (Randy Spangler) writes: >Before installingh Win 3.0, I defragmentized (via Norton SD 4.5) and backed >up my hard drive. Then I stuck a 3MB swap file on. >Since the swap file is a hidden system file as created, Norton SD >DOES NOT attempt to move it - In other words, yes, it's safe to run >SD on a drive with a swap file. Just check the options menu under >unmovable files, I think, to see if the 386spart.par is there. Speaking of disk optimization, where would the most efficient place to put this swap file? I have two drives, a 40 MB that will be used mostly for win and win apps, and a 60 MB. My hunch is that it should be put near the FAT on the 60 MB drive. I use FastTrax, so I can put the files in any specified physical location on the disk. With the swap file on a different disk than my apps, I'll be using two set of heads. BTW, my user files will be on the 60 MB drive. Putting it near the FAT should reduce head motion. This is for my home system. For my work machine, I have a single 322 MB drive. Would it be better to put Win and win apps nearest the FAT, then the swap file? Or maybe, swap file then win and win apps? Which will be used most often? The programs or the swap file? Just wandering (sic), Marshall -- Marshall L. Buhl, Jr. EMAIL: marshall@seri.gov Senior Computer Engineer VOICE: (303)231-1014 Wind Research Branch 1617 Cole Blvd., Golden, CO 80401-3393 Solar Energy Research Institute Solar - safe energy for a healthy future