Text Editor Performance Tests

Discuss anything related to portable freeware here.
Message
Author
TP109
Posts: 571
Joined: Sat Apr 08, 2006 7:12 pm
Location: Midwestern US

Re: Text Editor Performance Tests

#46 Post by TP109 »

Here are the results for NFOPad and J.A.N.E.:

NFOPad
F:\Utilities\NFOPad\NFOPad.exe
1.68
Empty - .26s
89MB - Unable to load file, locks up, freezes

JANE
F:\Utilities\jane\jane.exe
1.69
Empty - .22s
89MB - Locked up, froze, unable to load file

Again, I should emphasize that the purpose of these tests is not to identify the "best" editor. Just because an editor is the quickest to start or can load the largest file doesn't make it the best. These tests just measure a few things and doesn't take into account other features. However, If you are working with large files such as log files, you probably don't want to use certain editors for that. If you want something snappy for viewing small text files, there are several with launch times in the .25sec range that can do that easily. Also, don't take the numbers too seriously, you are not going to notice a difference in speed between .17secs and .18secs. Not accusing anyone of doing that, but just want to make that point.

User avatar
Midas
Posts: 6710
Joined: Mon Dec 07, 2009 7:09 am
Location: Sol3

Re: Text Editor Performance Tests

#47 Post by Midas »

:shock:
lintalist wrote:@midas: fyi UltraEdit now also comes in 64-bit just in case you want to try it out.
  • Thanks! I really hadn't noticed that, will have to test it ASAP... :oops:

TP109
Posts: 571
Joined: Sat Apr 08, 2006 7:12 pm
Location: Midwestern US

Re: Text Editor Performance Tests

#48 Post by TP109 »

Saw this discussed http://www.portablefreeware.com/forums/ ... hp?p=76894 and it looked interesting, so I did a quick test. One of those obscure editors that I didn't know about. Here are the results:

nPad2
F:\Utilities\nPad2\nPad2.exe
3.1.3.36
Empty - .18s
89MB - 15s
250MB - out of memory error, doesn't crash after closing error dialog

A fast and stable editor. Looks like it has good features for such a small size. Very quick launch and was able to load the 89MB test file, although 15s is not really a good load time. Probably wouldn't rely on this for anything over 25MB. Didn't crash or freeze like many, although it's not useable during load, which is normal behavior for most editors. Decided to try 250MB since it seemed stable, but out of memory error was consistant at this file size.

Very similar to Win32Pad in performance. Has more features than Win32Pad however since this is an actual code editor. Also could also be used as a notepad replacement in my opinion.

edit: Probably should clarify the term "usable" as used in these latest tests. It means the editor succesfully loaded the test file and remains usable - no freezing, locking up, slow or erratic behavior, etc.
Last edited by TP109 on Fri Nov 13, 2015 2:42 pm, edited 1 time in total.

TP109
Posts: 571
Joined: Sat Apr 08, 2006 7:12 pm
Location: Midwestern US

Re: Text Editor Performance Tests

#49 Post by TP109 »

I liked the snappy performance of nPad2 that I started using it. Looks like I was a little premature with the positive comments. I was on the PC with nPad2 open in the background. I had used it about an hour before to edit a small html file and just left it open and started doing something else. The PC started to overheat with high CPU usage. It was nPad2 at 92% CPU causing it. Surprising considering its small footprint and the first time I've seen this caused by an editor. Although I'm documenting issues found by switching among various editors, this was one of the most serious, so I thought I would go ahead and post it here in case someone else decides to use it because of my previous comments. It also helps to verify a problem when it happens on a completely different system. I will continue to use it to see if this is repeatable.

xor
Posts: 106
Joined: Sat Apr 18, 2015 11:02 pm

Re: Text Editor Performance Tests

#50 Post by xor »

Textadept would be a fine candidate to be tested for performance as well.

TP109
Posts: 571
Joined: Sat Apr 08, 2006 7:12 pm
Location: Midwestern US

Re: Text Editor Performance Tests

#51 Post by TP109 »

I'm looking at TextAdept now. Can't get it to work with test setup at the moment, but I will figure it out and post the results. I will test with the app launcher (yaP) and probably without it for a comparison.

I looked at nPad2 a little more and the version on Softpedia states it was last updated in 2006. One of it's top features is the ability to support files up to 2GB. I wonder what type of files those could be. I'm using plain text test files (windows 1252 encoding).

User avatar
Midas
Posts: 6710
Joined: Mon Dec 07, 2009 7:09 am
Location: Sol3

Re: Text Editor Performance Tests

#52 Post by Midas »

TP109 wrote:I'm looking at TextAdept now. Can't get it to work with test setup at the moment, but I will figure it out and post the results.
  • I think the following quote is relevant here...
    [url]http://foicica.com/textadept/FAQ.html[/url] author wrote:Q: Why can't Textadept handle HUGE files very well?
    A: Textadept is an editor for programmers. It is unlikely a programmer would be editing a gigantic log file. There are other tools for that case.

TP109
Posts: 571
Joined: Sat Apr 08, 2006 7:12 pm
Location: Midwestern US

Re: Text Editor Performance Tests

#53 Post by TP109 »

So they are stating that up front and it sounds so logical. Very clever. Yet other editors brag about their capability to handle large files. I suppose it's just a matter of who one wants to agree with. I wonder if if a statement like that would still be made it were able to handle large files. I doubt it. That capability would be showcased. So the statement is a pre-emptive strike of sorts.

I'm not agreeing one way or another, just comparing editors to one another in a standardized and objective manner. Like I said previously, choosing an editor is a complex and personalized decision. There are many factors involved and file handling is just one of them. If something is to be tested, it's impossible to test every characteristic at once. Specifc characteristics that can be measured and common to the group are chosen so that comparisons can be made. Comprehensive testing, a rare event for complex systems, has to be performed by separate groups testing different characteristics. It's just too complex to do so otherwise.

Nobody likes flaws exposed, even minor ones. Testing is a very touchy subject that instantly raises defenses. I understand that. So it is natural to attack the methodology, the justification for testing, and the credibility of the institutions and personnel involved.

I'm getting off topic here. Anyway, I looked at TextAdept and tried to find a way to integrate it into the test system, but failed. TextAdept is not a native Windows app and doesn't have the handlers that respond to commands through the Windows API such as WM_CLOSE, WM_SYSYCOMMAND, ALT-F4 and others used by AppTimer to control and measure varioius aspects of the application under test. Because of that TextAdept doesn't respond to any of those commands or provide any information to the system so it's impossible to determine its launch speed using the current setup. It can still be tested for large file handling and I will do that and post the results later.

Although TextAdept doesn't fall into the same category of editors that have been tested so far, I was curious to how it might compare to the others.

Anyway, the information stated was useful. If nothing had been stated, it could be found out; if claims were made, they could be verified as true or false.
Last edited by TP109 on Thu Nov 12, 2015 5:45 pm, edited 1 time in total.

TP109
Posts: 571
Joined: Sat Apr 08, 2006 7:12 pm
Location: Midwestern US

Re: Text Editor Performance Tests

#54 Post by TP109 »

The previous TextAdept data below is invalid because TextAdept was setup incorrectly. Had a lot of trouble getting TextAdept to work right with the yaP launcher. I instead used the executable in the App folder to bypass the yaP launcher for this test. Also, there was a space in the path that I didn't spot previously and that was removed. Changes to the configuration files were also made and I finally I got it to work.

Below are the results:

TextAdept (bypassed yaP Launcher and used App folder executable)
F:\Utilities\TextAdeptPortable_yaP\App\textadept.exe
8.4
Empty - .35s
89MB - 4s
250MB - Unable to load after waiting 300s. Used over 750MB RAM, GUI remained usable and stable.
Note: Linux Native (actual empty launch time may be inaccurate)
Used about 750MB RAM while loading 89MB test file, down to 7MB once minimized

Note the change in that it was able to load the 89MB test file with no problem.

----------------------------------------------------------------------------------------------------------
Previous Post:
Quickly performed large file testing on TextAdept

TextAdept
F:\Utilities\TextAdept Portable_yaP\TextAdeptPortable.exe
8.4
Empty - Not Tested. Unable to intergrate into test system.
89MB - 29s
250MB - Unable to load after waiting 300s. Using over 750MB RAM, but GUI remained usable and stable.

This editor was the most difficult to try to setup for testing. Nothing worked. Tried every configuration file combination. Others have been difficult, but I could eventually get them to operate properly.
Last edited by TP109 on Fri Nov 13, 2015 10:36 pm, edited 4 times in total.

User avatar
Midas
Posts: 6710
Joined: Mon Dec 07, 2009 7:09 am
Location: Sol3

Re: Text Editor Performance Tests

#55 Post by Midas »

Might you be interested in another local oddity? Have a look at Minimum Profit (http://www.portablefreeware.com/forums/ ... hp?t=21578)... :mrgreen:

TP109
Posts: 571
Joined: Sat Apr 08, 2006 7:12 pm
Location: Midwestern US

Re: Text Editor Performance Tests

#56 Post by TP109 »

The launch speed on Minimum Profit appears to be incredibly fast. Very noticable. Might try testing this, but like TextAdept, since it's a Linux native, it probably won't work. Otherwise, I like it anyway.

TP109
Posts: 571
Joined: Sat Apr 08, 2006 7:12 pm
Location: Midwestern US

Re: Text Editor Performance Tests

#57 Post by TP109 »

I tested the Minimum Profit editor in both the GUI and CLI versions. By tinkering around I got a configuration setting to work. With that success, I went back to TextAdept and retested that too, although I never was able to get it to work with the yaP launcher. I edited my previous post for TextAdept with the new info. Here is the Minimum Profit data:

Minimum Profit (GUI)
F:\Utilities\MinimumProfit\mp-5.exe
5.2.10
Empty - .29s
89MB - 57s
Note: Linux Native (actual empty launch time may be inaccurate)
Used about 200MB RAM while loading 89MB test file, down to 4MB once minimized. Interface not usuable while loading.

Minimum Profit (CLI)
F:\Utilities\MinimumProfit\mp-5c.exe
5.2.10-dev
Empty - .28s
89MB - 42s
Note: Linux Native (actual empty launch time may be inaccurate)
Used about 300MB RAM while loading 89MB test file, down to 3MB once minimized. Interface not usuable while loading.

This is a fast launching editor and I'm not sure that the empty launch times are accurate, especially considering these are non-native applications. But they are at least that fast and could be faster. At least my eyes tell me these are fast. AppTimer is supposed to measure when the application is fully initialized, so it's possible they are rendering quickly and not fully initialized yet. Or maybe these apps are not fully compatible with Windows and are unable to provide the correct information to the system. I don't know for sure.

That said, I can't certify if any of the launch times for any of the tests are completely accurate since I don't have a way to calibrate it against a standard of some type. In any case, all the testing is being performed with the same measurement applications and hardware, so it's accurate for comparing the results to each other.

User avatar
Userfriendly
Posts: 430
Joined: Tue Nov 27, 2012 11:41 pm

Re: Text Editor Performance Tests

#58 Post by Userfriendly »

New version of EmEditor 15.6 improves on performance. "Improved the speed of opening files. Compared to v15.5, the speed to open a large file (809 MB – 2.02 GB) became 12% – 23% faster."

I did a test of of copy and pasting 16 million lines of random text (about 2 gigs worth) and it survived. Did it pretty fast too. No freezing or hanging.

I like that it continues to improve on its main feature.

TP109
Posts: 571
Joined: Sat Apr 08, 2006 7:12 pm
Location: Midwestern US

Re: Text Editor Performance Tests

#59 Post by TP109 »

Thanks for posting the info on EmEditor. I looked at its features and it claims it can open files up to 248GB easily. See feature list here: https://www.emeditor.com/#features. I wonder how long it would take to load a file that big. What were the times for the tests you performed?

I didn't test EmEditor, but I did test UltraEdit. Being disk based, UltraEdit initializes slowly but after that, it is fast and can easily load huge files, and it's usable while loading. I expect EmEditor is probably similar. For small files, the freeware editors were much faster to launch and initialize.

I'm going to make a comprehensive listing of the test results soon. I need to retest the lastest editors (notepad class) using the same test file size I used for the 1st round of testing. 68MB instesad of 89MB so a fair comparison can be made. I won't go through a bunch of analysis, but will list fastest to slowest, lowest to most memory use, etc for all the editors. I think I have over 30 editors tested so far. If there are any more freeware, portable editors that haven't been included, please post them on this thread.

User avatar
Userfriendly
Posts: 430
Joined: Tue Nov 27, 2012 11:41 pm

Re: Text Editor Performance Tests

#60 Post by Userfriendly »

Well EmEditor is similar to UltraEdit as it uses the temp directory to load large files but EmEditor has a file size threshold limit option you can customize to tell the program when it should use the disk cache. By default EmEditor uses the disk cache for files 300mb and above. Anything below that will just use RAM. You can also disable the disk caching completely and it will behave like any other text editor and use only RAM. Though unlike other editors, huge files are still instantly usable while loading in ram.

With my setup, I'm using a 16gb ramdisk for temp files. I'm also launching the files from the ramdisk to eliminate the disk speed bottleneck. Used a 2048 txt file with 16 million lines of random text.

w/ Disk Cache Temp File
2048MB File - 6s load time - 350MB memory used - 2.21GB temp file.

w/o Disk Cache Temp File
2048MB File - 8s load time - 5.6GB memory used

It essentially performs a little slower for me with temp file disabled but maybe it would be faster for others who use regular hard drive cache.

I tried EditPad Pro again on this same file and it opened instantly but the downside is it doesn't cache the entire file immediately. I can only access the top first part of the file. Only way I can access the 16th million line of the file is to constantly keep scrolling down until it loads everything into ram and at the pace it was going it seemed like it would take about 5 minutes to do so. On the other hand, EmEditor takes 6 seconds for me to have access the entirety of the file.

My tests will by very different from yours tp109 since I have a high-end PC. 200mb files load in 1 second in Notepad++ for example. So I can only test legit huge files to notice a difference.

Post Reply