53 Replies Last post: Mar 27, 2012 6:04 PM by Andreas Vilinski  
Jakub Maly Newbie 24 posts since
Apr 24, 2008
Currently Being Moderated

Dec 17, 2011 6:45 PM

ReSharper performance

Yes, this will be another post about poor performance of R#.

 

I've been using R# since version 4. At that time, our work solution comprised of slightly above 100 projects and it was not possible to use R# with the whole solution. But when I unloaded the projects that I did not need, I cut the number to around 30 and R# was more or less usable (but it keeps underlining the code referencing types from the unloaded projects - why, when everything compiles fine? this problem continues up until now). I was enthusiastic about R#, recommended it to my colleagues and wrote a review to support what I regarded as good product.

 

Each new version of R# promised improved performance, but I never noticed a big difference.

 

As time went by, the solution grew even more, now it is about 150 projects. When I unload those projects that I don't need, I got about 50, but R# is not usable even with this limited solution. VS get unresponsive very often and after some time of work, OutOfMemory exception begin to pop up.

 

I am not using R#  any more, even though I love its features. I now regret the time I spent learning to use R# effectively and lost the habit of using the built-in capabilities of VS. They lack nice features of R#, but at least they don't make VS unresponsive each once in a while.

 

I uninstalled R# recently. I will look for alternatives (maybe Telerik product, as is suggested in this post http://devnet.jetbrains.net/thread/432674) or just wait until MS incorporates the most critical features into Visual Studio itself.

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Dec 18, 2011 9:38 PM in response to: Jakub Maly
Re: ReSharper performance

Hi

 

There are a number of threads about this problem. Indeed its seems the 6.1 EAP was pushed forward to at least show some alliviation of the problem.( the processing files issue)

 

I agree Just Code looks very promising , the guys at Telerik were very kind a while back and gave out free JC licenses to users who recommended their Just Decompile product .

 

I have had so many issues with R# that I have eventually had to uninstall on my production PC where I handle big solutions, including Db Projects. I am using a combination of Just Code and Whole Tomato's Visual Assist X (To make up for a lack of intellesense help in JC -- any time now they claim). They seem to lay well together.

 

I do appreciate trouble shooting these issues is difficult , but Jet Brains response always seems to be to send snapshots, videos , sample code etc

 

Its almost as if can't actually believe its their issue.

 

I am realtively new to R# (3 yrs) but I do smart at paying the initial V5 first licesne and then a further $ 120 for an upgrade not long after and it still limps along

 

My confusion is that I bought a product and I expect it to work . Sometimes I think its me then comments like yours pop up and give me some confidence to keep nagging Jet Brains.

 

I have tried most of the add ins over the years , and have this love hate relationship with R# , it is a superb product with lots of featured BUT , if it kills the IDE to the point of not working properly what use is it. When I uninstall I miss it , when I Install I curse it.

 

Another issue I don't like is that installing R# BLOCKS Just Code , it just dissaperas , so you can't even have them installed and switch betqween depending on performance needs.

 

It will stay uninstalled for now , until I get a warmer fuzzier feeling that all has come right.

 

Just Code is competing pretty well on most fronts , when intellesence appears Jet Brains must watch their market place .

 

Mike

Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Dec 19, 2011 10:13 AM in response to: Mike O'Neill
Re: ReSharper performance

Hello Mike

     The sole reason we ask for a snapshot or a sample solution is that we do not observe the mentioned behavior on test solutions that we have. Otherwise, it would have already been fixed. As you understand, it's almost impossible to fix a problem when there's no log information or at least some repro steps. As to "blocking" other products, there's absolutely no code in ReSharper to intentionally prevent competing products from working properly. We would really appreciate if you could find some time to try ReSharper 6.1 when it comes out. Thank you!

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Dec 19, 2011 10:34 AM in response to: Andrey Serebryansky
Re: ReSharper performance

Hi Andrey

 

As you can gather there is a  fair amount of frustartion in my reply

 

I seem to spend most of my time swapping between add ins trying to find the happy medium adding to the frustration.

 

The bottom line is that when it comes to the final analysis ,  R# is still the most fully featured product out there , and I inevitatbly give up and GO BACK , hence the Love Hate

 

It doesn't take the frustration away , if anything it makes it worse. I REALLY do want to use R# , please believe that . For all the hunting at near competitors none of em quite make the mark.

 

But with poor performance it just makes things worse. I am back (as ever) to R# this morning with another vow to perservere ...

 

WRT the Just Code block ,I appreciate its not deliberate but it is true , it may be courteous to your competitors to investigate why the clash exists and maybe work together to get around it . If for any reason I choose to Use JC instaed of R# , instead of a couple of ticks and a VS restart , its an complete install uninstall cycle, which takes time.

 

I am using the 6.1.29.21 EAP build currently (is their an better one ? )  and will keep monitoring

 

The Love Hate Continues

 

Mike

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Dec 19, 2011 4:21 PM in response to: Mike O'Neill
Re: ReSharper performance

Hi

 

Sorry to gripe BUT

 

I have now been using VS 2008 for 2 hours just general editting and  I am on my third Unhandled Out of

Memory exception and final crash and restart of VS 2008

 

its showing 141 Mb Managed Memeory and 444 Mb  In Task Manager for the ide

 

I thought this time I would submit the error BUT my account details seem not to allow it

 

Mike

Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Dec 19, 2011 4:39 PM in response to: Mike O'Neill
Re: ReSharper performance

Hello Mike

     I understand your frustration and I really hope that we will address this problem somehow. At the moment there's build #35 that is very close to release state and that has reduced number of assertions, logging and sensitivity to errors (which in terms should lead to better performance). Could you please install that build and write back in case it doesn't help with these problems? Thank you!

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Dec 20, 2011 4:48 PM in response to: Andrey Serebryansky
Re: ReSharper performance

Hi Andrey

 

Downloaded and installed build #35

 

Loaded up VS 2008 and a  VB Project , (funnily enough a different one than has been causing problems before , so its not obviously project specific.)

 

VS lasted 5 mins max   -- "Resharper has encountered a fatal error - out of memeory excpetion"

 

I tried to copy the error message for you but VS2008 bombed before I could . I restarted in Task Mgr IDE is showing 270 Mb with Resharper suspended and 600+ enabled, 190 managed

 

I have only Outlook open alongside VS , I can't do less.

 

I am a pretty loyal R# user but you guys are doing a great job of selling your competitors products

 

I simply can't work like this so its uninstall until you have a stable product , I am beginiing to wonder if its IF rather than WHEN

 

Sorry for the downer

 

Mike

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Dec 21, 2011 2:09 PM in response to: Mike O'Neill
Re: ReSharper performance

Hi

 

Managed to catch the exception , I was just generally editting , not building etc

 

I had to screen shot as I could copy from the dialog

 

Hope it points a directions

 

Mike

Attachments:
Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Dec 21, 2011 7:26 PM in response to: Mike O'Neill
Re: ReSharper performance

Hello Mike

     Thank you for this information! I've logged it under http://youtrack.jetbrains.net/issue/RSRP-287280 in our tracker and our devs will look into this problem. You're welcome to monitor the status of this problem.

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

Andreas Vilinski Newbie 8 posts since
Mar 24, 2010
Currently Being Moderated
Mar 9, 2012 1:06 PM in response to: Andrey Serebryansky
Re: ReSharper performance

I have also an annoying problem with my studio+resharper 6.1.1000.82. Sometimes it does not let me just edit my code. I must wait for a couple of seconds if i want to type a word. I thought first it happens only in some very big cs files ~4000 with a lot of inspection issues. But today i have the same issue in a relative small ~200 lines file with no bulbs. I have used dotTrace from EAP to generate a snapshot. During editing about 5 minutes and make some refactorings, dotTrace has generated a 5,5 Gig! snapshot. I have packed it to 1.8 Gig "bigCs.zip" and uploaded it per ftp to ftp://ftp.intellij.net/.uploads/ like described in http://confluence.jetbrains.net/display/ReSharper/ReSharper+Profiling+Instructions.

This is a .net 4.0 winforms solution with 20 cs projects

Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Mar 9, 2012 5:56 PM in response to: Andreas Vilinski
Re: ReSharper performance

Hello Andreas

     Thank you for feedback! At the moment I cannot find anything suspicious in the snapshot. Also, it shows that there are some other third-party extensions present and doing their work. First of all, could you please check if suspending ReSharper (under Tools | Options | ReSharper | General) helps to aviod these performance problems?

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

Andreas Vilinski Newbie 8 posts since
Mar 24, 2010
Currently Being Moderated
Mar 12, 2012 6:30 PM in response to: Andrey Serebryansky
Re: ReSharper performance

Yes, suspending R# helps definitely. But i will experiment with uninstalling of some VS extensions anyway.

Andreas Vilinski Newbie 8 posts since
Mar 24, 2010
Currently Being Moderated
Mar 27, 2012 6:04 PM in response to: Andrey Serebryansky
Re: ReSharper performance

well, seems that the villain is the other VS addin - GitScc! My suspection about resharper was probably my imagination... nevermind

Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Dec 19, 2011 10:03 AM in response to: Jakub Maly
Re: ReSharper performance

Hello Jakub

     We're sorry to hear that you're having such problems with ReSharper. Yes, performance is a common problem and we're doing our best to eliminate as many performance problems as we can (you can see the list of performance problems fixed in ReSharper 6.1 here: http://youtrack.jetbrains.net/issues?q=project%3A+RSRP+%23%7BPerformance+Problem%7D+%23fixed+fix+for%3A+6.1). Have you tried ReSharper 6.1? If yes, do I understand correctly that it didn't work fast enough for you?

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

Chris Marisic Newbie 83 posts since
Jan 13, 2010
Currently Being Moderated
Jan 6, 2012 12:27 AM in response to: Jakub Maly
Re: ReSharper performance

Now to provide a complete opposite repsonse, I use R# with solution analysis on multiple solutions with upwards of 4 visual studios open concurrently.

 

I never feel like R# impacts the performance of my machine.

 

My largest solution is only about 50 projects however.

 

Maybe the real issue is that your team needs to improve how you structure code that it has 150 projects in a solution? Or maybe your business needs to provide you with better hardware to do your job? For reference we develop in Win7 x64 on quad core 3.2ghz xeons with 12GB of ram. Have never in my life seen an OOME from R# since 2008 on a Windows XP machine with only 2GB of ram. If your hardware is crap, well what do you expect.

Chris Marisic Newbie 83 posts since
Jan 13, 2010
Currently Being Moderated
Jan 6, 2012 2:12 AM in response to: Jakub Maly
Re: ReSharper performance

Those specs are comparable to mine other than the ram being lower. There might be reasons to have 150 projects in the solution, doesn't mean it's correct.

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 6, 2012 9:27 AM in response to: Chris Marisic
Re: ReSharper performance

Hi Chris , I am really happy for you , that you can run a lovely spec PC as this

 

I work for a multinational corporate , PC are upgraded as part of our normal Capital Expenditure process , and we probably get a new one every 3-4 years. (99% of the company does email SAP , Web Apps and Office and needs no better)

 

Until literally 2 months ago , I was working on a Core 2 2.2 Ghz and 2 G of RAM. Thats it thats what you get . Add to that all the corporate baggage of network and antivirus protection etc , which takes up more resources.

 

A product like R# should carry a "Health Warning" if what you say is correct , "Do not even try to use our product on a Bog Standard PC"

 

I was upgraded in Nov , to a DELL Win 7 32 / i7 Quad 3.2 Ghz / 4G RAM/ SSHD , in anyones eyes a reasoanble PC , I still hit OOM Errors . At least this cycle my company recognised "Super Users" and we got better than the rest who got i5's

 

That said , how Jakub's compnay chose to structure their projects is their business , whether you or I believe it is wrong , we shouldn't have to change our ways of working to allow a product (like R#) to work.

 

I see memory leaks all the time while I am working which is also not right. Open Task Manager and look what memory Devenv.exe is consuming , say 400M to start with  , Now work away normally for 2-3 hrs and review the memory , its now up to 700-800 M .

 

Us mere mortals with 4G , only have 3G to play with anyway on win 32 and before we start with VS 2010 , Outlook and Windows , 2 GB is gone. The increase in 400 M caused by additional VS demand is a major part of our disposable memory , hence what I see and Jakub sees. As for multiple VS instances foget it , I use actively in my daily work any 2 of 2005 , 2008 and 2010, but I must manage it to run only one at a time.

 

I am sure that lots of factors come into play , no of projects in solution , size of individual files , how many files are open  etc. Solutio wide analysis to me is a no no , I simply turn it off because I know its going to cripple the IDE (SO why have it in the first place). My example of 400 M incresae is on a 4 project Demo Solution with about 40 classes , hardly rocket science.

 

I have bleated about this a lot in this forum but Jet Brains seems to not listen. I love R# dearly and feel "naked" without it , but while it is behaving like this and causing my productivity to drop NOT improve as it is designed to do , I just have to Suspend it (or uninstall it ) to carry on working effectively. And YES I do miss the functionality , but what comes first ? As an aside , if you get to the 800M stage and suspend R# , nothing changes , it still needs a restart of VS to clear the memory.

 

I am concerned that maybe R# has become "too fully funcioned" ( Bloated)  to the point where on normal PC's its becoming unusable , I don't think I am alone. Maybe Small Is Beautiful.

 

I have reviewed the opposition , and I already have a Visual Assist X license, so now I use VAX , and suffer the loss of functionality. 6.1 was promised as the panacea to all our perfrmance issues and they proudly show the list of performance issues they have fixed , I would ask why these were not identified in prelim testing. Even with these improvements ,, its clear from the comments in this forum that all is still not well in this area.

 

The Love Hate realtionship contiues and R# stays inactive. What a waste of money !!!

 

Every now and I relent and run R# , but inevitably I give up again

 

Mike

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 6, 2012 9:36 AM in response to: Mike O'Neill
Re: ReSharper performance

PS @ Andrey

 

Any move on the issue above , its marked as critical and has been sitting the re since well before Xmas ???

 

Mike

Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Jan 8, 2012 9:16 PM in response to: Mike O'Neill
Re: ReSharper performance

Hello Mike

     We've been looking into this problem, but I'm afraid we have no ideas at the moment. We'd need a memory dump or a sample solution to come up with something. Thank you!

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

Chris Marisic Newbie 83 posts since
Jan 13, 2010
Currently Being Moderated
Jan 7, 2012 12:23 AM in response to: Mike O'Neill
Re: ReSharper performance

Giving developers 32bit PCs is just silly. Your company is LITERALLY pissing away money. You're losing time due to your machine not having correct resources, time lost is money lost.

 

The cost of a PC in comparison to a software developer is measured in hours.

 

"we shouldn't have to change our ways of working to allow a product (like R#) to work."

 

That's also a silly statement.

 

R# is a gigantic database that sits on top of your code, to be able to expect R# to work well when your system has no available memory to host its database because your machine has too little memory and the solution has 150 projects in it (VS itself will probably take 1GB just from that, of your total 3gb if its 32bit).

 

Making these kinds of statements is like making the argument that SQL server is at fault for queries failing with OOME because the machine hosting it has 1GB of ram and your query is a god query that joins in every single table in the database together.

 

 

 


 

If anything I WISH R# USED MORE MEMORY. Put more in memory, less on disk, would make everything so much faster.

Eyal Shilony Newbie 76 posts since
Jul 12, 2011
Currently Being Moderated
Jan 8, 2012 12:45 AM in response to: Chris Marisic
Re: ReSharper performance

Maybe JetBrains should be more strict about the Requirements for R# to run smoothly ? both for Release and Debug builds.

 

The latest release build works pretty well for me.

Eyal Shilony Newbie 76 posts since
Jul 12, 2011
Currently Being Moderated
Jan 8, 2012 12:48 AM in response to: Chris Marisic
Re: ReSharper performance

Maybe Jetbrains should be more strict about the Requirements of the product for it to run smoothly ? especially debug builds.

 

The latest release build is working pretty good for me.

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 9, 2012 9:18 AM in response to: Eyal Shilony
Re: ReSharper performance

Hi Eyal

 

The problems I am encountering are with the 6.1 Release version , not EAP or nightly builds, I have been bitten a couple of times with "work in progresss" betas etc so I steer clear of them where I can

 

Mike

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 9, 2012 9:10 AM in response to: Chris Marisic
Re: ReSharper performance

  Welcome to the corporate world ...

 

I don't disagree , but while we are small development team (12 of us) in a BIG manufacturer of 15,000 employees , we have a small voice  too. Our prime business is manufacture , not software so our PC's are aimed at one size fits all , at least this time around we got a decent 32 Bit spec PC, up from the email and Word and Excel bashers

 

Life like that

Eyal Shilony Newbie 76 posts since
Jul 12, 2011
Currently Being Moderated
Jan 9, 2012 5:39 PM in response to: Mike O'Neill
Re: ReSharper performance

That's sucks. :|

Paul Bradshaw Novice 354 posts since
Aug 22, 2002
Currently Being Moderated
Jan 9, 2012 11:00 PM in response to: Mike O'Neill
Re: ReSharper performance

I too can vouch for those OOM errors going away with the switch to Win7

64-bit.  Even with only 4GB of memory, the problem was alleviated.

 

Any pressure you can provide to give your developers a 64-bit OS, and more

than 4GB of memory will be effort well spent (and investment for your

company well spent).  The amount of productivity we get back for that small

investment is more than worth it.

 

 

"Mike O'Neill"  wrote in message

news:23369969.52821326086015560.JavaMail.devnet@confluence.jetbrains.net...

 

Welcome to the corporate world ...

 

I don't disagree , but while we are small development team (12 of us) in a

BIG manufacturer of 15,000 employees , we have a small voice  too. Our prime

business is manufacture , not software so our PC's are aimed at one size

fits all , at least this time around we got a decent 32 Bit spec PC, up from

the email and Word and Excel bashers

 

Life like that

 

---

Original message URL: http://devnet.jetbrains.net/message/5448296#5448296

 

 

Charles Hooks Newbie 12 posts since
Sep 16, 2011
Currently Being Moderated
Jan 13, 2012 1:06 AM in response to: Jakub Maly
Re: ReSharper performance

Our company is also using 32bit OS.  Until we can upgrade all of hte legacey apps to 64 bits, which is still a year away...

 

We are stuck with 32 bits and R# is not usable for me with 6.1.  I'm lucky to get more than an hour of coding before an OOM crash occurs.

 

Without R# no crashes at all.  I had no problems before version 6 came out.  

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 13, 2012 9:00 AM in response to: Charles Hooks
Re: ReSharper performance

Hi Charles

 

Thats good (or bad ) to hear , I was beginning to think I was a lone voice out there. We are in the same boat , we have a large legacy app , half in VB ., half in C# . The SQL db project is also enormous.

 

Its unlikely we will upgrade this app much more , we just keep adding functionality . Its more likely to get replaced.

 

I find I keep 6.1 running by routinely restrarting VS every couple of hours , its a pain but it saves the crashes. I also reboot my PC daily to make sure I have no memory hang ups anywhere ( . we have new PC with solid state drives so this is no problem , a miniute or so , compared with our old PC's at 10 mins).

 

I do think that at the moment Jet Brains has abandoned hope on this , there is an incident in their system , since before Xmas which has received Zero attention.

 

As I have said so many times , R# IS the best tool , but not while it causes grief like this . Its a Love Hate relationship.

 

My suspicion is that because we shoudl all have super 64 Bit PC  etc with loads of RAM as our friend above comments, thats what they test on . They wouldn't see the pain we suffer.

 

Maybe some of their developers should Dog Food R# on a 32 Bit 4 Gb PC for a while , they may just wake up

 

MIke

Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Jan 13, 2012 12:27 PM in response to: Mike O'Neill
Re: ReSharper performance

Hello Mike and Charles

     There are certain limitations in memory management under 32-bit OS which can lead to memory fragmentation and consequent OOM exceptions. I'm afraid there's little we can do on our side in order to address this problem. By the way, have you tried using the /3GB switch which may help under some circumstances?

     Mike, our developers have reviewed that bug report, but at the moment we don't know what causes such big memory usage on your solution. We'd need a memory dump or a small sample in order to investigate this problem further.

    Thank you!

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 13, 2012 1:39 PM in response to: Andrey Serebryansky
Re: ReSharper performance

Hi Andrey

 

Sorry If I keep harping on ....

 

At least now there is something definitive said about 32 Win and OOM errors.

 

If I get this right , in effect you are saying that R#6.1 is not compatible with VS 2010 running under Win 32 if I am to use large solutions. (defifne large ?)

 

I have been using R# fro 3 years or so and I have never been aware that there is this limitation , is it described anywhere in documentation that I should have read before I bought it ?

 

I have been using 32 Bit Win for ever , and its unlikely I will change just to be able to use a specific tool, but I would like to have known about the limitation before I spent hard earned cash purchasing something which seems to be destined to cause me problems. Equally in my work environment I have no say over what PC I use , its simply provided.

 

As you can appreciate , I like the features of R# and I don't think there is anything else out there to come close to it , but while I get either OOM errors or have to restart VS every hour just to work with it , it seems a two edged sword , you can't easily work with or without it . Back to my Love Hate relationship.

 

I assume this limitation was not there in V5 , should I reinstall V5 maybe , would that help ?

 

Short of that, do Jet Brains have any suggestion as to what I should do , and more to the point what I should do with what appears to be a redundant product in my environment.

 

I apologise if this sounds irritating but I am irritated .

 

Mike

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 16, 2012 9:38 AM in response to: Mike O'Neill
Re: ReSharper performance

Hi

 

Another example of an OO Memory error, maybe to guide your thoughts

 

Been working around 45 mins , a variety of small editting tasks , and a compilation of a project . Then added a line of code and WHAM - does this help ?

 

---------------------------
ReSharper – Runtime Error – Technical Data
---------------------------
JetBrains ReSharper has encountered a runtime error.
Technical data follows.
Hint: Ctrl+C copies the text of system message boxes to Clipboard.
________________________________________________________________

Exception of type 'System.OutOfMemoryException' was thrown.

 

--- EXCEPTION #1/2 [OutOfMemoryException]

ExceptionPath = Root.InnerException

ClassName = System.OutOfMemoryException

HResult = -2147024882

Source = JetBrains.Platform.ReSharper.Util

StackTraceString = “

  at JetBrains.Util.TypeHierarchyMap`1..ctor()

     at JetBrains.Util.TypeHierarchyMapCached`1..ctor()

     at JetBrains.ReSharper.Features.Common.TreePsiBrowser.TreeModelBrowserPresenter..ctor()

     at JetBrains.ReSharper.Features.Browsing.CodeStructure.CodeStructurePresenter..ctor()

     at JetBrains.ReSharper.Features.Browsing.CodeStructure.CodeStructureBrowser.PerformBackgroundUpdate()

     at JetBrains.ReSharper.Features.Browsing.CodeStructure.CodeStructureBrowser.BackgroundUpdate.Work()

     at JetBrains.Application.InterruptableReadActivity.DoWork()

     at JetBrains.Application.InterruptableReadActivity.WorkerThreadProc()

 

--- Outer ---

 

--- EXCEPTION #2/2 [LoggerException]

Message = “Exception of type 'System.OutOfMemoryException' was thrown.”

ExceptionPath = Root

ClassName = JetBrains.Util.LoggerException

InnerException = “Exception #1 at Root.InnerException”

HResult = -2146232832

StackTraceString = “

  at JetBrains.Application.InterruptableReadActivity.WorkerThreadProc()

     at JetBrains.Threading.ThreadManager.PooledThread.ThreadProc()

     at JetBrains.Util.Logger.Catch(Action action)

     at JetBrains.Threading.ThreadManager.PooledThread.<.ctor>b__1()

     at System.Threading.ThreadHelper.ThreadStart_Context(Object state)

     at System.Threading.ExecutionContext.Run(ExecutionContext executionContext, ContextCallback callback, Object state, Boolean ignoreSyncCtx)

     at System.Threading.ExecutionContext.Run(ExecutionContext executionContext, ContextCallback callback, Object state)

     at System.Threading.ThreadHelper.ThreadStart()


---------------------------
OK  
---------------------------

Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Jan 16, 2012 12:13 PM in response to: Mike O'Neill
Re: ReSharper performance

Hello Mike

     This one comes from the File Structure window, so I'd suggest to check if working with this window closed helps to avoid OOM errors. Thank you!

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Jan 16, 2012 12:06 PM in response to: Mike O'Neill
Re: ReSharper performance

Hello Mike

     Perhaps I wasn't entirely correct in my previous post. What I wanted to say is: if one experiences OOM exceptions, one of the possible problems that we have found is memory fragmentation on 32-bit systems. In order to check if this is the case, one can run the system with a /3GB switch and make sure that this solves this problem.

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

Charles Hooks Newbie 12 posts since
Sep 16, 2011
Currently Being Moderated
Jan 13, 2012 6:22 PM in response to: Andrey Serebryansky
Re: ReSharper performance

I am not able to use the 3G switch, as it conflicts with some of the security and HIPAA software on my machine.

Paul Bradshaw Novice 354 posts since
Aug 22, 2002
Currently Being Moderated
Jan 27, 2012 8:36 PM in response to: Charles Hooks
Re: ReSharper performance

32-bit legacy apps should run just fine on Win7 64bit.

 

Visual Studio itself is a 32-bit app.

 

 

"Charles Hooks"  wrote in message

news:20760308.28271326402475889.JavaMail.devnet@confluence.jetbrains.net...

 

Our company is also using 32bit OS.  Until we can upgrade all of hte legacey

apps to 64 bits, which is still a year away...

 

We are stuck with 32 bits and R# is not usable for me with 6.1.  I'm lucky

to get more than an hour of coding before an OOM crash occurs.

 

Without R# no crashes at all.  I had no problems before version 6 came out.

 

---

Original message URL: http://devnet.jetbrains.net/message/5448853#5448853

 

 

Charles Hooks Newbie 12 posts since
Sep 16, 2011
Currently Being Moderated
Jan 28, 2012 2:39 AM in response to: Paul Bradshaw
Re: ReSharper performance

Unfortunately that is not true for most of our Borland apps.  They have custom interfaces with a variety of hardware to pirnters, inserters, scanners, bar code readers, etc.  We have tried to run them on x64 a couple of times and it has failed.

Chris Marisic Newbie 83 posts since
Jan 13, 2010
Currently Being Moderated
Jan 28, 2012 12:52 AM in response to: Jakub Maly
Re: ReSharper performance

I really have to wonder what the collective cost to these organizations that supplied their developers on this thread with inadequeate hardware to complete their jobs properly. I would really assume the time loss due to improper resources of just the few developers (and their related teams) in this thread would equate to hundreds of thousands, if not millions of dollars over a year.

Charles Hooks Newbie 12 posts since
Sep 16, 2011
Currently Being Moderated
Jan 28, 2012 2:34 AM in response to: Chris Marisic
Re: ReSharper performance

I don't see how your comment is relevant to the conversation, but for sake of argument...

 

Lots of companies have invested in software that only works on the x86 platform.  Many of these applications are expected to live for 5-15 years and possibly longer before replacement.

 

Now in the case of our company we are slowly porting various applications when they need changes...   Porting old Borland C++ win32 apps to Visual Studio.  We do build new apps for x64, but we still have to support the old.  Unfortunately, these old borland apps do not run on an x64 OS, even when we try x86 mode and we have 100's of these old apps.  So when we can we upgrade them, but until they're al upgrade we're stuck with x86 os.

Chris Marisic Newbie 83 posts since
Jan 13, 2010
Currently Being Moderated
Jan 28, 2012 9:52 PM in response to: Charles Hooks
Re: ReSharper performance

Excuses. Even with your organization being one of the rare needs for truly using x86, that still doesn't negate the need proper resources. With proper machines you could develop in Visual Studio in x64 and for legacy 32bit work develop inside virtual machine(s). This way multiple gigabytes of ram could be allocated even with XP's 3.2gb usable limit, with VMs you could easily run 3 VMs concurrently (even with only 12gb of ram) and dedicated 3.2gb to specific projects if they're gigantic.

 

My analogy that i used above is perfectly fitting. R# is a gigantic database that sits over your code. The larger the codebase, the larger the database. Blaming R# for using the memory it needs (which I would argue it uses substantially less than it should) and your system not being able to provide it, is a silly argument.

 

Would you seriously blame Redis, MongoDB or Sql Server that were in the exact same scenario that they're not provided enough resources to run the database it's told to run? Of course not.

 

Hardware is ridiciously cheap, wasted time is ridciously expensive as not only does it have direct costs on reduction on employee production it has opportunity cost attached to it making the time wasted a double cost. Lower productivity, and lower opportunity to do other things.

 

Regardless even if your organization refuses to replace your legacy hardware with modern hardware you still have numerous options, not using R# solution wide analysis, if that still doesn't push down the needs enough, at that point your projects need seperated. There's no need for solutions to have hundreds of projects. 50 projects is about the maxiumum that is ever reasonable in a solution. Past that you should be more concerned with sharing code in sane fashions aka nuget packages. There's no need for every project under the sun to be in 1 solution when you can easily share the dependencies it needs with nuget. Having solutions that collasal clearly shows no seperation of concerns, and falls into the god object anti-pattern. Code patterns and good design don't just apply to code, they can also apply to code structure and methodology.

 

You can also use nuget from the command line for sharing projects that aren't .NET projects. It wouldn't have as beautiful of an experience as it does inside Visual Studio, but it could definitely be used to share any code or dependencies in existense. Just the way you could use gems or npn to share anything.

Charles Hooks Newbie 12 posts since
Sep 16, 2011
Currently Being Moderated
Jan 28, 2012 10:08 PM in response to: Chris Marisic
Re: ReSharper performance

R# 5 worked fine, R# 6 crashed and died horribly.  Do a search of the forms and you'll see my request to back down to version 5, which went unanswered.

 

Regardless of your opinion on what a viable development shop should be or have.  There was no indication from JetBrians that upgrading to V6 would require such a drastic change in the development environment.

 

At this point, the lack of support and the virtol from posters like yourself is just pushing shops similar to the one I work in to other vendors.  You buy a product to solve problems, not create them and R#6 is creating more problems than it solves and changing our entire development environment is not an option.

Charles Hooks Newbie 12 posts since
Sep 16, 2011
Currently Being Moderated
Jan 28, 2012 10:13 PM in response to: Charles Hooks
Re: ReSharper performance

As for the comments on projects, all of our solutions have less than 10 projects with most just having 5 or 6.

 

The most common solution project listng is similar to the following listing:

Unit test Lib

Common Lib

Code Lib

Execution Harness (service, command line, etc...)

WiX installer

Fit Lib

Chris Marisic Newbie 83 posts since
Jan 13, 2010
Currently Being Moderated
Jan 28, 2012 10:42 PM in response to: Charles Hooks
Re: ReSharper performance

If that's the size of your solutions, unless there's like 8 billion code files in them, I would say yes R# should work on a 32bit machine. But this could directly relate to the # of visual studios open, the number of other apps running (are virus scanners messing with index files for R# etc)

Charles Hooks Newbie 12 posts since
Sep 16, 2011
Currently Being Moderated
Jan 28, 2012 11:24 PM in response to: Chris Marisic
Re: ReSharper performance

Only 1 visual studio is running.  However in our environment we have to use encrypted machines, so the whole drive is encrypted.  Plus the other 'special' hardware and the machines are dogs.

 

While I can run two instances of VS without R#, without problems a 3rd instance causes the same problems I get with 1 instance of VS with R# 6.0 or 6.1.  We are in process of getting solid state drives and some of my team members have them now, but that hasn't helped with the errors, though it has made everything so much faster that I am quite jealous of them.

Chris Marisic Newbie 83 posts since
Jan 13, 2010
Currently Being Moderated
Jan 28, 2012 10:50 PM in response to: Jakub Maly
Re: ReSharper performance

That's a compeltely inaccurate comparison. Visual Studio itself will occupy a large amount of memory. When you factor in you're probably running Web browser(s) or the apps your developing, Office products, Communication products, and any other software that is active on your system when you look at Windows XP you only have 3.2gb of available memory ever (assuming you have atleast that much physically).

 

Also there are many issues on Windows XP related to heap fragementation that even if your machine has enough bytes to store everything in memory that you can still get OOM exceptions because there isn't contigious blocks of memory large enough.

 

Your comparison with Firefox would only be accurate if the addon hoists a large a database on top of firefox.

 

In response to your statement "if your customers complain a product is slow", if my product was a database, I would have no issues informing them better hardware = better experience. Of course I also would likely make older versions available that are less robust for legacy hardware.

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 30, 2012 9:38 AM in response to: Chris Marisic
Re: ReSharper performance

Nice Debate , I missed it as I was sunbathing !!

 

While I do not wholly disagree with the resource loss argument from Chris , but after 40 years in a corpoarate world you learn that sometimes we have to live with what we are given and make the best. Debating 32 v 64 bit architecture with an accountant normally has little if any impact . In a dedicated software house the argument may wash  , I am not in that environment we are a FMCG comapany.

 

I do however take up the previous argument, if R# 6 was going to be so much more demanding that previously maybe warn us before we but it. Finding out when it crashes VS is not really the way to do it.

 

I do feel a bit agreived by Jet Brains " innocent until proved guilty "  approach they always want to see "proof" before they investigate. We all develop software , or we would be commenting here, and we all know how difficult it can be to root out some lurking performance issues, but Come Clean and admit they may be there maybe ....

 

I am working on 2 sorts of projects , big complex ones at work , small , hopefully well designed ones at home. The home ones are max 6 structured projects to a solution, and the one I am working on at the moment as my reference point has maybe 30 classes per project. I use this as a reference to see what changes , as this project doesn't. So the complaints based on Monolithic Solutions is not valid.

 

The real dilemma is that when you look for an alternative to R# , it isn't there . Visual Assist X is brilliant , but only covers 50% of R# . Just Code is good and improving but they seem to not see a need for Intellesence , CodeRush is Just Different....

 

So we get back to using R# and put up with the issues , but in our case why put up and shut up when we can debate as we are doing . Maybe Jet Brians will comment on this thread sometime , Andrey has commented but , with all due respect, without really joining the debate.

 

I am rapidly looking towards Visual Assist X and Code Rush Express to provide the best coverage without the memory issues.

 

Lets see what Jet Brains can come up with

 

Mike

Bobby Cannon Newbie 39 posts since
Jul 7, 2011
Currently Being Moderated
Jan 30, 2012 10:33 PM in response to: Mike O'Neill
Re: ReSharper performance

MikeONeill wrote:

 

... CodeRush is Just Different....

 

Mike

 

Yeah, I'm glad someone agrees. I tried to switch to CodeRush but I just couldn't. It's definantly a different way of thinking. Too bad I just can't think that way...

 

Thanks,

Bobby Cannon

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 31, 2012 11:02 AM in response to: Bobby Cannon
Re: ReSharper performance

Hi Bobby

 

Bit rude advertising competitors BUT ...

 

Maybe try a demo of Visual Assist X if you haven't already , its a bit C++ Biased and has a few holes in VB  Refactoring but for C# its fine. The Tools like Outliner match R# equally .

 

its very similar to R# but covers a lot less , and I think is less memory demadning as a consequence. I have used the freebie Coderush Express along side with no issues just to supplemt the shortfalls in Refactoring.

 

Mike

Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Jan 31, 2012 6:55 PM in response to: Mike O'Neill
Re: ReSharper performance

Hello Mike and All

     I'm afraid this discussion has gone off topic, because the main purpose of this community is asking questions about Resharper and reporting bugs.

 

     In order to summarize the problems mentioned here I kindly ask everyone involved in this discussion to install the latest build of ReSharper 6.1.1 from http://confluence.jetbrains.net/display/ReSharper/ReSharper+6.1.1+Nightly+Builds and reply to this message with description of the problems that you've encountered with this build. That will give us some common ground to continue working with.

     Thank you!

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 31, 2012 8:33 PM in response to: Andrey Serebryansky
Re: ReSharper performance

Hi Andrey

 

I am sorry if dicussions here hit a sore point , I have given up on R# , I am trying to find something that replaces your functionalty , take it as a complement that this is not easy

 

If I find a fellow sufferer then so be it

 

I have been a devoted R# user for years and find it very hard to give it up , but while I cannot use it without VS 2010 crashing what am I to do, I have often used the Love Hat comparison , I lobe the product but I simply cannot live with its foibles , Am I Alone ???

 

If you find my comments re Jet Brains painful then maybe you should take note of that  and do something about it.

 

I spent good money upgrading to V6 and I have had nothing but pain since , I had few peoblems with V5 , and even less with with your competitors products

 

I will stop contributing to this thread if that please you , after all censorship keeps the masses in the dark

 

GOODBYE , and goodbye to Jet Brains products if thats what you want

 

Mike

Mike O'Neill Newbie 46 posts since
Sep 17, 2010
Currently Being Moderated
Jan 31, 2012 8:39 PM in response to: Mike O'Neill
Re: ReSharper performance

PPS

 

Can I have a refund of my upgrade costs ?????

Andrey Serebryansky JetBrains 2,747 posts since
Jan 12, 2005
Currently Being Moderated
Jan 31, 2012 10:43 PM in response to: Mike O'Neill
Re: ReSharper performance

Hello Mike

     Please take my apologies for the poor experience that you've had with ReSharper. I sincerely hope that in the future you'll find some time to install newer versions and they will work better for you. Sure, you can have a refund. Please send me an e-mail to andrew dot serebryansky at jetbrains dot com.

 

Andrey Serebryansky

Senior Support Engineer

JetBrains, Inc

http://www.jetbrains.com

"Develop with pleasure!"

More Like This

  • Retrieving data ...