Memory leak found in Samsung SSD Magician

[P4-630]

Latest version

 

[trparky]

Latest version

View attachment 271621

It's not that that was causing an issue, you had to dive a lot deeper into the OS to notice the memory leak. Like I stated in my second post of this thread, you had to use specialized utilities to see the memory leak.

I was over a period of time watching my system memory get eaten so I decided to do some looking around and downloaded a utility called RamMap from SysInternals and I noticed a rather huge amount data being used for memory mapped files. So, I went to the Files Summary tab and sorted all memory mapped files according to size and there to my horror there was a history data file being used by Samsung SSD Magician located in "C:\ProgramData\Samsung\Samsung Magician" and it was using up to nearly 16 GBs of system RAM.

 

[P4-630]

It's not that that was causing an issue, you had to dive a lot deeper into the OS to notice the memory leak. Like I stated in my second post of this thread, you had to use specialized utilities to see the memory leak.

I had such issues years ago when I used hibernate/fastboot, since then never used hibernate/fastboot anymore and no more issues with apps eating my RAM...

 

[Mr Bill]

One might think that if you have enough system RAM you can disable the page file but that's not true at all. Windows still needs the page file no matter how much RAM you have. You could have 64 GBs or even 128 GBs of RAM, the page file is still needed.

Should I Disable the Page File if My Computer Has a Lot of RAM? (howtogeek.com)

This has been going back and forth for years. Since the pagefile is a disk-based system it would add more write cycles to your SSD. I've never used a page file with any ssd and never had any issues, go figure.

 

[ThrashZone]

Hi,
I'm nearly at system managed page file
I minimize to 16mb and max it at 5gb running 32gb memory ssd's still alive and kicking after years of use.

 

[Bill_Bright]

This has been going back and forth for years. Since the pagefile is a disk-based system it would add more write cycles to your SSD.

So what? You do make a valid point about this going back and forth for years. But clearly you failed to heed the lessons learned over those years.

Limited write cycles for SSDs has not been a problem for years now. And frankly, it never was for "normal" computer users. Continuing to spew such nonsense is just spreading ignorance and FUD.

If you have a fairly recent laptop, odds are it has a SSD only - no hard drive. And by default, it has a Windows managed PF on that SSD. And guess what? No issues either. Go figure.

I've never used a page file with any ssd and never had any issues, go figure.

Sure you did. But by refusing to let go of long-held, outdated, obsolete, and no longer applicable beliefs and take the time to learn the true facts, those issues are never realized.

The PF on even the slowest SSD provides superior performance compared to having the PF on the fastest hard drives. And 99% of use will never come close to reaching the write limits on current generation SSDs. It takes years of constant writes to reach those limits.

"Because I always did it that way", "I disabled it and didn't see any difference", and "limited number of writes on SSDs" are three of the lamest excuses ever - when it comes to disabling Page Files, current versions of Windows, and Page Files on SSDs.

"Because I just wanted to!" makes much more sense and logical excuse for dinking with the default PF settings than any of those three lame excuses.

Are there exceptions? Of course. But it is highly unlikely any computer used by any one on this site fits that category.

 

[eidairaman1]

So, here I am, and I found a memory leak in Samsung SSD Magician. And I mean serious one, the kind that only after three or four days, all of your system RAM will be used and potentially crash your system. How do I go about reporting this issue to Samsung because I can't for the life of me find a contact to send this to info to Samsung.

Oh sure, there's Customer Service but I need to get into contact with someone who can I send technical details about what's happening.

So dont have it running in background, only use it for firmware updates

 

[trparky]

So dont have it running in background, only use it for firmware updates

That’s exactly what I’ve done.

 

[Mr Bill]

So what? You do make a valid point about this going back and forth for years. But clearly you failed to heed the lessons learned over those years.

Limited write cycles for SSDs has not been a problem for years now. And frankly, it never was for "normal" computer users. Continuing to spew such nonsense is just spreading ignorance and FUD.

If you have a fairly recent laptop, odds are it has a SSD only - no hard drive. And by default, it has a Windows managed PF on that SSD. And guess what? No issues either. Go figure.


Sure you did. But by refusing to let go of long-held, outdated, obsolete, and no longer applicable beliefs and take the time to learn the true facts, those issues are never realized.

The PF on even the slowest SSD provides superior performance compared to having the PF on the fastest hard drives. And 99% of use will never come close to reaching the write limits on current generation SSDs. It takes years of constant writes to reach those limits.

"Because I always did it that way", "I disabled it and didn't see any difference", and "limited number of writes on SSDs" are three of the lamest excuses ever - when it comes to disabling Page Files, current versions of Windows, and Page Files on SSDs.

"Because I just wanted to!" makes much more sense and logical excuse for dinking with the default PF settings than any of those three lame excuses.

Are there exceptions? Of course. But it is highly unlikely any computer used by any one on this site fits that category.

Bill, I didn't expect a sermon, I was just speaking about my preference, what others do, is really none of my business.

 

[Bill_Bright]

It was not a sermon. It was reporting the facts and putting them in the proper perspective so others reading are not misinformed.

Implying, as you did, that SSDs today will somehow suffer if one keeps the PF on them is misleading at best - to put it nicely.

Expressing your preference is fine. I have no problem with that at all. In fact, I will defend your right, with vigor, to do so. But then it needs to be expressed as your preference or opinion. When you imply or suggest something that is not a fact, I will use the same vigor to ensure readers know the true facts.

To illustrate the facts, note this budget $60 500GB SSD, has a TBW (terabytes written) rating of 300TBW. That means, you would have to fill that 500GB drive up, then delete everything, and fill it up again 600 times before you reached its write limit. How likely is that even with a busy page file? Very unlikely considering most accesses to PF are reads, not writes.

FTR, with 1 TB equal to 1000GB, that is 300,000GB.

 

[trparky]

Implying, as you did, that SSDs today will somehow suffer if one keeps the PF on them is misleading at best - to put it nicely.

I have owned several SSDs over the last eight or so years and I don't treat them lightly; I use and abuse them. I don't treat them any differently than a traditional hard drive. I don't care if I'm doing unneeded write operations, I don't care if I unzip and process compressed files all the live long day, etc. And do you want to know something folks? Oh yeah, none of them have ever reached the point where I had to retire them due to write operation limits. Not once.

Hell, I still have two smaller SATA SSDs that are just sitting around doing nothing, yet they still have a hell of a lot of life left and I too didn't treat them any differently. I used and abused them too. Again, no limits reached. I only retired them due to upgrading to bigger drives.

So yeah, this worry about reaching some arbitrary write limit is nothing more than a scare tactic, bullshit at worst. Go ahead, use your SSD and don't be afraid. Trust me, it can take a hell of a lot more abuse than you think it can.

 

[Mr Bill]

It was not a sermon. It was reporting the facts and putting them in the proper perspective so others reading are not misinformed.

Bill, it's ok, don't get your blood pressure up, everything is going to be ok, trust me, I don't have any patent on misinformation on the internet.

 

[Deleted member 185158]

Bill, it's ok, don't get your blood pressure up, everything is going to be ok, trust me, I don't have any patent on misinformation on the internet.

Its not the accurate approach though.

Because of just the comment of using it manually disabled, this promotes shutting it off.

But as you can see, the wording and play are a thing.

I do it because I can. Because it's an option to do so.

When benchmarking and disabling all services, page file is also then disabled. I tried to enable it (w7) with all services disabled, the option is grey'd out. But benchmarking, as I plainly stated once in another thread is something that I do.

Then, once apon a time, try and reflect that the drive space is allocated. When you have enough system memory, it's unused allocated space.

In addition to, even supplied instructions how to measure usage through performance monitor.

I even made a thread and asked at my benching site.

The concensus is to just leave windows managed size.

Because the software engineers designed it that way. For memory management. Which I cannot see a difference when it's enabled or disabled for my personal preferences.

You just simply need to be careful how you express your actions on your personal set up. Often the meaning of what you said will be taken out of context.

I guess you could start by saying to people not to enable it, but I do Because it suits my personal needs. This does not mean it should become the norm for average users. Windows needs it even if optional to disable it.

 

[Mr Bill]

You just simply need to be careful how you express your actions on your personal set up. Often the meaning of what you said will be taken out of context.

I agree, but my intention was just an opinion from my own personal experience. I would hope that anyone asking questions on any forum would follow up with their own research, and not take one's opinion/advise as gospel, until spending some time researching other opinions. You've probably heard the old saying "You are entitled to your opinion. But you are not entitled to your own facts". Every time in the past 10 to 15 years I've personally googled/researched this question, I found pro's and con's to be around 50% for and 50% against, with that said, who do we believe? Sometimes we have to go with one side, I chose the "against" side and it's worked great for me, especially regarding the ssd drives. If I were doing some major opening and closing of programs that hog a lot of memory on my PC, I would put the paging file on another drive. Next time I get involved in a questionable thread on here, I will make sure I put from my own personal experience, but please do your own research, because your experience may differ.

 

[Deleted member 185158]

I agree, but my intention was just an opinion from my own personal experience. I would hope that anyone asking questions on any forum would follow up with their own research, and not take one's opinion/advise as gospel, until spending some time researching other opinions. You've probably heard the old saying "You are entitled to your opinion. But you are not entitled to your own facts". Every time in the past 10 to 15 years I've personally googled/researched this question, I found pro's and con's to be around 50% for and 50% against, with that said, who do we believe? Sometimes we have to go with one side, I chose the "against" side and it's worked great for me, especially regarding the ssd drives. If I were doing some major opening and closing of programs that hog a lot of memory on my PC, I would put the paging file on another drive. Next time I get involved in a questionable thread on here, I will make sure I put from my own personal experience, but please do your own research, because your experience may differ.

Every thread can be skewed at this forum, simply that you are not exactly entitled to your opinion, because it's not a fact.

Which is only a 50% bull blank statement and would only pertain if you make claim as it a fact.

Well your opinion is actually fact, because you make practice with the experience. And simply share it with others.

But so far, I couldn't get a screen shot of anyone with a pc that would be in this site that could produce page file usage in the described manor of actually loading up past the memory availability, pushing system memory to page file and share THAT experience.

None. Zip. Nadda. Zero shared page file usage experiences. Not a one.

 

[Mr Bill]

Every thread can be skewed at this forum, simply that you are not exactly entitled to your opinion, because it's not a fact.

Which is only a 50% bull blank statement and would only pertain if you make claim as it a fact.

Well your opinion is actually fact, because you make practice with the experience. And simply share it with others.

But so far, I couldn't get a screen shot of anyone with a pc that would be in this site that could produce page file usage in the described manor of actually loading up past the memory availability, pushing system memory to page file and share THAT experience.

None. Zip. Nadda. Zero shared page file usage experiences. Not a one.

Yeah, I did a little googling, this was about the best opinion or fact "you choose" I could find.

"This old chestnut crops up again and again, and you get a variety of responses, but none of the really get to the root of the issue.

Historically pagefiles were used to cope with programs needing more RAM that pc had installed, and most of the advice was to reserve (say) 1.5 x disk space than RAM.

However, in time, RAM has become much cheaper, so the dependence on pagefiling is less critical BUT it really depends on what user is doing e.g. some activities are very RAM hungry e.g. rendering large videos, or running huge database type applications, having lots of tabs, apps open etc.

The other factor is that now Windows 10/11 uses compressed memory to minimise diskpaging, offloading compressed memory if that physical RAM is needed.

The relaptionship between compressed memory and pagefiling is a dynamic one and so best managed by OS.


Also, people cite other chestnuts about reducing wear and tear on SSDs.

Again, perhaps some truth back in days of 1st Gen 64GB SSDs, but now most new ones are 512MB plus with much more wear and tear resilience, and the pagefile impact on wear is less significant to point of being almost irrelevant now.

So, there really is no serious driver to set pagefiles manually these days.

Ask yourself if you think you know better than the OS in setting pagefile size manually?

The crux of most peoples arguments is I have plenty of RAM, thus pagefiles are not relevant.

Of course, this is true EXCEPT for when it isn't!

In the end, its a bit like having a modern DSG automatic car vs a manual gearbox. Many claim you will get better 0-60 acceleration performance with a manual rather than a DSG. This may be true of the driver is a professional racing driver but for mere mortals, DSG acceleration will easily out perform manual acceleration operation anyday".

 

[Bill_Bright]

Historically pagefiles were used to cope with programs needing more RAM that pc had installed, and most of the advice was to reserve (say) 1.5 x disk space than RAM.

Ummm, I am going to assume you meant to say, most of the advice was to set the size at "1.5 x RAM". Not disk space. But I quickly add, that advice was arbitrary back then, and most likely bad, or at least not good, now. And more importantly, W10 and W11 are NOT Win3.0, W95/98 or WinXP. People need to stop treating them like they are.

And it was (and is) not just about programs want more RAM, but the OS too.

BUT it really depends on what user is doing

This is VERY TRUE! However, how many users do the exact same task, over and over again, day in and day out, week after week, year after year? Any one? Probably not. This means the demands we put on our systems are constantly changing. We see this in the widely varying demands on our PSUs, CPUs, cooling and memory. My point? The page file size is NOT a set and forget setting. It is why Microsoft wisely made it dynamic - and never automatically disables it, even though it could easily be coded to "IF" it was better to do so.

Ask yourself if you think you know better than the OS in setting pagefile size manually?

??? Have you asked yourself this question? And not just of the OS, but do you (speaking to the crowd) know better than the "teams" of PhDs, computer scientists and engineers, and programmers who code the OS? Why do I emphasize "teams"? To point out it is not just one nerdy computer programmer making these decisions. But large groups of bone fide experts combing over exabytes of empirical data, peer reviews, study panels, etc.

You make some great arguments for just leaving the defaults alone! But then you change them yourself - while admitting you are not a memory management expert. Ask yourself, does that really make sense?

The other factor is that now Windows 10/11 uses compressed memory to minimise diskpaging, offloading compressed memory if that physical RAM is needed.

Nope! That is NOT a factor here. The fact W10 and W11 use memory compression (because it is faster than paging to a disk, even a fast SSD) does NOT negate the advantages to having a PF too. Memory compression is NOT a replacement or substitute for a PF. Memory compression is an additional tool. Memory compression and the PF are complementary features that work together to help improve performance.

And while Windows can run without a PF if you have a large amount of RAM, that does NOT, in any way, suggest it is better to disabling it when there is a large amount of RAM.

For those unfamiliar, here is a good explanation of memory compression. I invite folks to read it, then follow the How-To Geek's link to his related article, What Is the Windows Page File, and Should You Disable It? and note where he reports (my bold underline added),

Myth: Disabling the Page File Improves Performance

Some people will tell you that you should disable the page file to speed up your computer. The thinking goes like this: the page file is slower than RAM, and if you have enough RAM, Windows will use the page file when it should be using RAM, slowing down your computer.

This isn’t really true. People have tested this theory and found that, while Windows can run without a page file if you have a large amount of RAM, there’s no performance benefit to disabling the page file.

However, disabling the page file can result in some bad things. If programs start to use up all your available memory, they’ll start crashing instead of being swapped out of the RAM into your page file. This can also cause problems when running software that requires a large amount of memory, such as virtual machines. Some programs may even refuse to run.

In summary, there’s no good reason to disable the page file — you’ll get some hard drive space back, but the potential system instability won’t be worth it.

The crux of most peoples arguments is I have plenty of RAM, thus pagefiles are not relevant.

Of course, this is true EXCEPT for when it isn't!

Therein lies the problem. This of course, is NOT true, except for when it is. For the vast majority of us, there is no advantage to disabling the PF, regardless how much RAM we have installed.

I agree, but my intention was just an opinion from my own personal experience

And that would be fine, had you expressed your opinion as an opinion. Instead, you expressed it as a statement of fact. If this were a newspaper op-ed column, that would be fine too. But this is a technical support forum. We, as advisers, need to be technically correct - especially when making statements that imply, or suggest something not true or correct.

 

[Mr Bill]

Yeah, I did a little googling, this was about the best opinion or fact "you choose" I could find.

"This old chestnut crops up again and again, and you get a variety of responses, but none of the really get to the root of the issue.

Historically pagefiles were used to cope with programs needing more RAM that pc had installed, and most of the advice was to reserve (say) 1.5 x disk space than RAM.

However, in time, RAM has become much cheaper, so the dependence on pagefiling is less critical BUT it really depends on what user is doing e.g. some activities are very RAM hungry e.g. rendering large videos, or running huge database type applications, having lots of tabs, apps open etc.

The other factor is that now Windows 10/11 uses compressed memory to minimise diskpaging, offloading compressed memory if that physical RAM is needed.

The relaptionship between compressed memory and pagefiling is a dynamic one and so best managed by OS.


Also, people cite other chestnuts about reducing wear and tear on SSDs.

Again, perhaps some truth back in days of 1st Gen 64GB SSDs, but now most new ones are 512MB plus with much more wear and tear resilience, and the pagefile impact on wear is less significant to point of being almost irrelevant now.

So, there really is no serious driver to set pagefiles manually these days.

Ask yourself if you think you know better than the OS in setting pagefile size manually?

The crux of most peoples arguments is I have plenty of RAM, thus pagefiles are not relevant.

Of course, this is true EXCEPT for when it isn't!

In the end, its a bit like having a modern DSG automatic car vs a manual gearbox. Many claim you will get better 0-60 acceleration performance with a manual rather than a DSG. This may be true of the driver is a professional racing driver but for mere mortals, DSG acceleration will easily out perform manual acceleration operation anyday".

Bill, just to clarify, all this in red is something I found and copied and pasted, none of this was my opinion or facts, I'm not smart enough or maybe dumb enough to even begin to argue page file pro's or con's. Again for all the forgiving folks here, I'm sorry I even got involved in this decades old argument, I'm sure no one here has ever went against the grain of someone elses opinion on TPU, I probably have a patent on that.

 

[Deleted member 185158]

Yeah, I did a little googling, this was about the best opinion or fact "you choose" I could find.

"This old chestnut crops up again and again, and you get a variety of responses, but none of the really get to the root of the issue.

Historically pagefiles were used to cope with programs needing more RAM that pc had installed, and most of the advice was to reserve (say) 1.5 x disk space than RAM.

However, in time, RAM has become much cheaper, so the dependence on pagefiling is less critical BUT it really depends on what user is doing e.g. some activities are very RAM hungry e.g. rendering large videos, or running huge database type applications, having lots of tabs, apps open etc.

The other factor is that now Windows 10/11 uses compressed memory to minimise diskpaging, offloading compressed memory if that physical RAM is needed.

The relaptionship between compressed memory and pagefiling is a dynamic one and so best managed by OS.


Also, people cite other chestnuts about reducing wear and tear on SSDs.

Again, perhaps some truth back in days of 1st Gen 64GB SSDs, but now most new ones are 512MB plus with much more wear and tear resilience, and the pagefile impact on wear is less significant to point of being almost irrelevant now.

So, there really is no serious driver to set pagefiles manually these days.

Ask yourself if you think you know better than the OS in setting pagefile size manually?

The crux of most peoples arguments is I have plenty of RAM, thus pagefiles are not relevant.

Of course, this is true EXCEPT for when it isn't!

In the end, its a bit like having a modern DSG automatic car vs a manual gearbox. Many claim you will get better 0-60 acceleration performance with a manual rather than a DSG. This may be true of the driver is a professional racing driver but for mere mortals, DSG acceleration will easily out perform manual acceleration operation anyday".

Well, before you do the back and forth with BB, which he knows his stuff, there's no denying it.....

In testing/practice, I have found page file to be allocated. In my personal experience, not actually being utilized. At least, not on a measurable level.

So at some point, I am going to load up my system memory and see if a 980pro can act like system memory and get a wonderful experience from it, but I think this might be a misconception. The page file is there to prevent data loss when all system memory is used.

Also, maybe some small read or writes through commit charge, but would have to be less than 1% of available storage allocated. Because it's not measured that I'm aware of, at least not measured within the allocated space.

 

[ThrashZone]

Well, before you do the back and forth with BB,

Also, maybe some small read or writes through commit charge, but would have to be less than 1% of available storage allocated. Because it's not measured that I'm aware of, at least not measured within the allocated space.

Hi,
To late for that

Not sure what the almighty mad ducks argument against disabling page file is really, beside another pointless argument of who's correct or not seems a tad off topic for sammy magician issue at hand

Many people disable page file even myself at one time think all my benchmark installs I do disable it and system protection as well
All installs I disable hibernation to so unless one has an issue, the argument against is pretty lame imho

 

[Deleted member 185158]

Hi,
To late for that

Not sure what the almighty mad ducks argument against disabling page file is really, beside another pointless argument of who's correct or not seems a tad off topic for sammy magician issue at hand

Many people disable page file even myself at one time think all my benchmark installs I do disable it and system protection as well
All installs I disable hibernation to so unless one has an issue, the argument against is pretty lame imho

Bill Bright is just making sure we don't share misinformation.

And he has every right to do so, and he's been factual about it too.

The reality of allocating drive space for swap file is an industry standard when the OS is installed. So the recommendation is to just leave it alone and let windows manage it. Which I agree with 100% for average users.

 

[Bill_Bright]

I'm sorry I even got involved in this decades old argument

And this, so much, is part of the problem.

My 1970 Toyota Corolla did this, so I am modifying my 2022 Corolla so it won't do that.

That's the crux of my point here, TZ.

You are correct, however, in that it is way off the topic of Samsung Magician. So time to move on.

 

[chrcoluk]

I agree - or at least feel one would hope.

To be sure, I typically am one to recommend users use the drive maker's own diagnostics programs too (rather than 3rd party tools beyond chkdsk) when they encounter problems. I base that on the fact I would assume the maker is the most qualified. So we are on the same page with that philosophy; that is, the maker knows best.

However, SM is not really a diagnostics program and I personally consider it as yet another unnecessary, bloated and intrusive [nearly] "crapware" program manufacturers constantly try to "foist" on us under the pretense it is needed - very much in the same manner laptop makers pre-load extras we don't need, and especially printer makers who attempt to foist all kinds of junk on us - when the reality is, all we really need is the basic drivers which very often are already integrated with the OS.

So the part that bothers me the most is how these otherwise totally reputable manufacturers market all these extra resource consuming programs as though (1) we need them and (2) how they often imply they "improve performance" - when in fact, neither is true.

Again, today's operating systems know how to optimally use and maintain SSDs just fine without any 3rd party software stuck in the middle, starting with the OS , running in real-time, consuming resources.

It is foisting of extras that often cause users to complain (and rightfully slow) that their computers are bogging down and no longer run as well as they used to. No wonder! Every time they install a new program piece of hardware, some "auto-update" program gets foisted on the system too - if the user is being watchful.


Yes, be we don't need yet another program running all the time, constantly phoning home to check to see if a new update is available. I note that firmware and driver updates (with the possible exception of graphics drivers) typically fall into the category of, "if it ain't broke, don't fix it!"

For sure, I don't have a problem running SM "manually" to check for updates, or to check S.M.A.R.T. status, or run a utility command on it like Secure Erase. I don't like, or see the need to have SM start with Windows every time we start our computers.

I can give the reasons I have used it.

LBA sector scan - easily done in linux, but windows doesnt have any software that I know off that does this, chkdsk /r only reads sectors that hold data, not all sectors.

Can run SMART self tests - dont know of any windows tools that let you do this, easy in linux though. Also have not found a way to run self tests on nvme drives in linux, the only way I know of is via the samsung magician (and possibly other vendors) tools.

Check for firmware update.

I dont keep it running 24/7 in background though, its just used on demand when needed.

 

[Bill_Bright]

chkdsk /r only reads sectors that hold data, not all sectors.

Huh? Not true at all and never was. Of course it reads all sectors, regardless if it currently holds data, or not.

Did you mean chkdsk /f?

chkdsk /r certainly does scan the entire drive and will attempt to fix any errors it finds. chkdsk with the /r switch will also locate bad sectors and attempt to recover any data from those sectors.

chkdsk /f, on the other hand, does not check each location. Since the /r switch implies (or includes) /f, I always recommend using the /r. If one enters chkdsk /f /r, the /f will be ignored.

This is simple to see for yourself. Get a brand new drive, or format an old and run chkdsk on it and watch the progress. On a blank drive, chkdsk /f will take seconds. But chkdsk /r will take minutes - even hours on a very large drive. Look for activity in Stage 5 of the progress.

Your other reasons to have Samsung Magician available are fine - I never disagreed with them.

I said, and you quoted me, I see no need to have it start and run in memory every time we start Windows. I stand by that.

 

[chrcoluk]

when I last used chkdsk /r on a zeroed drive it finished extremely quickly, I can try it again though to see if its changed.