Core i5 and Core i7: there is difference in a single pc for an home use ?

What? Are you sure it was moderate? Sounds like an insane amount of caching or something like that.

 

@J_L Yes, my system would sometimes benefit from having more than 4GB of RAM due to always having a very large number of browsers tabs open. But, I if I upgraded to 6GB, I'm certain that memory use would never exceed 74%.

 

About a dozen Chrome tabs (Chrome is a memory hog...every process is its own instance for security). Caching wasn't included in that figure--which is another Windows feature not discussed in regards to RAM: if you're using most/all of it for running processes, you aren't benefitting from caching/superfetching.

 

@Rolo42 I've got 30+ tabs open right now in 360 Extreme Explorer (using Chromium 45) under Windows 10 with 4GB of RAM, and my memory use is only at 54%.

 

That's not useful as we aren't browsing the same sites, which can be static text to presentations and players (which some of mine were) and we probably don't have the same extensions. Another issue is pagefile configuration: you'll use less RAM if some of those tabs were swapped (mine were not as they were all active).

I'm using almost 4GB RAM with six tabs and The Night of the Rabbit minimised (a small game). systeminfo output:

Code:

Total Physical Memory:     8,150 MB
Available Physical Memory: 4,236 MB
Virtual Memory: Max Size:  9,430 MB
Virtual Memory: Available: 3,451 MB
Virtual Memory: In Use:    5,979 MB

Wife typically has about 30 tabs open and her RAM is typically full (real-time CPU and RAM monitor on our keyboards). Our next builds will definitely have at least twice the baseline RAM installed.

My tasklist output (6 tabs):

Code:

C:\WINDOWS\system32>tasklist

Image Name                     PID Session Name        Session#    Mem Usage
========================= ======== ================ =========== ============
System Idle Process              0 Services                   0          4 K
System                           4 Services                   0     46,976 K
smss.exe                       340 Services                   0      1,176 K
csrss.exe                      556 Services                   0      3,788 K
wininit.exe                    644 Services                   0      4,628 K
csrss.exe                      652 Console                    1      7,960 K
services.exe                   716 Services                   0      6,744 K
lsass.exe                      724 Services                   0     50,024 K
winlogon.exe                   792 Console                    1      8,972 K
svchost.exe                    876 Services                   0     19,348 K
svchost.exe                    936 Services                   0      9,088 K
dwm.exe                         80 Console                    1     35,604 K
svchost.exe                    524 Services                   0     20,352 K
svchost.exe                    840 Services                   0     46,836 K
svchost.exe                    920 Services                   0     20,672 K
svchost.exe                   1052 Services                   0     27,224 K
svchost.exe                   1148 Services                   0     12,024 K
nvvsvc.exe                    1244 Services                   0      8,908 K
svchost.exe                   1340 Services                   0     30,244 K
nvxdsync.exe                  1420 Console                    1     19,912 K
nvvsvc.exe                    1428 Console                    1     12,832 K
WUDFHost.exe                  1708 Services                   0      7,180 K
spoolsv.exe                   1996 Services                   0     16,568 K
svchost.exe                   1644 Services                   0     32,284 K
Cloud.Backup.Scheduler.ex     2144 Services                   0     52,816 K
SASCore64.exe                 2152 Services                   0      5,128 K
DfSdkS64.exe                  2160 Services                   0      1,420 K
svchost.exe                   2168 Services                   0     30,280 K
mbae-svc.exe                  2304 Services                   0     13,668 K
SecUPDUtilSvc.exe             2356 Services                   0      3,564 K
svchost.exe                   2364 Services                   0     17,572 K
svchost.exe                   2404 Services                   0     10,664 K
dasHost.exe                   2412 Services                   0     12,828 K
sqlwriter.exe                 2420 Services                   0      7,284 K
MsMpEng.exe                   2440 Services                   0    108,312 K
mbae64.exe                    2960 Services                   0      5,240 K
conhost.exe                   2984 Services                   0      3,588 K
sihost.exe                    3268 Console                    1     21,180 K
NisSrv.exe                    3892 Services                   0      7,772 K
taskhostw.exe                 3956 Console                    1     15,084 K
explorer.exe                  4088 Console                    1    108,720 K
MSIAfterburner.exe            3344 Console                    1      3,468 K
RemindersServer.exe           4104 Console                    1     19,804 K
RuntimeBroker.exe             4852 Console                    1     36,008 K
ShellExperienceHost.exe       4936 Console                    1     65,752 K
SearchUI.exe                  5500 Console                    1     91,488 K
SearchIndexer.exe             5820 Services                   0     34,244 K
GoogleCrashHandler.exe        2940 Services                   0      1,052 K
SettingSyncHost.exe           3008 Console                    1      3,364 K
GoogleCrashHandler64.exe      5616 Services                   0        936 K
RTSS.exe                      3496 Console                    1      1,576 K
EncoderServer.exe             5932 Console                    1        800 K
RTSSHooksLoader64.exe         5008 Console                    1      1,020 K
svchost.exe                   4880 Console                    1     25,068 K
RAVBg64.exe                   5368 Console                    1     13,288 K
LCore.exe                     6308 Console                    1     54,264 K
LCDClock.exe                  6728 Console                    1     17,392 K
googledrivesync.exe           6812 Console                    1      3,884 K
googledrivesync.exe           6868 Console                    1     85,152 K
OfficeClickToRun.exe          6888 Services                   0     50,644 K
Ditto.exe                     6960 Console                    1     18,804 K
Steam.exe                     7012 Console                    1     81,024 K
Dropbox.exe                   7092 Console                    1    146,508 K
steamwebhelper.exe            7120 Console                    1     27,636 K
SteamService.exe              1384 Services                   0      9,776 K
mbae.exe                      6520 Console                    1     13,848 K
jusched.exe                   5252 Console                    1     11,656 K
fontdrvhost.exe               6468 Console                    1      2,808 K
svchost.exe                   5056 Services                   0      5,244 K
ApplicationFrameHost.exe      1472 Console                    1     28,272 K
Microsoft.Msn.Weather.exe     7744 Console                    1     48,116 K
steamwebhelper.exe            4296 Console                    1     23,740 K
rabbit.exe                    7816 Console                    1    511,728 K
audiodg.exe                   7088 Services                   0     19,388 K
GameOverlayUI.exe             7944 Console                    1     44,152 K
taskhostw.exe                  480 Console                    1     14,952 K
WUDFHost.exe                  2608 Services                   0      9,584 K
chrome.exe                    4788 Console                    1    184,332 K
chrome.exe                    2760 Console                    1      9,464 K
chrome.exe                    5684 Console                    1     76,656 K
chrome.exe                    4868 Console                    1    177,288 K
chrome.exe                    2020 Console                    1     94,080 K
chrome.exe                    1252 Console                    1     37,836 K
chrome.exe                    4148 Console                    1     41,784 K
chrome.exe                    4756 Console                    1     44,064 K
chrome.exe                    5724 Console                    1    147,248 K
chrome.exe                    4460 Console                    1     39,880 K
chrome.exe                    6428 Console                    1     45,064 K
chrome.exe                    6324 Console                    1    130,088 K
chrome.exe                    3424 Console                    1     42,000 K
chrome.exe                    7240 Console                    1    106,456 K
chrome.exe                    8044 Console                    1    140,864 K
cmd.exe                       5588 Console                    1      4,120 K
conhost.exe                   7652 Console                    1     16,352 K
chrome.exe                    4912 Console                    1    134,092 K
chrome.exe                    6896 Console                    1     95,312 K
chrome.exe                    7248 Console                    1    102,344 K
chrome.exe                    7484 Console                    1     95,820 K
TrustedInstaller.exe          6924 Services                   0      6,504 K
TiWorker.exe                   656 Services                   0     10,436 K
tasklist.exe                  7688 Console                    1      7,336 K
WmiPrvSE.exe                  2712 Services                   0      8,052 K

 

It's a challenge to discuss an issue with someone who misunderstands points in order to create conundrums out of thin air. No one but you said "changing is not needed". I've been very clear in my points with quite a degree of repetition.

I'm not going to have a 6gb pagefile (larger than the combined size of all installed applications on my machine) when I can set it to a fraction of that without a direct change in performance. Whether or not I can afford this waste is irrelevant to me, although for some machines (e.g. those with 64gb SSDs) then it's a better choice to leave that space free for over-provisioning.

 

Not sure to understand: do you use 10 GB of RAM using your pc ??

 

This is true. You still must ensure compatibility and such. But the fact it, it is an easier user task and in relation to other hardware upgrades, is typically less expensive too - even if you have to remove the old RAM and install all new RAM (often the case with motherboards with only 2 RAM slots). Also, adding RAM rarely requires upgrading the power supply whiles upgrading the graphics card, for example, often requires a bigger supply.

That said, I always try to fully populate my motherboards with more RAM than I think I will need during the initial build. Then I don't have to worry about it later or have to retire smaller RAM to my storage shelf and buy all new. Also, if you decide 3 or 4 years down the road to upgrade, the RAM you want may be harder to find and thus even more expense. It will not be too long before DDR3 is no longer in production, supplanted by the emerging DDR4 standards. Just like DDR2 and DDR today, finding old, compatible RAM is hard and expensive.

So a little extra investment today can save you money down the road.

As I said long ago in this thread, 16Gb is probably overkill for my needs, but that's what I put in this system when I built it because I know I will never need to add more. And for sure, I know of no scenario where more RAM degrades performance. However, because my board has 4 slots and I went with 2 x 8Gb, I can double my RAM with little problems. If I had gone 4 x 4Gb, I would have to retire some to add more. Not cost effective over the long run.

 

How about most of the world? By far, most experts say to just leave it alone! In fact, no one here has shown where any expert is advising normal users of W10, W8, or W7 to change from the defaults.

The challenge is trying to understand someone who keeps contradicting himself who instead of debating the facts on a mature level, resorts to personal insults.

You said post #64 above!

So I ask again, if it does not improve performance, why change it?

 

It seems as though every single one of these debates ends up in you advocating that 'for the average user...'. Anyone who browses this forum isn't an average user. You can make generalizations all you want about the average user, but we aren't average users.

There are other factors than just performance. Maybe you want the security benefit of being sure that your encrypted container key isn't being saved to disk. Maybe you want to save the disk space. Maybe you have enough extra ram that disabling the page file is perfectly viable and WILL increase speed. You cannot deny that accessing data from RAM is faster than accessing from disk. That said, I'm aware that there are many optimizations that Windows has that make a case where accessing something that Windows has not yet loaded into ram from the pagefile more minimal, but it isn't perfect.

I personally have no page file on my desktop (16Gb RAM). I've been running that way for about a year now with no problems. The most ram usage I've seen was about 12Gb and that was with 3 games open (Counter Strike Source, Heroes of Newerth, and Grand Theft Auto 4) and in a skype call along with multiple tabs open in Chrome. Between disabling the page file and disabling hibernate, my install size is about 11Gb (though it has become about 20Gb since installing Windows Updates).

 

You don't know that. Most are not advanced users, and many are complete novices. And more importantly, when giving advice in an "open forum" it should not be assumed all readers are advanced. And any case being an advanced user, or even an expert on computers, that IN NO WAY means they have expertise in memory management to know how to properly analyze commit rates.

As for your what ifs? Security? If someone with that sort of forensic expertise to read the data in your PF has that sort of physical access to your computer, you have more problems than an abandoned PF containing something you don't want others to see!

That is such an extreme exception, it borders on silly. Exceptions do not make the rule for the many.

I already agreed that disk space may be a factor, if 6G of disk space really matters, you need to uninstall programs or buy more disk space if you want it done right.

No it won't! These are the myths that keep getting perpetuated that cause folks to follow unwise advise. Only in very rare circumstances for a very few people does this ever happen.

No, but you are then assuming the data on disk is of a higher priority. It does not work that way. If the OS needs data to be cached, it will rack and stack the higher priority data in RAM way before thinking about putting it in the PF. So that is not a valid reason.

I did not say running without a PF when you have large quantities of RAM will cause problems (though there is evidence it might). I said if you have one you can get dump files and Windows will cache data it may need in the PF - lower priority data for faster retrieval later.

Understanding the Windows Pagefile and Why You Shouldn't Disable It.

Ask Leo! Can I delete pagefile.sys? What is it?

I want to reiterate what I have been saying - if you know what you are doing, fine. Manage the PF yourself. But (1) it is not a set and forget thing, the commit rates need regular reassessments and the PF settings adjusted as the user changes. And (2) do not advise others to set a user managed PF unless you know they are fully capable of properly analyzing their commit rates and have the discipline to keep regular tabs on that. And (3) as others have noted too, because there are no performance advantages (except for a tiny few under very limited circumstances), just leave it - Windows knows how to manage this well. Microsoft has been at this for over 20 years. They are not dummies, in spite of what some may think. Their marketing department may be full of dummies, but not the folks doing the development.

 

"The herd is always wrong."
"If everyone in the room is thinking the same thing, then somebody's not thinking."

"It is well known", "The experts say", "So-and-so says" are non-arguments as they are not quantitative (these are specific, observable, measurable technical things here; technical details aren't subordinate to opinion).

Anyone browsing a technical forum asking about i5 vs. i7 is not an average/ordinary/herd user. Friends don't let friends settle for ordinary anyway. Blanket generalities have no place here.

It does when a platter drive is involved. The only question is "by how much?" and that depends on what the user is doing.

You don't have to keep continual statistics on an endpoint's paging in order to tune it--you're using sever performance logic.

You've made your point: "They" say not to change it. Others provided technical explanations that disagree with your point/the relevance of it ("They" do not know my specific application) to those who want to know and understand their computer better. I find it strange you spend a lot of discourse to discourage pursuing technical knowledge on a technical forum.

 

The fact that someone actually takes the time to research their security setup instead of just buying the current AV BestBuy is pushing makes them lean towards the advanced side, in my opinion

Ah yes. It's impossible to be completely secure so let's just not bother securing anything.

You say it won't then you say it will in certain circumstances. Which is it? Also, that's more or less exactly what I said when I noted that Windows' prediction works well the majority of time but isn't perfect. Assuming you have enough ram, it can only improve performance when you disable your pagefile, no matter how small of a gain it is.

If you wanted to immediately dismiss any reasons that don't support your viewpoint on this, why did you ask?

I never once told anyone to set a user managed pagefile. I simply said 'this is what I do and it works fine for me.'

This x100000.

 

Bullfeathers! You see these type questions all the time. If someone comes in and says help me decide between the i5-4670K Haswell 3.4Ghz and the i7-4770 Haswell 3.4GHz, then I would agree.

It is clear you guys are convinced you are right. So be it. And you may be right - for you!

But don't assume what is right for you is right for others without analyzing their commit rates over an extended period of time. That is just wrong. And don't assume someone with expertise on the i5 or i7 is an expert in memory management or page file configuration. That is also wrong. And finally, exceptions don't make the rule. Only for a very few under rare circumstances does not in anyway suggest it is right for everyone.

 

I don't discourage it all. But I do encourage verifying because there are many folks who have swapped out a PSU, added RAM, assembled a computer or did something that did not break their system and suddenly they think they are experts in all areas of IT.

My goal is for OPs to have ALL the information they need to make informed decisions for them. Telling someone they don't need a PF because they have 8G or 16Gb of RAM installed without analyzing their commit rate, or simply because they never exceeded 12Gb on their system with their programs suggesting that then applies to anyone with 16Gb is offering informative advice. Besides you cannot tell just by RAM usage. You have to look at commit and peak rates - which are not even selected for view by default.

And again, being an expert in security, or knowing the different between an i7 and i5, in no way suggests they are experts in memory management, or analyzing commit rates or setting up user defined Page Files. And none of you have shown us anything that says disabling the PF or setting a user defined PF is what people should be doing to improve performance, or security.

 

You have yet to provide a way for someone to log their usage in order to analyze it or even mention what to look for so as far as I can tell, that's not what you're doing.

Never did I once tell anyone what to set their pagefile to. In fact, I never even gave any recommendations on what to set it to.

Yea, you're not wrong here. So how is someone supposed to become an expert or even just more educated in the area if we are told to just accept what the 'expert's' have already determined is best for us? This automatically assumes we both, use our computer the same way as the average person, and have the same hardware as the average person. You're telling us to not give out all this information because we can't assume we know what's best for a user and yet you're advocating Microsoft knows what's best for all users.

 

Of course not! I recommend people leave it alone! If someone starts a new thread asking how to analyze and set their own PF, then I would help them.

I was not accusing you of doing so. But when some posters see regulars say that is what they do, it can be inferred that they should do it too. And when there no mention of how to determine the posters actual requirements, that is a problem because every user and computer is unique.

Good point. It is done by doing a lot of research. Not just listening to one person telling how they do it. Look for the bonafied experts, like Mark Russinovich who went into great detail how to analyze memory utilization.

BUT YOU DIDN'T! Where did you show how to use commit rates and peak rates to analyze needs. Where have any of you said that reanalyzing whenever your computing habits change is needed? You didn't! None of you have. Only after I started to raise a fuss did Mark's Blog appear. And no one has yet to show where anything but Microsoft managed is better with W10, W8, or W7.

I am advocating that they know what is best for the vast majority of users AND I am advocating they know how to program Windows to do the analysis too, then set the PF for that computer and how that computer is being used. And they know how to program Windows regularly reanalyze those needs and adjust the settings as needed.

You are implying that Microsoft as set a one-size-fits-all value. That is not true at all. If they had, then we would not be having this discussion because I would be 100% behind you - as long as you were telling people how to analyze their needs.

These modern operating systems are not XP, 98, 95. They are intelligent programs that know how to optimize themselves based on how the user uses the computer. Again, don't hold the developers accountable for what the Microsoft marketing or MS execs do. Give the coders some credit. They are the ones doing the grunt work and they do it well.

 

• You haven't encouraged verifying; you encourage "Leave it alone! Because the experts say..."
• We aren't talking about "many folks"; there are a handful of specific individuals here who appear to know of what they speak. I can't help but see this as a thinly veiled ad hominem argument

I didn't say to eliminate the pagefile based on installed RAM and I don't remember anyone else stating such.

Nobody has said this either. Additionally, one doesn't have to be a "memory management expert" (whatever your definition of that is) to effectively tune one's PC much like one does not have to be an expert auto engineer to tune one's car. Finally, I'd argue said experts are less useful when their prime directive to to "save users from themselves". Nothing ventured, nothing gained...

You're the only one asking for proof. We're offering advice. Take it or leave it; it's your PC. Mine already works swimmingly.

 

I believe I've already explained my reasoning in the post you just replied to, as well as congruently in each of my posts preceding that one.

I've given no general advice, rather have merely discussed my own approach to managing the pagefile and how this approach works for my needs. Other people are responsible for how they apply this information to their own contexts. For whatever reason you've taken exception to this.

You've already explained that you don't understand how paging works, so therefore you believe that I and others should leave it alone and trust the defaults from Microsoft. I listened respectfully - but since you've given me no technical reason for managing the pagefiles on my machines differently, then I've chosen to dismiss your advice. I'm no longer interested in discussing this with you.

 

What? Where?

Actually, I know how it works very well. I'm no spring chicken and if you follow the link in my sig you will see I've got some experience supporting Windows based computers. And ONCE AGAIN I am saying if you know how to properly analyze your commit and peak rates (not just RAM "usage") AND you have the discipline to regularly reanalyze and readjust as necessary, then go for it, if that is what you wish.

But if you don't know how to analyze those values properly and you don't have the discipline to regularly reanalyze those values, just let Windows do it. As noted by me and others, there is nothing to suggest user managed page files (or no page files) will improve performance.

And lastly - and my biggest concern as a technical advisor and forum helper - when giving advice to others, "we" must not assume posters (and everyone else reading) have the necessary expertise in memory management to properly analyze their commit and peak rates (and NOT just RAM "usage" as some here have indicated) - even if they have demonstrated advanced expertise in other areas.

Setting a user managed PF size is NOT a set-and-forget thing! Yet that is typically what happens when users change those setting from the defaults. Even with users (even advanced users) who know how to properly analyze the applicable rates.

Now if you disagree with any of that, then I am sorry. But until you can show what I have said is wrong, I stand by it.

 

Here ya go, Bill: https://www.wilderssecurity.com/threads/memory-optimizing-bad-for-ssd.380192/

I think we'd agree.

 

Yeah, snake oil is right. Thanks! I needed the laugh.

 

I'll make it easier for you, here's a comparison of Performance with 4 of the latest Skylake Processors

https://www.ple.com.au/ViewItem.aspx?InventoryItemId=620470
https://www.ple.com.au/ViewItem.aspx?InventoryItemId=620291
https://www.ple.com.au/ViewItem.aspx?InventoryItemId=620469
https://www.ple.com.au/ViewItem.aspx?InventoryItemId=620290

The K versions are preferable to Gamers, as they can open up the speeds by overclocking the system

 

When you stated that "the fact of the matter is, unless you are Mark Russinovich..." then we're not as qualified as Windows to manage our own pagefiles. I just followed your logic to its conclusions, since I gathered that you're not Mark Russinovich

I agree and since I was the first to point that there's no measurable difference in performance, then I'm wondering why you keep debunking "performance improvement" in each post as though it was ever my goal? That's a straw-man argument, yet another logical fallacy you seem to make.

Appropriately set pagefiles give the same performance as the Windows recommended size, but with a fraction of the space permanently allocated. So yes, I manage the pagefile size better than Windows - and happily rely on Windows to appropriately adjust the size dynamically.

Then simply provide that information to promote understanding, given this is a technical forum.

Really though, the sky won't fall for an inappropriately set pagefile: Windows will inform of low resources, system performance will degrade, and the application might crash. This will prompt the user to consider allowing a higher range for a dynamic pagefile. The user will learn from their mistake, no big deal.

You may have forgotten, but this whole discussion was because you specifically disagreed with the way I had managed my own pagefiles, and advised me to leave the settings for Windows to manage.

For this reason the onus was always on you to provide a technical justification for this, which you never did.

 

This discussion was between me and several of you who have changed your settings from the defaults without providing and technical reasons for doing so - and it was not started by you!

And sorry, but you cannot seem to follow the logic. Saying I am not Mark R is not saying I don't understand PFs. I do, very well frankly. And because I know very well how virtual memory in Windows is managed, I let Windows manage it on all my computers and all the computers I am responsible for because,

1. I know Windows can manage it well,
2. I have lots of free disk space,
3. I have no desire to keep reanalyzing my virtual memory usage, and
4. There is NO performance advantage.

My objection with you and others, as I have repeatedly said, is your implying what is right for you is right for others when (1) you have not analyzed their commit and peak rates and (2) you will not be there to frequently re-analyze their commit and peak rates as their computing tasks and habits change and (3) you do so without telling folks it is not a set-and-forget thing.

No big deal Tell that to the person who runs out of memory in the middle of his project and the OS crashes. It seem apparent that to you, your computer is just a toy. To MANY others, it is an important tool that has become an integral part of their lives and they are very enthused to keep it running properly. You cannot assume everyone who comes to these forums has the same mindset as you.

I provided links to multiple experts showing why most users should just leave it alone. But you ignored them.

Want more? Okay.
Computer Hope: Is it a good idea to change my Microsoft Windows page file size? No.
HowToGeek: What is the Windows Page File and Should You Disable It?

You changed from the defaults. You (none of you) provided zero supporting evidence from ANY expert that says, "users should manage their own virtual memory". None!

I am asking for ANY supporting evidence that your "advice" is sound advice! And you have given none!

You (speaking to all who believe they are smarter than Windows) need to justify (but not here - we've gone OT enough) with technical data, how you are smarter than the PhDs at Microsoft and all the experts I quoted that your way is better. And you should start by justifying to yourself without bias or pride stepping in the way, that changing from the default IS providing positive results for your computer. "No discernable negative results" is not a valid reason to change from the defaults.

Now blacknight started this thread to discuss i5 and i7 CPUs. We have run this OT long enough. I have backed up my comments with links to recognized experts. If you don't believe any of them, including Mark Russinovich who says you MUST analyze your commit charges and peak values to do it right, then as I said above, I am sorry. And until you can show that I, Mark R, and all the other experts I quoted are wrong, I stand by it.