Build Advice New build questions ?

Page 15 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
progress update, I ran into problems with the 2nd Windows 11 install, its just an experiment so I could just reformat it. but I decided to try and fix it in order to learn how to deal with this.

where that Windows 11 wouldnt boot at all. but the other Windows versions would boot. but I had to go via the Linux boot to reach them as that 11 was the default boot and was bricked up!

after a lot of failed experiments, I got Windows 11 to boot, via booting with the Windows 11 boot bluray, and selecting R for repair at the start, eventually reaching "startup repair", which then allows me to select any Windows install. now the identical looking options from when you boot to Windows, where it shows the various Windows versions, and then another option which you then click, and navigate from there, that
only allows you to repair the current default version, so is no use. it skips a step which the installation bluray gives when you click R for repair. they should have a V option also for vendetta!

now after this, Windows 11 did boot, but I then found that Notepad no longer worked!

to deal with that, I had to run Setup.exe from the Windows 11 bluray, from the problem Windows 11 desktop, and then select an option that preserves data and programs. that took hours to complete, I could tell it hadnt frozen up by looking at properties for C:\ where I could see the bytesize was continually growing. eg at one point it was stuck at 46% for ages, but the bytesize was continually increasing. so that is just an ineffective progress indicator rather than no progress.

now long before the repair finished, Notepad started functioning!

much later, I went to bed, and at some point I thought the repair had crashed. but in fact the screensaver had activated!


most of the tips online are only for fixing a Windows install that boots to the desktop, when it doesnt boot at all, the above MO works. I am going to try this later to fix the 2nd Windows 10 install, as that one had problems, and I will photograph every step, and thus will test if it works as it will need the Windows 10 bluray which might not be the same. I will test both MOs that worked for Windows 11. in particular I want to see if the current installs of software are preserved, after each of the 2 stages.

this way when I do the permanent installs, I will know how to get out of jeopardy.

its lucky I decided to do these temporary installs, as I have gotten into major jeopardy en route and found how to sneak around the jeopardy, sometimes too late!

I learnt also how to make icons remain at the top of the desktop, which is to right click on the desktop and select View, and deselect the auto arrange and align icon options. then I can put the more useful icons at the top edge.


I have found for example that if I copy the mobo flash drive to a directory, that the mobo installation mostly instals, but one installation fails. I have now copied all that stuff to a bluray at the top level, and will see whether it all installs from the bluray.


when I installed Windows I gave all passwords as nothing, ie the empty string, that way any request for password I just press return!

I have just now migrated everything from the other 2T M.2 to the 4T M.2, as I have a futher experiment later, which is to put at least one windows install on that M.2, to see what happens when Windows is installed on more than one drive. I will also leave a dummy partition in place on that M.2 to see what happens to it if I install Windows.

Before I install the permanent installs, I will do a compressed sector backup of the entire 2T M.2 drive with all the OSes to the 4T drive. then reformat and begin the permanent installs.

with the Linux Mint 21.1 install, which includes a lot of stuff, I can only see the sizes of the partitions from Linux, Windows doesnt show how much in use. I found that only 19G is in use. so for the 2nd Linux Mint install, which is the next work, I will give 100G or less, eg maybe 50G. with the permanent installs, I can thus set say 512G for win10, win11, 100G for Linux, and then still have enough room for maybe a future Win12, and Win13, and meanwhile maybe some scratch installs of Win10 and Win11.

before reformatting, I also want to test what happens if I just delete one of the Windows install partitions, will the other instals work?

if that works, I will then delete a Linux installation partition, and see what happens.

and eg what happens to the Linux loader called Ubuntu.

if everything goes to pot, hopefully it doesnt matter, as I can maybe reformat the entire disk, as long as I can still arrange for the Windows bluray to boot.



That specific one is Thermaltake Pacific W7 Plus CPU water block,
specs: https://www.thermaltake.com/pacific-w7-plus-cpu-water-block.html

It is just a small part of custom water cooling loop. Once completed, there will be tubing going to and from CPU water block.
It would look like so:

Eghdl2-UMAYPmMj


123727503_4298103666882941_8698418553509091720_n.jpg



G.Skill Flare X5 (2x 16GB) 5600 Mhz kit is the freebie one he showcases briefly.
Specs: https://www.gskill.com/product/165/396/1661842219/F5-5600J3036D16GX2-FX5

what are your arguments for and or against this specific product?

does the product supply the water, or you have to supply that?

some years ago I bought a water distiller, its very useful eg for steam irons, as tap water will soon block up a steam iron. someone gave me a blocked up steam iron, and using distilled water, it soon cleared up!

I use the distilled water with lint free cloth from ebay also for cleaning monitors and glass eg mirrors. as it leaves no residue.

For cloning Samsung has different dedicated software: Samsung Data Migration Tool. It's one of the best out there (if not the best).
I have to study that sometime, but I need to do a lot of other things that I need to do sometime first!

at the moment I use the Linux shell for cloning, namely the dd command, which treats disk hardware as if it were a file. where an SSD will be something like /dev/sda but the M.2 drives have more complicated names, but also of the form /dev/??????

but you mustnt clone the current Linux disk, to do that, I run the Linux try without installing DVD, eg when I clone the entire drive with OSes I will do that from the try without installing DVD, to avoid coherency problems. eventually I might try the backup software which USAFRet recommended as that is probably much more efficient. I have to regulate the new things I do, to introduce one at a time gradually.



Have you actually looked if you can define any amount, or does UEFi give you some options (e.g 6-8) in different sizes?
Since usually, there are predefined options and one just can't allocate all available RAM to iGPU.

ok, I did check, with Games Mode, you cant set the amount, where UMA Frame Buffer Size is ghosted:

http://www.directemails.info/tom/graphics_card/UEFI_Game_Mode_no_options.jpg

if you press + repeatedly, each is ghosted till you reach Force,

http://www.directemails.info/tom/graphics_card/UEFI_force.jpg

then if I click the UMA Frame Buffer Size I get this menu:

http://www.directemails.info/tom/graphics_card/UEFI_force_options.jpg

and then after selecting 16G I get this:

http://www.directemails.info/tom/graphics_card/UEFI_force_16G.jpg

note that increasing things doesnt necessarily make them better, eg larger memory can be worse for caches, so with such things, one needs to benchmark what allegedly is better, to confirm or deny if it really is better.

there are different forms of benchmarking, the one is to use a benchmark program eg unigine, and with the Amiga I think there was one called Sysinfo, which would place your system's ratings against some standard ones eg factory supplied Amiga 1200, Amiga 4000 etc. But the other is to do personal benchmarking, which is where you try something ad hoc. there are 2 forms of personal benchmarking:

1. doing a real task, eg with disk drives to do a sector backup of a drive.
2. doing a contrived task, eg an assortment of dummy files of different sizes, or eg to do a file by file copy of the Windows system partition, usually C:\, where its a real directory, but where the task is contrived as doing a file level backup of Windows is of limited value. I would categorise this latter one as less contrived.

one of the differences between older people and younger people, is older people tend to consider realistic scenarios, whereas younger people get caught up in unrealistic contrived scenarios. the older person will look at a circumstance, and consider for each part a realistic option. eg if you say to an older person "what would you do if someone gave you $10m and a jumbo jet", the older person will say "someone wont give me that, it is unrealistic". a realistic scenario is someone gives them $100 and an unrepairable bicycle.


for an assortment of dummy files, it would be best to eg fill these with repeatable random numbers, eg use a pseudo random number generator with seed. where the same seed value generates the same sequence. that way you can verify the files copied correctly and also it wrecks any compression, as random numbers generally wont compress. compression only works where there is order which gets factored out to chaos. a well compressed file should look like total chaos. With Windows you can make a partition increase capacity via compression. it would need benchmarking to determine if this is faster than uncompressed for a specific disk arrangement. its all a question of whether the latency of compressing is faster than the latency of writing a bigger file! eg compressed sector backups to the M.2 drive are hours faster than uncompressed on this machine. but with my Amiga 1200, compressed copying was much slower than uncompressed.

or you can use realistic files, such as all the photos of your camera, which for me is some gigabytes as I photograph in max resolution.


to benchmark different systems across the internet, the benchmarking needs to be standardised, which is then things like unigine. but any test at all yields information. its like to understand a country, you could follow some protocol, to spend a year touring the country. but even if you just never go beyond the airport departure lounge on a transit flight, that also yields information.

with a benchmark, you need to be careful about copying the same file repeatedly, as the source file may well get cached, and if small, the repeated destination files could also get cached, where the target disk is unchanged. doing a real task is more useful as that gives a realistic indication.

with that fake 64T drive, when I copied files, they'd be there, but in fact only in the Linux or Windows disk caches! if I rebooted, the files vanished!

when Apple was PPC based, Adobe would optimise say jpeg loading, and show this was faster than Windows, but the thing is most people arent even aware that such things take time. they load the jpeg, and think the computer is busy, they expect everything to take ages, and Windows by default hides the jpeg suffix, so for them its just a photo and Adobe's benchmarks are over their heads! Eventually Adobe made their wares available for Windows, and Apple became Intel based.


where he got those DDR5's for free,

I didnt catch what capacity those freebie ones were. with time the higher capacity ones will become cheaper, and noone will buy the lower capacity ones!

G.Skill Flare X5 (2x 16GB) 5600 Mhz kit is the freebie one he showcases briefly.
Specs: https://www.gskill.com/product/165/396/1661842219/F5-5600J3036D16GX2-FX5

that is then the problem, that if you are going to buy some upmarket DDR5 machine, you dont want to then ruin the dream with 2 x 16GB!

because 16GB is more an upmarket machine of long ago!

but the 2 x 16GB may be alright if its for free, as you can then delay buying a larger amount of memory.
for ordinary uses it will be fine, but software will start using too much memory, eg a bluray burner will store the entire bluray in memory and will want 25GB just for that, and then you'll need maybe 8G for all the software, which then is 33GB. thus probably 64GB is the minimum to consider. the bluray burner software I got by Cyberlink, copies all the source files to a temporary directory in C:\, which is a bit inefficient. I am sure the Roxio software for my 2010 PC just left the source files where they are, but I am not 100% sure.

As i said, 2 GB is plenty for iGPU. Your dGPU (RTX 4060) has 8 GB of dedicated VRAM, separate from the rest of the system.

but the 16GB mentioned earlier, is that entirely for the iGPU?

the 4060 has an unchangeable onboard 8GB?



This was discussed and solution was found. Earlier in this topic:

yes, the sounds being produced were unbearable, and setting voltage to 0 was the only satisfactory option.


I wouldn't neccessarily want temps to go up to 65C. But i didn't make a bespoke profile, i just edited slightly the profile i was already happy with.

ah yes, but you dont know why I selected 65C!

I looked at the temperature graph when all fans at full power, and even with all at full power the graph would hover a small margin below 65C, so it is futile with my system to cool below 65C as it wont work.

so instead to start cooling only when it goes a small margin beyond the normal range, but because low voltages cause the fans to try to power up then fail, then try to power up then fail, repeatedly a bit like a fluorescent light where the starter is kaput, or a bit like in the film ELF where Santa Claus' sleigh wont take off because not enough people believe in santa, the sleigh is powered by people's belief in Santa!

I think those supplied profiles are nonsense, with my system they are unusable garbage.

just the CPU one I have left unchanged as that one works out of the box.



1x120mm Noctua NF-S12B Redux-1200 exhaust

120mm could be a way to mitigate the lack of space for the base fan for my system here, at the moment I mitigated by removing the lower 3.5" cage.

The reason that annoying sound exists probably has to do with the combination of case fans,
false



false
room size,
false
location of PC in the room,

false

and maybe the alignment of Jupiter as well.

false

I think these are all wild guesses!

the annoying sound is entirely what Aeacus said, which is that the voltage isnt enough to start the motor spinning, there will be some minimum force needed to spin them successfully,

this is a general problem that occurs in the real world, but the only example I can think of is the defunct fluorescent light starter, which leads to the light making these continual sounds as it tries to unsuccessfully start the light, where the light is "on the blink".

Aeacus probably can confirm or deny this, but probably both the fluorescent starter and the fans are relying on capacitors, and the problem is the capacitor charging up till it has enough charge to send some to the next part of the system, but it isnt enough, and it starts charging up again. that is my conjecture!

I dont know if the fluorescent starters rely on capacitors, but they look suspiciously like capacitors!
 

35below0

Respectable
Jan 3, 2024
1,727
744
2,090
the annoying sound is entirely what Aeacus said, which is that the voltage isnt enough to start the motor spinning, there will be some minimum force needed to spin them successfully,
Nay, i was talking about my fans. When they spin at 740rpm and there is a barely noticeable resonance. Dropping to just 720rpm makes it dissapear.
That is the reason i customized my fan curve.
I looked at the temperature graph when all fans at full power, and even with all at full power the graph would hover a small margin below 65C, so it is futile with my system to cool below 65C as it wont work.
Where is 65C being read? Which sensor reports 65C?
 
Last edited:

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
some very time consuming experiments, I backed up the entire 2T M.2 drive at M2_4 with the OSes so far, namely win10 ; win11 ; Linux Mint 21.1 ; win11 ; win10, doing a compressed sector backup to the 4T M.2 at M2_3, and it took about 2.75 hours. I was taking a risk that it could take say 8 hours!

Tue May 7 17:44:01 UTC 2024
3907029168+0 records in
3907029168+0 records out
2000398934016 bytes (2.0 TB, 1.8 TiB) copied, 13566.7 s, 147 MB/s
Tue May 7 20:30:10 UTC 2024
$

the compressed drive is 182.39GB.

then I installed a 2nd Linux Mint 21.1 to the same M.2, taking photos of each step, including steps which didnt work, where I get an error message. some steps dont allow one to return to the previous page, so I have to photo it even if its wrong. where I photo the error message.

I had to load GParted to comprehend the existing partitions, as the installers view of these is unintelligible, where it gives a different partition count from Windows for the same drive. I carefully made notes of what was where, and verified with sizes allocated when installing the various OSes, also GParted gives different numbers from the installer.

one puzzling thing is I have to say where the boot loader should be, where the default was on another M.2 disk, instead I changed that to this M.2 disk, rather than a partition on this disk, where I am kind of guessing. The risk is that Windows stops booting, but I had backed up the entire drive so can re-establish the earlier arrangement.

this time I used the normal Linux installation, and not the OEM one of last time.

once everything was complete, the machine ejects the install disk and shuts down.

I rebooted and it did go to a Linux boot screen, but no UEFI options. Now I dont know if this is because its a 2nd install, or if because the boot loader was to the wrong place, or if because it isnt the OEM install.

I have tested out the different options, it has an option for both Linux installs, advanced options for each, but these dont lead to the UEFI, but just load the 2 installs of LInux, and it has a windows option, eg I booted to do this message via the windows option of the Linux bootloader.

the windows options of different versions of Windows have all got the partition numbers changed, so it re-establishes those each time, and doesnt just use a fixed version, which is good.

I allocated just 50GB to this 2nd Linux install, as it is just an experiment, and I am keeping a bit of unused diskspace left for any further experiments.

before all this, I also tried the Windows repair via the windows 10 iso, of the 2nd Windows 10 install which earlier had malfunctioned, this worked but it said it couldnt repair it, possibly because the reset I did some days ago had already repaired it. I then did the update from the desktop of that windows install, and it was much faster than the earlier equivalent update of windows 11, which took many hours.

and afterwards, all the software installs were fine.

whilst waiting for the sector copying of the M.2, I experimented a bit with Firefox of the try without installing DVD, and found it can import bookmarks, but cannot import passwords. This is where MS Edge is ahead of Firefox, that you can export bookmarks, passwords and history, but I am not 100% sure if you can import history, but you can import passwords and bookmarks.

it seems to me blatant that you should be able to import + export, bookmarks + history + passwords.

the Linux Firefox says I can import those via synching, I refuse to synch such things, as my passwords and history will be out there on the internet, and I dont know who could be raking through it all.

I tried a passwords manager for Firefox, but it says I have to sign in, but that could mean my passwords are at some internet server, what's to stop them using my passwords and taking all my money?

I think it is ridiculous that password managing is done by 3rd party add ons, passwords are your money, they must think I am very stupid!

these experiments are to calibrate what you can and cannot do. there is a saying that "a stitch in time saves 9", which means if you fix a problem early, you have less work to do. these experiments are the stitches in time, making the mistakes on a set of scratch installs. the totality is very technical and fiddly.

I am considering doing a 3rd install of Linux tonight, this time an OEM one, to see if that gives me an UEFI option, as that will determine if the current lack of this is because it was a non OEM install. if it does give me an UEFI, it means the OEM install is better than the desktop install and my earlier opinion that the OEM install was a mistake was wrong.
 

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
Nay, i was talking about my fans. When they spin at 740rpm and there is a barely noticeable resonance. Dropping to just 720rpm makes it dissapear.

for your ones maybe, resonance may well be the case.

but the problem here wasnt resonance, but I think was some kind of ongoing clicking sound, its difficult to remember this kind of thing. it was also difficult to localise where it was coming from, I literally had to place my ear on the tower case like a stethoscope to eventually determine where it was from. even after disconnecting the fan, another fan was clicking. disconnecting all I then got silence.

originally I thought the sound was coming from the back panel IO socket zone, but in fact it wasnt.



Where is 65C being read? Which sensor reports 65C?

if you look eg at the graph on the right of this screenshot from many days ago:

http://www.directemails.info/tom/mobo/uefi/MSI_SnapShot_33.bmp

you will see it starts at maybe above 65C, there is a graph for each of those things on the upper left,
namely CPU Core, System, MOS, Chipset A, where Chipset A is selected.

that graph gradually moves leftwards,

at the time, I had a careful study of these, and decided 65C was a small margin above where the graph was keeping, so I make the voltage rise from 0 at 65C, and rise very rapidly to max allowable at 75C.

because at the lower voltages you get the noise problem.

this arrangement does work, where most of the time no fans other than the CPU ones and the PSU, but those are super silent and super well designed. where I have the silence I originally requested in the earlier topic.

with the copying of the M.2 drive, I had a sleep, and eventually I'd be woken up by the fans starting but I had no problem returning to sleep. as those fan sounds are like a strong breeze, once sleeping I dreamt the sound was a breeze in some palm trees by the sea!

what I dont like is perpetual noise, as that with time will impair hearing. it wont cause deafness, but it might cause partial deafness to some sounds.

for a more useful graph, I'd have to set all fans to max power, and let the graphs develop, and then screenshot, but I did all this at the time, and arrived at 65C being an optimal start temperature for this specific machine. if where you live is colder, you might well be able to go for a lower temperature.
 

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
looking at the photos for the OEM install of Linux, I find something unusual:

when I first did the install, there was no UEFI option in the boot screen:

http://www.directemails.info/tom/mobo/linux_boot_options.jpg

but later on in time, there is an UEFI option!

http://www.directemails.info/tom/graphics_card/how_to_get_UEFI.jpg

all I can think is that the later 2nd Windows 10 install must have caused the Linux boot screen to have an extra option!

so I may instead install a 3rd Windows 10 or 11, and see if that causes this.

this does show the value of photographing each step, as there is no way I would remember this!
 

Aeacus

Titan
Ambassador
what are your arguments for and or against this specific product?

does the product supply the water, or you have to supply that?
Overall, there are 4 types of consumer grade PC cooling. Actually two types, where each type has two branches;

Air
* Passive cooling - only heatsink, without a fan, relying on natural convection. E.g Arctic Alpine 12 Passive
Quite rare cooling method nowadays since modern PC hardware (CPUs, GPUs) output too much heat for passive cooling being effective. But there are some passive cooled PSUs, e.g Seasonic PRIME Fanless TX.
* Active cooling - heatsink with a fan. E.g Arctic Freezer 50 and your Be Quiet! Dark Rock Pro 5 also falls under here.
Most common cooling method around. Also the least maintenance with longest lifespan. Cheap too.

Liquid
* AIO (All-In-One) water coolers - pump, tubing, radiator, fans and liquid all come in one sealed package. E.g Arctic Liquid Freezer II 280 A-RGB
2nd most common cooling method nowadays. Same level of maintenance as with active air cooling, but considerably shorter lifespan (2-5 years), with more noise and far more expensive. Often double of what active air cooler would cost.
* Open-loop water cooling - pump, tubing, radiator, fans, fittings, reservoir and liquid all come as separate pieces, where end user has to assemble all of it, according to their needs. E.g EKWB, the go-to shop for all custom loop water cooling parts. Thermaltake CPU block is just one part of it.
This is the most "fancy" cooling method, since many different water cooling loops can be made, including custom colored liquid. This is very maintenance heavy, whereby every 6 months you should drain the loop and fill it with new liquid and after 1 year, you need to completely drain the loop and take it all apart, to clean the insides. And then putting it all back together, while ensuring that there aren't any leaks (if there are, you can fry your entire PC). It is also the most expensive out of the 4 listed, since with custom water cooling loops, parts are expensive.

--

As far as practical cooling goes, they all are equal, since in the end, they all are cooled by ambient air.
(Only the usage of cryogenic fluids, e.g LN2, doesn't depend on ambient air but those are used only in record breaking.)

Now, when to compare two most common cooling methods, active air cooling (air cooler) and all-in-one water cooler (AIO), it goes like this:

For equal cooling performance between AIOs and air coolers, rad (radiator) needs to be 240mm or 280mm. Smaller rads: 120mm and 140mm are almost always outperformed by mid-sized air coolers. Single slot rads are good in mini-ITX builds where you don't have enough CPU cooler clearance to install mid-sized CPU air cooler.

Here are the positive sides of both (air and AIO) CPU cooling methods;

Pros of air coolers:
less cost
less maintenance
less noise
far longer longevity
no leakage risks
doesn't take up case fan slots
additional cooling for the RAM
CPU cools down faster after heavy heat output

Pros of AIOs:
no RAM clearance issues*
no CPU clearance issues
CPU takes longer time to heat up during heavy heat output (about 30 mins)
* on some cases, top mounted rad can give RAM clearance issues

While how the CPU cooler looks inside the PC depends on a person. Some people prefer to see small AIO pump in the middle of their MoBo with tubing going to the rad while others prefer to see big heatsink with fans in the middle of their MoBo.

Main difference between AIO and air cooler is that with AIO, you'll get more noise at a higher cost while cooling performance remains the same.
Here's also one good article for you to read where former king of air coolers (Noctua NH-D15) was put against 5x high-end AIOs, including former king of AIOs (NZXT x61 Kraken),
link: http://www.relaxedtech.com/reviews/noctua/nh-d15-versus-closed-loop-liquid-coolers/1

Personally, i'd go with air coolers every day of the week. With same cooling performance, the pros of air coolers outweigh the pros of AIOs considerably. While, for me, the 3 main pros would be:
1. Less noise.
Since i like my PC to be quiet, i can't stand the loud noise AIO makes. Also, when air gets trapped inside the AIO (some AIOs are more prone to this than others), there's additional noise coming from inside the pump.
2. Longevity.
Cheaper AIOs usually last 2-3 years and high-end ones 4-5 years before you need to replace it. While with air coolers, their life expectancy is basically unlimited. Only thing that can go bad on an air cooler is the fan on it. If the fan dies, your CPU still has cooling in form of a big heatsink. Also, new 120mm or 140mm fan doesn't cost much and it's easy to replace one. While with AIOs, the main thing that usually goes bad is the pump itself. And when that happens, your CPU has no cooling whatsoever. Since you can't replace pump on an AIO, you need to buy whole new AIO to replace the old one out.
3. No leakage risks.
Since there's liquid circling inside the AIO, there is always a risk that your AIO can leak. While it's rare, it has happened. It's well known fact that liquids and electronics don't mix.

Passive air cooling

maxresdefault.jpg


Active air cooling

UNcL1iE.jpg


AIO water cooling
(Radiator mounted front, in push-pull configuration.)

cpu_and_gpu_aio_placement-ZJWcTn.True


Open loop water cooling (aka custom loop)
(Many open loops also include GPU in the loop. In this image, there is dual-GPU in the system, where both GPUs are also included into same loop.)

why_liquid_cooling-1-1.jpg
Here is tutorial video and build log for open loop water cooling for beginners, made by JayzTwoCents (watercooling guru):

View: https://www.youtube.com/watch?v=5ZnDESqJdyU


And of course, the video showcasing the maintenance part of open loop water cooling, required to do at least once a year, made by LinusTechTips:

View: https://www.youtube.com/watch?v=AfzwEF5yr3k


the 4060 has an unchangeable onboard 8GB?
Yes.

but the 16GB mentioned earlier, is that entirely for the iGPU?
16 GB is usual amount of total system memory, most PCs nowadays have. With 16 GB RAM total, at most i'd allocate 2GB of RAM for iGPU.

This much is actually said in your BIOS as well, whereby:
* total RAM: 4GB to 6GB - 512MB for iGPU.
* total RAM: 8GB to 16GB - 2GB for iGPU.
* total RAM: above 24GB - 4GB for iGPU.

Since you have 96GB of total RAM, not much point to allocate more than 4GB for iGPU. 4GB is still a lot for iGPU VRAM especially since it won't be used for heavy graphical workloads, where it would need that much VRAM.

In your case, i'd allocate 2 GB for iGPU. Or you can leave it into "Game mode", whereby system automatically allocates as much VRAM to iGPU as it needs, depending on a task.

one of the differences between older people and younger people, is older people tend to consider realistic scenarios, whereas younger people get caught up in unrealistic contrived scenarios. the older person will look at a circumstance, and consider for each part a realistic option.
Benchmark programs (Unigine, Cinebench, 3DMark, CrystalDiskMark, memtest86, Prime95, AIDA64 etc) are synthetic and on purpose, unrealistic.
Idea is, when PC is stable on very high, unrealistic workload, it for sure can handle real world scenarios, which are far less demanding on the hardware.

To give an example: In my apartment building, our water system (drinking water), is under the 3.0 BAR pressure, so that water can reach higher floors and have good flow pressure when opening the water tap. A month ago, we had pressure test made, where water pressure was increased by 0.5 BAR, to 3.5 BAR for 72h, to test if there would be any water pipe leakages. (Essentially to force out any potential leakages while everyone were alert, monitoring water pipes.)
As far as i know, test went smoothly and no leakages occurred. After 72h, the water pressure was put back to the normal of 3.0 BAR.

Essentially same is with PC hardware benchmarks. If PC is stable on high workload, all is good. But if PC throws errors or benchmark program crashes, system isn't stable and it's time to start looking what the issue might be. Since better to have issues pop up in controlled environment, than during real workload, where stability matters and instability issues are catastrophic.

but the 2 x 16GB may be alright if its for free, as you can then delay buying a larger amount of memory.
With free included hardware, there is always a catch.

E.g MoBo that comes with G.Skill RAM. Since MoBo has 4x RAM slots and one gets 2x RAM DIMMs for free, to upgrade RAM (e.g add more), one would likely buy the same brand RAM as what they already got. So, one would be buying G.Skill RAM.
Essentially, with this freebie, G.Skill is trying to catch people to buy their RAM in the future, rather than competitors (e.g Corsair, Kingston, Crucial, TeamGroup etc).

Aeacus probably can confirm or deny this, but probably both the fluorescent starter and the fans are relying on capacitors, and the problem is the capacitor charging up till it has enough charge to send some to the next part of the system, but it isnt enough, and it starts charging up again. that is my conjecture!
Well, i don't know the exact minute detail as of why fan motor only clicks when it has power, but not enough power to turn over, thus producing this "click" sound. But there are plenty of similar cases out there. One of them is the fluorescent lamp you said. Another one, perhaps more relatable is car battery.

Namely, car battery holds 12V charge and needs to have plenty of charge left, to start the car engine. Now, if battery is drained, all the lights in the car interior will still light up, including headlights (sometimes headlights can be dimmed as well) but there isn't enough charge to start the engine. Either there is nothing when you turn the ignition key, or you'll hear car starter to go "click-click-click" loudly. This is what i've personally experienced (had to call car tow service, where technician then charged my car battery, until i was able to start the engine).

it seems to me blatant that you should be able to import + export, bookmarks + history + passwords.
Importing/exporting bookmarks is common. History not so much, but no issue there either. But passwords import/export is something i wouldn't do, since the file where all the passwords are stored is high security risk. As far as i know the import/export file isn't encrypted and if somehow malware gets their hands to that file, all your passwords are out there (in the internet) instantly.

Speaking of internet security, i've been part of MalwareBytes Premium for some time now and they often come up with good articles and tips regarding online security (i get newsletter from them).
Once such latest endeavor, which i think is worth sharing with others, is their new digital footprint portal, which:
By entering an email address, anyone can discover what information of theirs is available on the dark web—and what information is collected from data brokers on the internet. From our safe online portal, you can view past breaches, exposed passwords, risks to personal identifiers including SSNs, national IDs and more.
Link: https://www.malwarebytes.com/digital-footprint

Note: Not sharing this as an advertisement of MalwareBytes, but instead as a free tool to check up on your internet security.
Did try it out myself and did find some troubling info about my e-mail addresses. Ended up cleaning my house, removing several unused accounts + made password changes.

I tried a passwords manager for Firefox, but it says I have to sign in, but that could mean my passwords are at some internet server, what's to stop them using my passwords and taking all my money?
Yes, there are password managed add-ons for Firefox. Though, i don't use any of them. Instead, i use the password feature built-in to the Firefox itself, where i have also set a primary (master) password. Meaning that any site i go to, that have usr and psw credentials saved in my Firefox, can not be used before i enter my primary password. I get a pop-up to insert my primary password once i reach that site.

Further reading about that feature: https://support.mozilla.org/en-US/kb/use-primary-password-protect-stored-logins
 

35below0

Respectable
Jan 3, 2024
1,727
744
2,090
Pros of AIOs:
no RAM clearance issues*
no CPU clearance issues
In addition to AIOs, some air tower coolers are offset away from the RAM. They are asymmetrical.
Like the Scythe Fuma 3 and Mugen 6 for example. Very nice design in my opinion.
https://www.scytheus.com/fuma-3
Well, i don't know the exact minute detail as of why fan motor only clicks when it has power, but not enough power to turn over, thus producing this "click" sound.
I've read an excellent article on this but unfortunately did not bookmark it. Wish i remembered the details well enough to explain.
Interestingly both DC and PWM fans have a minimum speed, but for different reasons.

These explain PWM and indirectly DC operation.
https://resources.pcb.cadence.com/b...rategies-for-cpu-cooling-and-case-ventilation
 

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
Personally, i'd go with air coolers every day of the week. With same cooling performance, the pros of air coolers outweigh the pros of AIOs considerably. While, for me, the 3 main pros would be:
1. Less noise.
Since i like my PC to be quiet, i can't stand the loud noise AIO makes. Also, when air gets trapped inside the AIO (some AIOs are more prone to this than others), there's additional noise coming from inside the pump.
2. Longevity.
Cheaper AIOs usually last 2-3 years and high-end ones 4-5 years before you need to replace it. While with air coolers, their life expectancy is basically unlimited. Only thing that can go bad on an air cooler is the fan on it. If the fan dies, your CPU still has cooling in form of a big heatsink. Also, new 120mm or 140mm fan doesn't cost much and it's easy to replace one. While with AIOs, the main thing that usually goes bad is the pump itself. And when that happens, your CPU has no cooling whatsoever. Since you can't replace pump on an AIO, you need to buy whole new AIO to replace the old one out.
3. No leakage risks.
Since there's liquid circling inside the AIO, there is always a risk that your AIO can leak. While it's rare, it has happened. It's well known fact that liquids and electronics don't mix.

what I would say then with this specific mobo, is that you need to install the graphics card first before the cooler, that way you can access the latch properly.

(Many open loops also include GPU in the loop. In this image, there is dual-GPU in the system, where both GPUs are also included into same loop.)

why_liquid_cooling-1-1.jpg

this would then explain the plumbing I can see in the graphics card. but maybe they should do 2 versions of the graphics card, the one with the plumbing for people who intend to use liquid cooling, and one without the plumbing for people who wont use that. then the graphics card for the latter group can be made much smaller.


Yes.


16 GB is usual amount of total system memory, most PCs nowadays have.
a 32 bit PC cannot go beyond 4GB, as 2^32 = 4 gigabytes, thus with 32 bit windows, no point in having more than 4GB. I think I got 8GB, which was just in case I installed 64 bit.

I think there is a similar problem with IDE that it has an upper limit.

now this means the market has been held back at 4GB for the 32 bit era, some people would buy 8GB not realising it couldnt be used! I bought it just in case I ran 64 bit XP. also I was programming the hardware directly in 64 bit, where my programs could go beyond 4GB.

but now with 64 bit, you really want to have at least 8GB, so 16GB is then the next step.
but the problem is 16GB and even 32GB are the first steps of a new era, so as the years go by at some point this will be too little. you can do everything you would normally want with say 8GB. but the problem is the firms out there artificially force you to use more than is needed.

remember that a 2 hour film on a DVD is less than 5GB, so 5GB is a ton of storage.

With 16 GB RAM total, at most i'd allocate 2GB of RAM for iGPU.

This much is actually said in your BIOS as well, whereby:
* total RAM: 4GB to 6GB - 512MB for iGPU.
* total RAM: 8GB to 16GB - 2GB for iGPU.
* total RAM: above 24GB - 4GB for iGPU.

Since you have 96GB of total RAM, not much point to allocate more than 4GB for iGPU. 4GB is still a lot for iGPU VRAM especially since it won't be used for heavy graphical workloads, where it would need that much VRAM.

In your case, i'd allocate 2 GB for iGPU. Or you can leave it into "Game mode", whereby system automatically allocates as much VRAM to iGPU as it needs, depending on a task.
that sounds a better MO, to allocate whatever is needed. good design is about avoiding fixed limits as much as possible, where people have invented MOs for doing this.


Benchmark programs (Unigine, Cinebench, 3DMark, CrystalDiskMark, memtest86, Prime95, AIDA64 etc) are synthetic and on purpose, unrealistic.
Idea is, when PC is stable on very high, unrealistic workload, it for sure can handle real world scenarios, which are far less demanding on the hardware.

that may be the idea, the reality might be something else!

its not an exact science! we are in the realm of correlative logic, where it is words such as "usually", "often", "most people". eg modern medicine is based on correlative logic, that a medicine has at least a 97% success rate, or something like that. there will always be people for whom a medicine doesnt work.

main problem is something cant be everything, so if the data is mainly huge files, then it isnt mainly small files, and vice versa.

eg I have some drives which are almost exclusively for sector backups of drives, so the drive consists of files which are maybe at least 50GB, but each is accompanied by a tiny text file which is the shell output of the commands creating the big file. where the small file might be a few 100 bytes.

and in fact this is the only place where speed really matters, because eg backing up the 2T M.2 to the 4T M.2 took approx 2 hours 45 minutes. but if I transfer photos from my camera to hard disk, that might take minutes at worst.


main thing about benchmark programs is they do the same test on all systems, where you can thus compare like with like.

for a fair test, you need to do the identical test on both systems, but with real data this isnt feasible as you dont want people on the internet experimenting with your personal files!


To give an example: In my apartment building, our water system (drinking water), is under the 3.0 BAR pressure, so that water can reach higher floors and have good flow pressure when opening the water tap. A month ago, we had pressure test made, where water pressure was increased by 0.5 BAR, to 3.5 BAR for 72h, to test if there would be any water pipe leakages. (Essentially to force out any potential leakages while everyone were alert, monitoring water pipes.)As far as i know, test went smoothly and no leakages occurred. After 72h, the water pressure was put back to the normal of 3.0 BAR.

why not send the mains to each foor unrestrained, and then limit the pressure per floor, where that then redistributes to each apartment? they then would just need to test the outer system.

you'd need to check more carefully the lowest floor as that would have the highest pressure,


Essentially same is with PC hardware benchmarks. If PC is stable on high workload, all is good.

with some things maybe, but with files, the test cant be everything.

it cant be an entire disk of tiny files of a few bytes, and at the same time be an entire disk of huge files.

in general you would have a spread with say x-axis being filesize, and y-axis being total amount of bytes from files of that size.

where you could do it as a bar chart with the x-axis in steps eg 100 bytes to 200 bytes as one step on the x-axis, and y-axis is say 1MB of data from files of that size.

the resulting graph is the spread. the problem is the spread cant be all things. it is either say a uniform spread, but say you have 1 file of size 1TB, then that means for the spread to be uniform you'd need 1TB for each other step, and there isnt enough storage to do that!

at the other extreme, for a uniform spread, if you put 1MB for each size range, the problem is you cant then have a file of 2MB. so the spread has an upper limit.

now you could try testing the hardware on just files of 100 to 200 bytes, then test again for files of just 200 bytes to 1 kilobyte, and then test again for files of just 1kilobyte to 500 kilobytes etc.

but the problem is you arent testing how the system fares when files are diverse sizes, because eg this can lead to problems of fragmentation. eg you fill the disk with files of 100 bytes. but you delete all except a small amount, but these ones fragment the space where you cant fit 1MB anywhere contiguously.

you are now testing how the filesystem is able to split files into scattered segments, and this kind of thing is a black art, where there is no right or wrong way, and there are many ways to go about things, with some ways much better than others.

so eg if you copy files to a freshly formatted disk, then that generally will be very efficient, because the filesystem can just shunt the data consecutively. but if the target disk has been in use for some months, it could be highly fragmented. and copying the same stuff now could be a lot slower, as the filesystem has to scour around for space, and has to fragment files into lots of parts.

the other option is the system isnt a uniform spread, but then what spread? there isnt any well defined non uniform spread which stands out. should the focus be on really tiny files? or on files of a few kilobytes, or files of a few 100 kilobytes, or files of 1 megabyte.

for any spread you have a median size, what should the median size be?

100 bytes?
1000 bytes?
10kilobytes?
100kilobytes?
1mb?
10mb?
100mb?
1gb?
10gb?

etc,

there are 2 questions here, the one is how fast is the hardware,
but the other is how fast is the software, because with a real filesystem, if you often delete stuff, then you run into the problem of fragmentation.

and eg if you test copy a load of files, then delete all, then test copy another load maybe of a different size, you arent testing for fragmentation. you'd need to only delete some of the files, but there are so many ways to do this. fragmentation is a difficult phenomenon to describe or control.

hardware and system software (eg filesystem part of the OS, firmware), generally will be better for some scenarios than others.

eg optical disks are generally best used for large reads and large writes. for small reads and writes they are a bit cumbersome. its partly because optical disks only rotate when accessed like with floppies.

with AmigaOS floppies, the hardware would always read an entire 360 degrees at a time, that enabled them to fit more data, 880K rather than I think 720K. because they just need 1 zone of synchronisation data. whereas with more localised access MSDOS disks, they need synchronisation data for each subset.

this also means the AmigaOS floppies are more efficient for larger files, and inefficient for tiny files eg say environment variables where the file can be say 2 bytes.

eg if you wrote large files to a newly formatted disk, the writing could continue in one uninterrupted sweeping action, or a few such. whereas with a HD, if a read or write is highly fragmented, the led will flicker like crazy!

Windows does give a defragmenting option, which visually shows it nicely.

now they could redesign optical to permanently rotate, this leads to faster access times as you dont have to wait for the disk to get up to speed. also with hard disks, they use parallel disks with internal readers, rather than external readers used for optical and floppies. parallel disks enables the read and write speeds to be multiplied up, eg 2 disks presumably means 2x the read and write speeds.


But if PC throws errors or benchmark program crashes, system isn't stable and it's time to start looking what the issue might be. Since better to have issues pop up in controlled environment, than during real workload, where stability matters and instability issues are catastrophic.
if it fails that is different, and we are in the zone of diagnostics and troubleshooting.

I am talking of where everything is functioning fine.

there are 2 levels of testing, the one is to test if the system works, with this you can stress test, eg create files of random sizes and random contents, and copy these, and then verify if they copied directly.

for verification it is best to power down the machine, even at the mains and reboot, because otherwise you might just be reading the caches. I ran into this problem with the fake 64T drive, where until a reboot all the files were there, and had correct contents. but if I rebooted the OS, they vanished. its coz the OS wasnt reading the disk just the caches.

the second level of testing is where the system is assumed to work, and we are testing eg the speed.

you want to also write files, delete some, write further files, delete, to test if the system can handle fragmentation.

a system might work perfectly until you delete one file, and now start malfunctioning because they forgot to handle fragmented storage.

when a system fails, that could be the hardware failing, or it could be the software failing.

you could connect up a brand new disk to see what happens, if that doesnt fail, it could be the hardware failing.



With free included hardware, there is always a catch.

E.g MoBo that comes with G.Skill RAM. Since MoBo has 4x RAM slots and one gets 2x RAM DIMMs for free, to upgrade RAM (e.g add more), one would likely buy the same brand RAM as what they already got. So, one would be buying G.Skill RAM.
Essentially, with this freebie, G.Skill is trying to catch people to buy their RAM in the future, rather than competitors (e.g Corsair, Kingston, Crucial, TeamGroup etc).
they are also getting brand recognition, I have never heard of G.Skill before, but have heard of Corsair, Kingston and Crucial,

so just by giving the G.Skill for free, the customer is now becoming familiar with that brand name! and will have positive thoughts because they got it for free!


Well, i don't know the exact minute detail as of why fan motor only clicks when it has power, but not enough power to turn over, thus producing this "click" sound. But there are plenty of similar cases out there. One of them is the fluorescent lamp you said. Another one, perhaps more relatable is car battery.

Namely, car battery holds 12V charge and needs to have plenty of charge left, to start the car engine. Now, if battery is drained, all the lights in the car interior will still light up, including headlights (sometimes headlights can be dimmed as well) but there isn't enough charge to start the engine. Either there is nothing when you turn the ignition key, or you'll hear car starter to go "click-click-click" loudly. This is what i've personally experienced (had to call car tow service, where technician then charged my car battery, until i was able to start the engine).
I suppose that is the inertia of electrons percolating through the large battery.


Importing/exporting bookmarks is common. History not so much, but no issue there either. But passwords import/export is something i wouldn't do, since the file where all the passwords are stored is high security risk. As far as i know the import/export file isn't encrypted and if somehow malware gets their hands to that file, all your passwords are out there (in the internet) instantly.

for financial things it is no longer a problem, as all UK banks and the german bank I deal with need a mobile phone or the physical card to verify.

with credit cards, they first check if its the billing address, but even with that, if its an expensive purchase, they either send an OTP (one time password) to my mobile phone as a text, or I have to use their app which is linked to my SIM.

I could literally tell you all my banking passwords for my 3 bank accounts, one in Germany, and you wouldnt be able to login to any!

because the first login from a new browser, you have to insert the card in a special machine, and eg with the german one they have flashing lights on the screen which the machine has to read to generate a one use number.

so I am not worried about malware getting hold of the file. I have done this lots of times with no problem.

malware is only interested in money, so they are only interested in the financial ones, but those are all locked out for money. eg investment sites only allow money to be withdrawn to the sign up bank account. to change that they will personally take verification steps.

Speaking of internet security, i've been part of MalwareBytes Premium for some time now and they often come up with good articles and tips regarding online security (i get newsletter from them).
Once such latest endeavor, which i think is worth sharing with others, is their new digital footprint portal, which:

Link: https://www.malwarebytes.com/digital-footprint

but the problem is Malware bytes itself could be malware! you are trusting them with your passwords, all bunched under one password, so they have all your passwords in the bag!

their name is a bit dubious: malware bytes, they are kind of telling you they are malware!

beware also of confidence tricks, where someone explains to you how to keep your stuff safe, but that is to build up your confidence, that they are on your side, before they cheat you!

its the old scam of someone getting attacked by person A, then person B comes in and rescues them and chases of A. but in fact A is an accomplice of B, where its all fake!

when Russia invaded Afghanistan, they said they were invited in by the afghan government, and it was to protect Afghanistan!


Note: Not sharing this as an advertisement of MalwareBytes, but instead as a free tool to check up on your internet security. Did try it out myself and did find some troubling info about my e-mail addresses.

but they could fake that, they know your email addresses and passwords, so they can fake those being out there on the dark web! which is similar to the scam mentioned earlier, where you create a fake threat and then come in to protect from that fake threat!

I have to use the browsers for passwords, so I have no choice but to trust them, I dont want to trust anyone beyond them, and I am ok with unencrypted password files for migrating browsers.

what's to stop malware from recording all the keys you press and trawling out passwords from these, eg with AI it must be much easier for them!


Ended up cleaning my house, removing several unused accounts + made password changes.


Yes, there are password managed add-ons for Firefox. Though, i don't use any of them. Instead, i use the password feature built-in to the Firefox itself, where i have also set a primary (master) password. Meaning that any site i go to, that have usr and psw credentials saved in my Firefox, can not be used before i enter my primary password. I get a pop-up to insert my primary password once i reach that site.

I am less paranoid, they cant get any of my money from my passwords. now in 2004 they probably could, but the banks have all wised up and their MO today is bulletproof. main loophole is if you get phoned up by a scamster pretending to be your bank, where they are in fact pretending to be you and getting the code numbers from the SIM or credit card from you.


 

Aeacus

Titan
Ambassador
this would then explain the plumbing I can see in the graphics card. but maybe they should do 2 versions of the graphics card, the one with the plumbing for people who intend to use liquid cooling, and one without the plumbing for people who wont use that. then the graphics card for the latter group can be made much smaller.
GPU PCB is reference design from GPU chip manufacturer (AMD or Nvidia) and AIBs doesn't have a say over the PCB design and size. All what AIBs can do, is improve power delivery, which in turn improves GPU OC and also better cooling, over stock design.

That, and also the fact that custom loop builds are ~1% of all PC builds out there. So, no point for GPU manufacturer to cater to the niche of PC building.

In a similar sense, why would car manufacturers cater to these people:
Engine tuning. Looks essentially the same as custom loop water cooling for PCs.

40482b26b8f54e48994df1232d55d902.jpg


RICE.

Poser.JPG


Which means: Race Inspired Cosmetic Enhancements,
but i like to say: Radically Immature Car Enthusiast.
malware is only interested in money, so they are only interested in the financial ones
Besides what USAFRet already said, malware can also delete/corrupt files on your PC. Some malware is less severe, playing custom sound or making your desktop look weird.

Here's something neat;
Open up Google and type into search bar: do a barrel roll
and then press Enter. Look if you see something interesting. :)

Few other neat commands you can try in Google search bar are:
askew
blink html
google in 1998

but the problem is Malware bytes itself could be malware! you are trusting them with your passwords, all bunched under one password, so they have all your passwords in the bag!
I think you misunderstand. Malwarebytes isn't a password manager. It is actually legit anti-malware software. And it is one of the best ones out there.

For 2nd opinion, Malwarebytes free version is also listed as one of the ones to consider,
article: https://www.techradar.com/best/best-free-antivirus

And even the payed version of it, Malwarebytes Premium has reviewed well,
article: https://www.pcmag.com/picks/the-best-antivirus-protection

Full review of Malwarebytes Premium: https://www.pcmag.com/reviews/malwarebytes-premium

There is an old saying: "If you spend more on coffee than on IT security, you will be hacked. What’s more, you deserve to be hacked."

In this day and age, one must be cyber security expert to protect their PC and data from all and any malicious software.

their name is a bit dubious: malware bytes, they are kind of telling you they are malware!
Malwarebytes Inc. is a legit company,
wiki: https://en.wikipedia.org/wiki/Malwarebytes

but they could fake that, they know your email addresses and passwords, so they can fake those being out there on the dark web! which is similar to the scam mentioned earlier, where you create a fake threat and then come in to protect from that fake threat!
Malwarebytes doesn't know your passwords. All what that portal does, is scanning the web of any matches regarding your email address and if any account data has been leaked via the data breaches. If there are leaks, portal will show you what those leaks are, so you can go and deal with the issue. It's an awareness tool, free to use. It is no different from e.g you typing in your full name to Google and then looking what Google has found about you from the internet.

I have to use the browsers for passwords, so I have no choice but to trust them
Well, you do not have to save passwords within browser itself. No-one is forcing you to do that. Might as well keep all usr and psw credentials in your head and type them in once you reach the site where you have an account. Saving passwords within browser itself is just convenience.

what's to stop malware from recording all the keys you press and trawling out passwords from these
Well, that is a keylogger.

Here's a question: How can you tell if your PC is infected with malware? Other than using legit anti-malware software to scan the file system for malware (which Malwarebytes software is all about)?

main loophole is if you get phoned up by a scamster pretending to be your bank, where they are in fact pretending to be you and getting the code numbers from the SIM or credit card from you.
This looks to be country specific.

Here in Estonia, we never get a call from bank to ask for our credentials. Banks just doesn't operate like that here.
And even when scammer somehow imitates my voice via AI, even then they get nothing when they call my bank, since transactions/withdraws can not be done over the phone.

Online security is very tight around here. After all, Estonia is forerunner in the world regarding online security.
According to the National Cyber Security Index, Estonia is 4th in entire world (UK is 15th),
link: https://ncsi.ega.ee/ncsi-index/?order=rank

why not send the mains to each foor unrestrained, and then limit the pressure per floor, where that then redistributes to each apartment? they then would just need to test the outer system.
From the city, we get 2.0 BAR on our water main. This is most what city will offer us. Since our apartment building is 5 stories high, we have had issues of hot water reaching 5th floor. So, we had to install water pressure pump to our main water line, which we can define the pressure with. And we put it to 3.0 BAR, which is enough for water to reach 5th floor at good pressure.
 
  • Like
Reactions: 35below0

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
False.

Malware can cause your system to become part of a botnet.
Or run a coin mining program and sending the results elsewhere.

Or ransomware. Encrypt your whole system, and make you pay to get it unlocked.

but they can presumably do all this without needing the passwords?

and the last one:
Or ransomware. Encrypt your whole system, and make you pay to get it unlocked.

is my comment that they want your money.

I still want the liberty to export or import to unencrypted files otherwise its a nanny state where Microsoft or Firefox tell me how to run my life and where they want me to keep my stuff on the cloud, where unknown entities can trawl through everyone's stuff.

maybe they should stop people having money because they might spend it on something not good for their health.

if I was really paranoid about some password I wouldnt save it at all.

at the end of the day I wont use their synch mechanisms, and will just manually re-establish passwords, they just arent going to get my passwords on the cloud. also I will migrate to browsers which allow me to do what I want to do.


if you keep to safe sites, I think its much more difficult to get malware on your system. eg if you keep to amazon, ebay, and other established sites eg tomshardware.

for many many months now the virus checkers havent found any malware at all on my system. but in the old days they'd find a lot. the server companies now have much more effective malware protection.

with all my bank accounts, they always text a one time passcode for transfers to new recipients, so someone cannot get any money out with my passwords. they need my SIM card or debit card to complete the transaction. for online transactions my bank and credit cards will text me for verification for larger payments to new recipients.

my more valuable information I keep on disks which are never connected when I am online, and which I backup from time to time, so with ransomware they wont get any money from me. because with such people there is no middle ground, once you give in to their demands, they will come for more.

the main things of value on my system are eg bookmarks of useful links, whilst I wont want to lose those, I wont pay 1 pence for them!

most people have no money, so the work of a scamster is very frustrating, as you cant scam someone who has no money!

with some financial sites, they email me every time I login, that way if someone on the dark web logs in, I will be notified and merely change my password.
 

DSzymborski

Curmudgeon Pursuivant
Moderator
but the problem is Malware bytes itself could be malware! you are trusting them with your passwords, all bunched under one password, so they have all your passwords in the bag!

their name is a bit dubious: malware bytes, they are kind of telling you they are malware!

Because it has Malware in the name? So I should stay away from the Sloan Kettering Cancer Center because their name is telling me they give me cancer? And I should not call 911 in the event of a fire because they're called fire departments, not extinguishing departments?

I'm sorry, but if this is how you approach a problem like this, Microsoft or Google telling you what to do with your PC is a way safer option.
 
  • Like
Reactions: Aeacus

Aeacus

Titan
Ambassador
I'm sure that @Richard1234 posted another reply, since i read it but decided to reply at later time. But now, the reply is gone "poof". :??:
One of the things asked, in that now gone reply, was the definition of AIB, a word i used in my previous reply.

AIB = Add-In Board. A printed circuit board that acts as an accessory for another device.
It is common term for 3rd-party graphics card vendors, as well as informally used to refer to graphics cards from such vendors.

Dedicated GPUs come in two forms: FE and AIB.

FE = Founders Edition. A dedicated GPU that is fully produced and released by the GPU chip maker themselves.
Few such examples include:
Intel Arc A770, specs: https://www.techpowerup.com/gpu-specs/arc-a770.c3914
Nvidia GeForce GTX 1060 Founders Edition, specs: https://www.techpowerup.com/gpu-specs/nvidia-geforce-gtx-1060-founders-edition.b6559
AMD Radeon RX 580, specs: https://www.techpowerup.com/gpu-specs/amd-radeon-rx-580-8-gb.b5129

And AIB counterparts for the listed ones would be:
AsRock Arc A770 Phantom D OC, specs: https://www.techpowerup.com/gpu-specs/asrock-arc-a770-phantom-d-oc.b9798
MSI GTX 1060 Gaming X, specs: https://www.techpowerup.com/gpu-specs/msi-gtx-1060-gaming-x.b3703
Gigabyte Aorus RX 580, specs: https://www.techpowerup.com/gpu-specs/gigabyte-aorus-rx-580-8-gb.b4426

E.g the MSI RTX 4060 Gaming X is also an AIB (aftermarket) model, made by Nvidia board partner: MSI.

Though, not all GPUs have FE (or vanilla) version of them, only AIB (aftermarket) versions.
E.g some such include: Nvidia RTX 3050, AMD RX 6600 and some more.

And then, there are even GPUs that only come in FE version, without any AIB version of them ever being created. Few such rare and notable ones include:
Nvidia Titan X Pascal, Nvidia Titan Xp.

Main differences between FE and AIB models are:
FE editions have much better binned GPU chips, allowing for far higher OC. E.g when Nvidia makes the GPU chips, it takes the best of them (cream of the crop) for itself and usually makes FE editions of the GPUs, while selling the rest of the GPU chips (lesser ones) to AIB partners.
Also, GPU chip manufacturers are very strict (especially Nvidia), on how much AIB partners can change within the produced GPU. Currently, what AIB partners do better than chip manufacturer themselves, is better power delivery for the GPU and also better cooling for the GPU. And of course, fancier looking as well. FE editions are dull and plain looking, without any fancy RGB LEDs. But many AIB versions have fancy looks (color theme and RGB LEDs for eyecandy as well).

Sad thing is, that GPU chip manufacturers, especially Nvidia, arrogance and rudeness has no bounds. Things got so bad, that the longest running AIB partner for Nvidia: EVGA, throw in the towel and completely stopped making GPUs, starting from RTX 40-series.

Full story here:

View: https://www.youtube.com/watch?v=cV9QES-FUAM
 
Last edited:

COLGeek

Cybernaut
Moderator
This thread has gone off the rails and has become a meandering, off topic collection of random wanderings.

Please cease and desist on sidebar and off topic discussions, feel free to continue those via private message.

What started as a typical thread about a build, has now gone on for 15 pages.

Complete the build, please.

Specific issues with the build deserve their own succinct threads, not this endless exercise in verbosity.

Are we clear?
 

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
This thread has gone off the rails and has become a meandering, off topic collection of random wanderings.

Please cease and desist on sidebar and off topic discussions, feel free to continue those via private message.

What started as a typical thread about a build, has now gone on for 15 pages.

Complete the build, please.

Specific issues with the build deserve their own succinct threads, not this endless exercise in verbosity.

Are we clear?
yes!

progress report: I have been experimenting with installing Linux and Windows scratch installs to a 2nd drive, to see how they co-exist with the installs on the first drive. this is so in the future if an install becomes broken where I dont want to reformat in case I need to salvage stuff, that I dont find myself painted into a corner variously, better to test what can and cannot be done now before I put too much time and installations on permanent installs. eg I might have a win10 + win11 on disk 1, then win10 becomes broken, and install win10 to disk2 and then find win11 on disk 1 is lost.

I decided to see what happens if I put partitions at the start and end of the 2nd 2T M.2 drive, both of 100G called Dummy1 and Dummy2. Because when I installed win10 to the first disk, the installation put some extra partitions, so I want to see what happens if I obstruct that. does it work around them or does the install fail.

I then installed Linux Mint in order to also see what happens with Windows installs if you install Linux first.

I selected for the boot loader to be that same 2nd disk

where in the photo, that 2nd disk is /dev/nvme0n1 where the highlighted item is where the install will go.

Linux Mint did install, but the mobo's loader options are only the first M.2 drive, either the Windows loader or the ubuntu loader on the other drive. there arent 2 ubuntu loader options.

but I get all 3 linux installs from the ubuntu loader, 2 on the first disk, 2 on the 2nd disk.

I then decided to see what happens if I hem in the Windows 10 install by arranging partitions thus:

Dummy1 ; Ballast1 ; for_windows11 ; Ballast2 ; for_windows10 ; Ballast3 ; Linux_Mint ; Dummy2

where there is no unallocated space.


with this, the Windows install gives me an error message:

A media driver your computer needs is missing ....

rebooted in case it was a read write error of the optical drive, but same message.

so now I deleted the for_windows10 partition via another Windows desktop, and this time didnt get that error message. so one shouldnt set aside a partition for a Windows 10 install, but should set aside unallocated space.

I selected the unallocated space which used to be for_windows10 to install Windows 10, and got an error message with error number 0x80004005

I then tried various things, eg subdividing the zone with some 100G before and after the allocation zone, but same error. then tried also deleting Dummy1, in case it wants space at the start of the disk, but same error.

so it looks like to get Windows 10 64 bit installed on a disk, the first thing done with that disk needs to be either a Windows 10 or 11 install, if you create other partitions first, this will be problematic.

by knowing this, there is no point setting aside a partition on an already used disk for a future Windows 10 or 11 partition, unless the first thing done with that disk was such a partition. eg I have 3 windows 10 installs and 2 windows 11 installs on the first disk,

before reformatting that disk to put a Windows 10 as the first install, I will remove the other M.2 to see whether the Linux Mint runs. this is to test whether everything is tied to the other disk, or whether it is only forwarding to that disk.

the original advice was to put each OS on a different disk, but based on the experiments so far I will definitely not do this, and will have all the permanent OSes on the same disk, as I have tested this to be 100% viable and practical, as I currently have three Windows 10 installs, and two Windows 11 installs and two Linux Mint installs on the same disk, and no problem at all, I can boot directly to any of these 7 OS installations by reconfiguring, without needing a reboot to access any, other than when reconfiguring.

ie some reboots to reconfigure to a different OS, but now the computer will always boot directly to that OS.

but I need to also determine what happens when OSes are on more than 1 disk, where I will just test the case of 2 disks, as I dont want an existing install to get marooned. and the experiments so far show there is jeopardy in installing Linux first, that then 64 bit Windows 10 + 11 wont install.

the original advice was to put each OS on a different disk, so installing Windows 10 on the 2nd disk, will give me a preview of that scenario.

but right at the end I will reformat the first 2T drive, and install the permanent installs there, as that one is at M2_4 which is easy to access. the 2nd 2T drive is at M2_2 (or is it M2_3?) where the 4060 graphics card eclipses the access screws of the frozr cover plate, so is unsatisfactory as a location for the system drive.

I have already done a compressed sector backup of that drive to the 4T M2_3 drive.

when I reformat the 1st disk, I will firstly delete install partitions, and see what happens, do the other installations still work? if things go to pot, I will then reformat the entirety. I am stress testing the multi OS aspect, its not that I want lots of installs, but I might need to do a 2nd install and dont want to get painted into a corner.
 
Last edited:

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
Bingo. If an OS install would cause you to lose any important data at all, something's seriously broken about your backup routine, and thus the upkeep of your PC.

I agree that if I used automatic backups that I wouldnt lose important data,
but I still want to know how the OSes coexist, eg say the win 10 install became full. I might want to install a 2nd win 10 install on another disk, where I still want to be able to run the earlier win 10 eg some data can be stuck on a specific software install, where you cant shunt the data to another install.

at the moment I am running into a problem trying to install Windows 10 to another disk.

before trying that again, I removed the first disk with the various OSes, and Linux on the remaining disk doesnt boot, even though I requested the boot loader to be the same disk. also with the mobo UEFI there is no ubuntu load option at this point. so Linux isnt forwarding the loading.

reinserting the disk, back to normal,


I then deleted all the experimental partitions on the 2nd disk, the dummy partitions and also the Linux install partition, and both the windows 10 and 11 install disks give me the same error of cannot read the <ProductKey> setting from the unattend answer file.

windows 10 install error message:

http://www.directemails.info/tom/win10/unattend_win10.jpg

windows 11 install error message:

http://www.directemails.info/tom/win10/unattend_win11.jpg

I want to do a manual install, so I dont want to arrange some unattend answer file.


Only remaining thing I can think of is to remove the first disk with all the existing OSes, then maybe they will install. I got through that error message when I had the Linux install on that disk, but then ran into a different error message later.

unless I temporarily install Linux to get past that, then from the win 10 install delete the Linux install and then continue the win 10 install.



the people who put each OS on a different disk, I dont know if they run into this problem when they install the 2nd version of windows on the 2nd disk. eg say they install win10 on disk 1, do they run into this error message when they try to install win11 on disk 2?

is the problem because I have several installs of both windows 10 and windows 11 on the first disk?

eg would say win10 on disk 1 and win11 on disk 2 work, but not win 10 on disk 1 and then win 10 on disk 2?
where different license keys for those two win 10's, I have 8 license keys currently.

I have a photo here also of how with the frozr cover plate of M2_2 and M2_3, only one screw is accessible, the other is eclipsed by the 4060 graphics card:

http://www.directemails.info/tom/win10/frozr_eclipsed.jpg
 

USAFRet

Titan
Moderator
is the problem because I have several installs of both windows 10 and windows 11 on the first disk?
Probably.
It is very easy to get things in a totally screwed up state.

I strongly suggest starting over.

One single drive, one OS.
During the install, delete ALL existing partitions.

Get it up and running.
Use it for a few days.

Then and only then, branch out.
 

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
Probably.
It is very easy to get things in a totally screwed up state.

I strongly suggest starting over.

One single drive, one OS.
During the install, delete ALL existing partitions.

Get it up and running.
Use it for a few days.

Then and only then, branch out.
I need to think carefully before trying this, eg which disks to put an OS on, and which OS for which disk, and what sizes,

also to be doubly sure that I have all data elsewhere.

the 4T disk currently is one huge 4T partition, as that one is a workspace, I have to think whether to say install an OS to each of the two 2Ts, then shunt the 4T data to data partitions on those two, then reformat the 4T, put an OS at the start, then the remainder as one large partition, then shunt the stuff back to that.

or whether to just leave that one as a data only partition.

I will probably test out a few other things first, before trying the above as it is major upheaval.

eg removing the first disk would test whether it then installs to the 2nd one, and to see if trouble when the first one is reinserted,

but that wont deal with getting an OS to all 3 without removing the graphics card.

also I plan to delete partitions one by one, and see whether the remaining OS installs still load, and once all deleted to try out your idea of one OS on each disk, and only then to put more than one, at least to see what happens.
 

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
As recommended...ONE OS, ONE drive.
Work with that for a while.

Currently, you seem to be trying to do too many things from the start.
With never having a stable point to proceed from.

I will try your suggestion, but just a bit of troubleshooting first, because when I remove the M.2 drive at M2_4 with the OSes on, the windows 10 install bluray gives the unattended answer file error message.

I need to also assess the different disks, in order to make a plan,

the 4T drive currently has 445.58GB of data on it,

I have a 2T WD Blue magnetic drive in the tower unconnected, that I could shunt all that to. I had a look at what is on that drive, and I think I can reformat it as it was just experiments of some weeks ago.

I need to devise a plan before acting on it as there is danger of painting myself into a corner.


I am going to try and create a thumbdrive windows 10 install, to verify if the bluray is corrupted, by seeing if the thumbdrive gives the same error message. because further back in time the win 10 install bluray sometimes would give an error message, but if I rebooted then no error message. could be the drive itself is corrupted.

the win 10 installer disk is supposed to run for existing win 10 installs in order to repair them, with the option right at the start of the install. I dont understand why it isnt functioning when I remove the M.2 drive with the OSes.

I tried deleting all partitions on the 2nd disk, and reinstalling OEM Linux from scratch, but now the win 10 disk gives that unattended answer file error message, whereas earlier on it didnt. the linux install functions fine.

and as mentioned, removing the drive with the OSes also has the windows 10 install bluray BDR giving the error message.

what I am guarding against is that if I delete all the partitions on the drive with the OSes, I dont want to then find the windows 10 installer doesnt work! I want to determine it does work before deleting the partitions.

eg I dont see why as an experiment I cannot just remove and store temporarily the M.2 with the OSes, and try your suggestion on the other 2 M.2 drives, the 2T and the 4T under the inaccessible frozr for M2_2 and M2_3.


when you install a brand new drive I think it has to be initialised. will deleting partitions return it to the state it is after initialising, or do I have to do some lower level re-initialisation?

I have scrutinised the partition layout of the disk with the OSes, and it has a different structure from the 2nd M.2 disk where originally I setup data partitions, which I more recently deleted. the disk with the OSes has at the left an "EFI system partition", and it has a recovery partition to the right of the first OS, win 10,
but not after consecutive partitions including the first win 11, but when I set the 2nd win 10 and win 11 at the far end of the drive, recovery partitions appeared to the right of each of those.

whereas the other M.2 when setup as data partitions only, didnt have any of those extra partitions, neither an EFI system partition nor a recovery partition.

tomorrow as a first step I will try to create a thumbdrive for installing win 10, to see how that fares.

I just wonder also whether the system doesnt like the fact that I moved the M.2 with the OSes from M2_3 (or was it M2_2?) to M2_4, maybe I should return it to the socket it was installed at?
 

USAFRet

Titan
Moderator
whereas the other M.2 when setup as data partitions only, didnt have any of those extra partitions, neither an EFI system partition nor a recovery partition.
That is as it should be.

The problem here seems to be your assumptions of how you think things should be, vs how they actually are.
Especially in regards to multi boot configs, and especially multi boot on a single physical drive.
 

Richard1234

Distinguished
Aug 18, 2016
277
5
18,685
That is as it should be.

The problem here seems to be your assumptions of how you think things should be, vs how they actually are.
Especially in regards to multi boot configs, and especially multi boot on a single physical drive.
I have done some troubleshooting, but no progress yet. right now the thumbdrive for win 10 is downloading.

via the Linux mint DVD for trying without installing, I deleted the 2nd M.2 and also the 2T WD Blue SATA drive, so both are available to test installing one OS. I then removed the M.2 with the OSes, but the windows 10 install gave me the unattended answer file error. I then also put the frozr plate on without the drive, and same error.

now maybe the disks have to arranged in the right sequence of M.2 sockets,

the next test will be to try the installation from the thumbdrive.

your suggestion ought to work if the disk with the OSes is removed from the PC, but so far no progress, I will keep troubleshooting this, eventually to remove the graphics card and put just an empty drive at one M.2 socket under the frozr plate under the graphics card. with no other disks.

but before that I will test the thumbdrive, to determine if the problem is with the bluray drive and disk.

I am assuming things because I dont have any info on 64 bit windows multi boot, whether from one or several drives!

so instead I have been testing what worked with 32 bit Windows XP, and 32 bit Windows 10, and that functioned until recently!

I dont completely trust documentation, things often are different from the documentation, eg I didnt upgrade my 2010 PC to win 10, because many years ago I had read the requirements and my PC failed that. but then a few years ago I couldnt install Norton as it no longer worked on XP. on contacting them, they said win 10 will run on my 2010 PC, I tried that and it worked. so I went all those years because I trusted the official info!

also when the desktop icons didnt work on the 2nd Win 10 installation, the Microsoft website said I needed to do a reset. but now I could no longer access the UEFI and most software uninstalled. I eventually found a youtube vid saying you mustnt do a reset, but there is an option to fix things where all your software should continue to function, where I think this option is below the reset option!

so although I make assumptions, I also test those! to refine the assumptions.

before deleting the multi OS disk, if I remove that, then your idea ought to work on the remaining disks, where if it does work, I can then reinsert and reformat the multi OS disk. right now this is what I am trying to do, and no luck so far. I have an empty 2T M.2 and an empty 2T SATA WD Blue magnetic disk for this.

the troubleshooting will clarify where the problem lies, if the thumbdrive doesnt work, then I will remove the graphics card, remove all drives except have just 1 empty 2T M.2 drive. and then try again.

one general problem is everything is way too complicated, a lot of this complicatedness is unnecessary, where it all needs an overhaul.
 

USAFRet

Titan
Moderator
your suggestion ought to work if
Sir...what I've outlined is Troubleshooting 101.
The very basics.
All the other junk is counterproductive and needlessly complex.

For the OS install:
 
Status
Not open for further replies.