So you don't think it's the mobo that's faulty... Cause it's stranger that the first time It booted up and I was also able to install Windows on one of my SSD s
More recently with everything built in china :-) i build in slower baby steps, on the "bench" first. then onlly add things after it is showing a pic on (thanks for that) the built in GPU that might be available.
by being out of the case itself, you reduce any issues that might exist, one time i had a board that didnt want one of the standoffs to be grounded ??? WTF found out later it was a known (but very rare) problem for that one board.
How old was your PSU when you started? or is it a PSU bought for that build?
at any rate, when you build on a (non-conductive) open surface, you can see things, check things, and get things running without destroying things putting them in and out.
many times people have put into a case that had standoffs placed already, by somone other than the builder :-) am i going to trust anyone else NO!
get all crammed up in a case, and even the best of us (guilty) will forget one of the wires ,ooops. less is always better, there are only about 3 things needed on the motherboard to test operations. so if you have onboard or on-die GPU, prefer to not stuff powersucking gpus on until the thing already produces a picture, and you can get in the bios/uefi .
@Daniele i cant be sure, but we do know that large percentages of people who send back a failed motherboard, get another new one, and fail wit it too :-) so we do know that humans can have more possibility of error than motherboard fails .
@Daniele Also, what was the mobo, CPU and RAM model?
@Aibobot I'm not entirely sure @Daniele actually removed the RAM... it's perfectly possible to remove and check the mobo without taking the RAM (or CPU) out
@Daniele asus probably has a RamOK button? that will low-clock high latency the ram so you can get funtinality, then you can get in the bios and tweak it to operation.
@CausingUnderflowsEverywhere like every 210 & 212 used in the universe, they musta sold hundreds of thousands of them, once they were good enough, and $20
@MichaelFrank I was running in single channel mode instead of dual channel mode for the first three months because it felt like each colour pair is a channel so dual channel would be having the ram in two different colours, but in reality I was supposed to have them both in the same colour
@MichaelFrank not quite... well, depends. here PO Box is specifically at the post office and cannot be used by external services. then again auspost also does something similar with parcel lockers...
PO box is usually permanently rented
parcel locker is allocated temporarily when you receive something
From what I just read in a SO answer. So like the program is like alright I might use up to this much memory.. "states virtual size". I need this much of the memory available right now "private bytes" I'm really using this amount of my private bytes right now "working set" ?
So in a computer, we have the CPU , CPU Cache (insanely fast), RAM (pretty fast!) , HDD (super slow :( ). Alright I just upgraded from 4GB to 16GB. The most commit charge I ever had on 4GB was like 6GB. You know what, I have such plentiful RAM, I can remove the slow page file on disk cache. Now all the pages can be in RAM even if they aren't currently in use, it's all good. I'm sure there's a tiny speed boost since nothing need be paged from the page file
@Bob i think once i figured, there was nothing wrong with it being "virtual" that allocations are not FIxed , but there was something wrong with extending it outside of the ram spaces, Instead of people being more carefull and optomised with thier use of it, AND applying thier own paging if thier progam is the type that needs so much ram, that it would run from fake ram.
@CausingUnderflowsEverywhere ...no, there's a massive speed loss because now frequently-accessed files must be fetched from disk every time because you no longer have any space for the disk space in RAM
Now we STILL have out of memory errors , and insanely complex memory messes, and continual leaking and failures, BUT we have all the Fake ram anyone could ever want.
A lot of the time, this committed memory goes completely unused.
The program might request 50 MB, but only use 25 MB of it.
(More practically, that's actually because malloc allocates a large chunk of contiguous memory because lots of small allocations is actually slower. Then it can chop up those chunks to use.)
But Windows doesn't care how much is used. Only how much has been allocated.
Say Firefox commits 1.5 GB. Windows will now have to reserve 1.5 GB of its total virtual memory for Firefox. Firefox might only ever use 1.0 GB. That's 500 MB reserved but never used.
@Bob which does not explain why MS themselves has said more than once that people should use the "Dynamic" allocations instead of Pigging out an allocation, because your program leaks like a seive :-) but not being a programmer , i still blame Microsoft for starting it .
Say Firefox reserves 1.5 GB of virtual memory by committing.
@CausingUnderflowsEverywhere I'll get to your questions in a sec, want to finish writing this first.
Because Windows has no page file, that's directly 1.5 GB of physical memory reserved.
If 1.0 GB is used, that's 500 MB of your physical RAM locked up. It cannot be reserved by any other process.
There's a saying here: "unused RAM is wasted RAM"
Because physical memory that's not used by a process can be used by Windows to cache frequently-accessed files so you don't have to waste time on slow disk access.
(That's why "free" memory should be nearly zero while "available" memory should be high - "available" is mostly disk cache that can be discarded and used for process memory very quickly.)
Now, consider the other situation, where you have a page file.
Now Windows can put 1.0 GB (used) into RAM, but reserve the completely unused 500 MB in the page file.
This unused memory is reserved, but not accessed. There's no delay from the page file, because it's never actually written to or read from. It's just Windows knowing it has this extra space, just in case a program actually decides to use all it allocates.
It's the idea that in the absolute worst case of processes trying to use all virtual memory it's better to lag from paging than to randomly kill processes.
That is why a page file should still be used.
In the past, yes, a page file was used to compensate for lack of physical RAM, and heavy paging (and lag) was common.
These days, it's used as extra commit space, and not relied on as a memory replacement.
Now, there is a bit of an additional case where Windows can decide that a program you have not accessed for a very long time (minimised, etc.) is taking up RAM that can be better used for disk cache, and will page it out. That can cause a second or two delay when you next use the program. The tradeoff is that everything else you're actively using will speed up because you have more disk cache now.
This is the slightly more questionable part that some people aren't happy with.
imagine how much faster it would be to "run out of memory" if all this trash wasnt there. we got 25Gig memory speeds, and the OS handling and extensive managment and complexity of it probably brings the reality of that down to 2 . . . K :-)
@CausingUnderflowsEverywhere Private bytes are those allocated from virtual memory that are not shared with other processes. Shared bytes are shared (e.g. shared libraries). Working set is the bytes actually in physical RAM. Private working set is the private bytes in RAM.
@CausingUnderflowsEverywhere RAMMap should tell you how much of your memory is private, how much is shared, and probably how much is wasted too.
