Root Access

allquixotic

17:00

@OliverSalzburg why there?

Braiam

the best phrase that the obstinated can say: unless someone proves me wrong, I'm right

Psycogeek

@Boris_yo No wakey wakey? mabey you tried to wake up up when it was in delta sleep mode. Thats what heppens to me too, cant see nothing for a few minutes :-) How is it connected? there are reported incidents of bad wakup from HDMI , where they reconnect the wire to get it going again? I think people adjust all the s1-s7 sleep states to solution such things.

Jimmy Hoffa

@allquixotic what do you think of the idea of murdering swap on a windows 8 boxen with 8gb of ram? It should be fine, but will it be particularly faster running all in ram ya figure?

Ideally it'll lower my IO reqs too so things like loading levels etc in games should be faster cuz they're no longer just reading from-disk-to-disk as swap causes

Darth Android

17:17

@JimmyHoffa no.

Psycogeek

@JimmyHoffa have you ever seen any diatribes on that that really make any sence? "if it doesnt get paged it will have to get it from the disk again" no kidding? and where is it being paged out to? the disk. ok so your saying that this improves it because instead of comming from disk, it will come from disk. (in 4k sections instead of whatever cluster size you might have set for the disk)

Darth Android

Killing swap/pagefile will not make your system faster.

2

Swap allows Windows to page out applications that are not in use so that it can increase the size of the disk cache, and hold more files (like your levels) in memory

If you turn off swap, then things like Windows Update, the print spooler, and other system services that are not currently in use start to take up valuable memory, even when not in use.

Bob

Also, even with 8 GB, if you don't have a page file you can still run out of virtual memory quite quickly.

Psycogeek

@JimmyHoffa Then the SSD user , does not want it on the fast SSD, because it would write the SSD to death (supposedly) . so now they have a total handle on the situation, by paging to the slower disk :-)

Bob

Processes committing memory will "use up" virtual memory, but not necessarily ever actually use it.

The space must be reserved somewhere.

Darth Android

17:22

For example, the Memory Comitted on my laptop right now is 9GB

The Physical RAM in use for applications is only 8.3GB though

which means I have an extra 700MB of disk cache thanks to my page file

Oh wow, nope, apparently Task Manager rounds, I actually have 9.5GB comitted.

Bob

There's memory committed but never used (reserved, but never written to), and then there's memory that's been used and then paged out (e.g. not accessed in X time and memory can be better used for something else).

Darth Android

so my page file is effectively giving me more than 1GB of extra memory to keep files cached in RAM

Bob

They're actually different things.

Darth Android

I thought the "Commit Charge" was the total sum of all memory requested by the system o.0

(system and its applications)

( In that Committed memory is always > in-use memory, and max committed memory = size(physical memory)+size(page file) )

"Committed Memory" in Windows' terms doesn't just mean memory committed and unused

Hennes

Jimmy Hoffa

17:28

@Psycogeek But you only read from disk once this way and nevermore... paging reads from and writes to disk multiple times... Iduno it seemed like my machine was faster when I did this with 2gb of ram, and now I have 8gb...

Darth Android

@JimmyHoffa Unless you are using something other than Windows, Linux, or OS X, I can assure you that disabling swap will not make your system faster.

If your system is reading the same data from the page file over and over again, than that means you've over-committed your physical ram, and if you disable the page file, you simply will not be able to run what you're running now

You will get out of memory errors

Jimmy Hoffa

@DarthAndroid I can verify quite assuredly that: All terminology relating to meaning of various memory statistics is highly fucked, confusing, screwed, and totally fucking horrible to find out. Worked on a distributed data processing service once where we had to manually control and respond to the available memory on systems so that things were distributed and balanced right; and holy shit was it a miracle for us to figure out what each of the numbers actually are (note: many are misdocumented)

Bob

@DarthAndroid No, it means all memory reserved by a process, whether used or not.

Jimmy Hoffa

@DarthAndroid Untrue, I ran on 2gb without swap, I don't see why I couldn't now that I have 8gb

should be fine

Darth Android

@JimmyHoffa You can, yes, but it won't be faster.

Jimmy Hoffa

17:32

(I ran on 2gb in windows 7 ultimate just ~a week ago before I upgraded)

Bob

I ran into that problem when I tried to spin off >1000 threads for stress testing...

Jimmy Hoffa

@DarthAndroid you just said I can't... and it seemed faster when I did it with 2gb...

Psycogeek

@JimmyHoffa There are some programs out there that will fail with totally goofball errors if it does not exist in some way. an MS blog finnaly cleared things up more by saying something like " with todays memory ammounts , there is no need to have the 1X or 1.5X sizes anymore. and that logical paging sizes would be ok for logical ram sizes, depending on your use. So many people set a size relative to thier use. Like setting a simple 1/2g or 1Gig WHEN they have sufficient memory.

Darth Android

@JimmyHoffa No, I said that your page file is most assuredly not reading/writing the same data over and over again.

Bob

on a 64-bit system, the stack size defaulted to 4 MB and was unchangeable when using the TPL

that immediately committed 4 GB, and my virtual memory (physical + page) was only 8 GB.

each thread used < 1 MB, but committed 4 MB for the stack

for each thread, >3 MB (> 3 GB total) was never used

Darth Android

17:34

@Bob That's what I said, isn't it? In Task Manager/ Performance Monitor / Process Explorer, "Committed" memory is the sum of all memory reserved by processes

whether used or unused

Jimmy Hoffa

@DarthAndroid Alright, well I'll go with leaving the paging alone I suppose. Just thought I'd throw it out there, sounds like y'all have looked into this more than enough and all come to the same conclusion indepdently

Bob

@DarthAndroid yea

that machine only had 4 GB of physical RAM - if I didn't have a page file, I would have hit an out of memory error much sooner

Darth Android

@JimmyHoffa The same goes for RAM disks - You can make one thing really, really fast with a RAM disk, but it will come at the cost of overall system slowdown for everything, and I mean everything else.

Jimmy Hoffa

@Bob Yeah there is reason to this 4mb per stack, I recall reading some details from .NET team at one point detailing how and why all the memory is allocated per thread to the explicit purpose of explaining: Do not think threads are lightweight. They are not.

Bob

sure, i only used 1 GB - but Windows would have refused to let me commit any more than the total virtual, which = physical + page

Darth Android

17:36

The best alternative to a PCIe SSD with everything on it in terms of having a fast system is an SSD cache over your main disk

Bob

@JimmyHoffa not sure if I came across that - I spent a good few hours looking for an alternative

I was trying to stress-test a JMS message broker - which could block for several seconds per request, and the goal was to test hundreds or thousands of concurrent requests

and the API was synchronous only

lots of fun

Darth Android

Non-blocking I/O :)

oh

Bob

@DarthAndroid I wish :P

Jimmy Hoffa

@Bob at a previous position I started the performance and infrastructure team, as a part of that me and a colleague spent a ton of time reading about all the .NET system usage shit like that, the innards of the GC and such. Biggest lesson I didn't expect: Avoid finalizers like the plague.

Darth Android

Also, NB I/O is a royal pain without some sort of yield construct

Bob

17:39

@DarthAndroid best I could do was simulate it by spinning off a bunch of threads to do the waiting

@JimmyHoffa hm?

all I really know about them is to use them to dispose of unmanaged resources

Jimmy Hoffa

@Bob Yeah, but if you can get away with it and trust yourself: Use the dispose() methods instead and always use your using() blocks at which point you'll never rely on the finalizer and you can get rid of it

Bob

@JimmyHoffa But with unmanaged resources, I really can't trust that.

Psycogeek

@JimmyHoffa Problem is many people have been running without a designated paging file set to a spot on a disk, and they do just fine. you wont be able to explain that to anyone who it did not work fine for :-) I have seen the system make up a chunk on disk anyways without being told. i believe it was for the purpose of the memory dump. ( because i was not going to go over 8+g of memory dump, i just switched to small memory dumps)

Bob

Well, shouldn't.

Or so I've heard.

Jimmy Hoffa

If an object has a finalizer, the GC needs to know exactly where it is all the time and so to ensure it knows where it is, when that object is instantiated it's pinned. A pinned object has all objects it references pinned. And then GC compaction just goes to shit and you end up with a ton of empty space that can't be recollected

Bob

17:43

Don't really want to have open handles left lying around, do I?

Darth Android

I was always amused in Java how you could resurrect dead objects from the GC.

Bob

o.O

Darth Android

@Bob I think what Jimmy is saying is that you need to use using(){} and dispose() to preemtively close file handles when you're done with them, instead of waiting until the object is being GC'd

Jimmy Hoffa

@Bob No, but then so long as you trust yourself and your code you can remove fair amounts of finalizers... Depends on how much you trust your system not to explode violently in a fit of flames.

Bob

@JimmyHoffa Oh, no violent explosions. Just slow, painful memory and handle leaks.

Jimmy Hoffa

17:45

@DarthAndroid What I'm saying is if you have lots of objects that are referenced by finalizers you will find yourself spending a ton of time in GC

Bob

Then again, I don;t really trust myself, nor do I trust others who might use that code.

Sammy

What's the best S.M.A.R.T. program?

Jimmy Hoffa

@Bob Hopefully you don't have those :P Violent explosions are the only time finalizers are relevant. So long as your using() disposables you only need to worry about leaking handles if your process goes boom so hard that the finally's don't get called (this is generally like an OOM, or stackoverflow, or power-down at which point leaks don't matter anyway)

Bob

@DarthAndroid It's always good to do that, but the whole point of finalisers is to make sure that unmanaged resources are always released.

Jimmy Hoffa

@Bob Are they released in a stackoverflow situation?

Bob

17:47

@JimmyHoffa I'm pretty sure the OS will handle (heh) that if the process ends.

Darth Android

@JimmyHoffa That's the point of try{}finally{}, isn't it??

Jimmy Hoffa

@Bob nope, stackoverflow means "Ack, I can do nothing, I die now and do not clean up or do anything of value."

Bob

The problem is if such objects are repeatedly instantiated and disposed of while the process is running.

@JimmyHoffa The process won't clean up after itself, but the OS will make sure any open handles, etc., are closed. At least that's what I remember.

Jimmy Hoffa

@DarthAndroid OOM, stackoverflow, or power-plug-un = no finally execution (or any cleanup for that matter) and using() is just sugar over a finally

Darth Android

@Bob Not necessarily, those kinds of objects need to be associated with something that has a longer lifecycle

Bob

17:48

Memory leaks, barring major OS bugs, don't persist after a process ends.

Darth Android

Wait, why wouldn't a using trigger on a SO or OOM?

Jimmy Hoffa

@Bob That depends on the resources. With that logic if you don't have finalizers then the OS will clean everything up too

Darth Android

(I could see it not running properly in an OOM if the using statement tries to alloc a bunch of stuff )

Bob

@DarthAndroid But if I call CreateFile somewhere, and never call CloseHandle, then the handle will remain open until my process ends.

The GC has no idea how to release unmanaged resources.

Jimmy Hoffa

@Bob That's what using is for

Darth Android

17:49

^

Bob

@JimmyHoffa That's assuming Dispose will be called. Which cannot always be guaranteed. Especially if unknowns are using (heh, again) the code.

Darth Android

@Bob if using(){} doesn't call a Dispose(), I'm pretty sure that means your process is about to be terminated anyways

Bob

The whole reason finalisers exist is for unmanaged resources. If a using or Dispose was always enough, then finalisers wouldn't exist.

Darth Android

And I wouldn't expect a finalizer to be called either

Jimmy Hoffa

using(FileStream myStream = new FileStream("Arrr"))
{
  // code!
}

is de-sugarred into:
FileStream myStream = new FileStream("Arrr!");
try
{
  // Code!
}
finally
{
  IDisposable disposeThing = (IDisposable)myStream;
  disposeThing.Dispose();
}

Bob

17:51

@DarthAndroid What I mean is that's assuming the consumer of your code uses (hey, a third time) a Dispose or using.

You are trusting humans.

Very bad idea.

Jimmy Hoffa

using() is compiler sugar to give you a finally block

Bob

@JimmyHoffa Cleaning up on process close is very different from cleaning up during execution.

Jimmy Hoffa

@Bob Like I said, if you can trust your situation enough to get rid of finalizers, do it. If you can't then well there ya go

Darth Android

@Bob Perhaps, but using finalizers and not giving the consumer of the code a choice can be worse, is what Jimmy is saying

That Help Vampire Guy

Silly question: are 2x 1GB DDR3 sticks on a dual channel mobo faster than 1x 2GB DDR3?

Darth Android

17:52

@ThatHelpVampireGuy faster for what?

That Help Vampire Guy

@DarthAndroid Hm... anything?

I forgot how dual channel works.

Bob

while (true)
    new MemoryLeak();

...

class MemoryLeak
{
    public MemoryLeak
    {
        CreateFile(...); // let's assume this is a DllImport
    }
}

Darth Android

If it's one page of data at a time, then no

Jimmy Hoffa

@DarthAndroid Trusting consumers is bad.. I'm talking about don't put finalizers in the code when you are the only consumer and you know you'll do it right. It's a serious optimization, but it is a risk. It's a risk worth taking if you know what you're doing and why it works the way it does

Bob

there, you have a problem.

it'll be solved by the OS on process end, but wile the process is running the OS can't do anything except throw an error back

Darth Android

17:54

if it's multiple pages, then half will go to one channel and half will go to the other channel

Bob

@DarthAndroid what kind of choice would they expect?

@JimmyHoffa it also sounds a lot like microoptimising with a lot of potential to introduce problems down the line

Darth Android

@Bob the "I'm smart enough to manage myself, don't leave pinned objects all over the heap" option

Jimmy Hoffa

@Bob do you understand what the cost of the finalizer is though?

Darth Android

@JimmyHoffa To some degree you have to trust your consumers, though.

Jimmy Hoffa

@Bob Not true, I've seen it necessary before. As a rule I always put the finalizers in unless I see a situation where it's going to be dangerous. It's particularly bad for long-lived objects.

Darth Android

17:55

Yes, if they write shit code, then they'll have a shit app

Bob

@JimmyHoffa As with all microops, it's necessary in some limited situations.

Einsteins Grandson

so i took that acer

but

Bob

But they're also generally things you'd never recommend for every situation.

Einsteins Grandson

the windwos 7 in it is not english

Darth Android

As long as it's well documented that you need to dispose your objects, I would expect/trust consumers to call it.

It's not a library's job to handhold the consumer

Einsteins Grandson

17:57

is there a way to use the licence key of windows 7 for some different language, install english windows 7 and use the same key , so that it would work?

Darth Android

Just to make whatever task your library is designed to accomplish easier to accomplish than with no library

Einsteins Grandson

i think it is impossible

Darth Android

@EinsteinsGrandson Try it?

Jimmy Hoffa

@Bob I didn't, I said if you really trust that your code is going to always do the usings. The .NET team has said they regretted having ever put finalizers in, and that if they had done the using() and idisposable interface correct the first time they wouldn't have had finalizers. They are not "sort of" heavy, finalizers have a severe impact on GC times

Einsteins Grandson

i would have to install new windows

;o)

another thing

this is win 7 64 bit

Darth Android

17:58

@EinsteinsGrandson That is what you described doing, yes.

Einsteins Grandson

should i install and use 32 or 64 bit applicatiions?

like IE

Darth Android

@EinsteinsGrandson Whatever you want.

Einsteins Grandson

or any other software

but which is one is better?

Darth Android

I prefer 64-bit personally, but unless they're exhausting address space, it doesn't really matter.

Einsteins Grandson

why do they even let 32 bit apps to run on 64 bit OS/

?

Darth Android

17:59

Because compatibility

Einsteins Grandson

ok

jokerdino

@Bob how would you rate this?

Einsteins Grandson

and 64 bit versions are OK right? no causing any problems

Darth Android

If you couldn't use 32-bit on 64, no company of any significant size would ever upgrade to 64-bit

Einsteins Grandson

i remember i installed 64 bit ubuntu 7 years ago.... and it was crazy

Darth Android

17:59

and that's where MS makes their money, so...

stuff was only crazy if you were running XP 64

Einsteins Grandson

also win 7 4 years ago.... and it was better using 32 bit aplications back then

ok

Bob

@jokerdino Decent but slightly overpriced?

Darth Android

better? There's no difference

Bob

Well, decent for web browsing.

Darth Android

(no significant difference, anyways)

Einsteins Grandson

18:00

also, how much of RAM does win 7 need?

i will use some vmware

Darth Android

I mean, I can make all sorts of benchmarks that show that either 32-bit or 64-bit apps are better, based on what you want me to make them say

Bob

@DarthAndroid That almost killed Itanium...

Einsteins Grandson

ok

Darth Android

for an end user though, it just doesn't matter in the real world unless you're doing something that requires more than 2GB of memory.

(Photoshop, WoW, etc.)

Bob

Rule of thumb is to install a 64-bit OS is you have or plan to ever have >= 4 GB or RAM

below that, it's a bit more optional.

As for programs, the majority of programs probably wouldn't notice.

How I wish they'd come out with 64-bit Firefox, though...

Darth Android

18:03

No multi-process and no 64-bit is kinda killer in this day&age :/

allquixotic

@Bob that's been "a thing" on Linux for a long time, and you can compile it on Windows, but you won't have Flash

That Help Vampire Guy

Weird: this answer says dual channel is always better, while this other says the difference isn't much.

Which one do I trust?

Bob

@allquixotic Well, I meant an official build. I never did get around to compiling my own.

Jimmy Hoffa

@ThatHelpVampireGuy They both say it's better, ergo...

Bob

Also, if Shumway moves along a bit faster, the lack of Flash wouldn't really be a concern.

Darth Android

18:04

@ThatHelpVampireGuy Better in numbers or Better in "ZOMG THIS FEELS LIGHTNING FAST?"

Because duslchannel is always better in numbers

jokerdino

@Bob with the inflation rate, i have to bite the bullet.

Darth Android

But you probably won't really notice the difference

Not without two systems side-by-side

Bob

Ya. Memory bandwidth in general isn't going to affect experience that much.

That Help Vampire Guy

@DarthAndroid No, OMGWTFBBQFAST is SSDs or so I heard.

allquixotic

you'd notice if you're running Star Citizen lolz

Bob

18:05

Bottlenecks are generally elsewhere.

Darth Android

@ThatHelpVampireGuy You heard correct

That Help Vampire Guy

@Bob Thought so.

allquixotic

@DarthAndroid any idea how I would add an SSD cache to an 8TB (2 x 4TB HDD) hardware RAID array? (Adaptec 6405E)

Darth Android

Put a nice SSD cache on that system, or if you can afford it, install a large PCIe SSD and move all your data to it

allquixotic

on Win32

That Help Vampire Guy

18:06

@DarthAndroid I witnessed it on a friend's i7 8GB laptop. Boots faster than the screen turning on.

Darth Android

@allquixotic Uh, for that your best bet is FancyCache

Bob

The secondary storage is one of the bigger ones, but moving to an SSD severely limits your storage - so SSD caching is the way to go where possible.

Psycogeek

@ThatHelpVampireGuy it is better/faster the quantity is not that much. A ram thouroput would show that it has incresed much, anyone showing how many FPS it changes in a Game instead, should be taken out back and shot for not understanding what ram speed provides :-) Oddly more people are checking for FPS based on ram speed, than for Reloading the game level from Caches :-)

Darth Android

or whatever they renamed it to

PrimoCache

allquixotic

looking at maybe a 256GB SSD, enough to make a dent in that 8 TB, especially considering I very rarely go outside a given 256 GB data set in a given day

Darth Android

18:07

@allquixotic PrimoCache will actually give you exact hit/miss stats

Bob

@allquixotic Heck, even a 64-128GB SSD would give a noticeable boost.

Darth Android

that's how I found out that my system has a 97.5% hit rate with a 36.7GB cache

Bob

You may hit 256 GB a day, but that doesn't mean you use all of it often.

allquixotic

@DarthAndroid so what, I just buy an SSD, install PrimoCache and it does its thing?

Bob

Only the most frequently accessed really needs to be cached.

That Help Vampire Guy

18:08

@Psycogeek Well, it will be for a file server (I'll probably be running freenas, nas4free or openmediavault on it), so I'm not expecting cpu/memory intensive activity from it, and definitely no game FPS =P

Darth Android

@allquixotic Preeety much. You install the SSD and PrimoCache, you tell primo cache to enable caching for all volumes on your raid Disk, and then you enable the L2 cache and point it to your SSD

allquixotic

@DarthAndroid cool

Darth Android

And tell it to use 250GB of cache

Bob

@DarthAndroid Does that work during boot?

Einsteins Grandson

i have 2 PCs

will connect them

to create a network

need to copy 20 gb from one pc to another

Darth Android

18:09

@Bob I think it does work doing boot, it loads as a driver in the kernel

Einsteins Grandson

how long will it take?

20 minutes maximum?

;o)

Darth Android

@EinsteinsGrandson too broad

depends on far too many factors

Einsteins Grandson

it's like 10 mbit per second?

first one is old pc

Darth Android

what network speed? How fast is the sending drive? how fast is the receiving drive?

Einsteins Grandson

second one is that acer notebook

Bob

18:10

@DarthAndroid And presumably falls back nicely if the SDD fails/is removed?

Darth Android

What programs are also using the drive while they're copying?

@Bob if you have write caching turned off, yes. If you have write caching turned on... eh, I didn't test that

It will only let you use SSDs flagged as internal drives by windows though

so you shouldn't have to worry about them being removed, just failing

allquixotic

hmm, it's a beta :/

Bob

Hm. I have one spare 3.5" bay and can afford a 32GB SSD... hrm...

Darth Android

It's been in beta for several years I think now

Boris_yo

►

My problem documented kind of

Darth Android

18:12

Which, this all moot if you have a Intel chipset with RapidResponse and all the drives you want to accellerate are connected to the chipset ports

Because that is far easier to configure, and much more tested

Boris_yo

It worked 3 hours previously today and suddenly it is not.

Darth Android

And works during boot/install of windows

allquixotic

@DarthAndroid problem being, my hardware RAID setup would go poof :P

Bob

@DarthAndroid Mine's a P67, that's only with Z68+ :(

I'm on Intel's fakeraid otherwise, which would have made it perfect :P

@allquixotic Assuming your hardware RAID controller can't do SSD caching?

Psycogeek

@ThatHelpVampireGuy Myself I will always go for the larger stick of ram, for the sole reason that later i can have 2 larger sticks of ram :-) And i have been stuck with smaller ram , and had to pull it ALL out to upgrade. Which just leaves the minor problem of the ram should all be the same (as much as possible) . So the answer is 2 stick of large ram :-)

Darth Android

18:14

So, just for those curious the Holy Grail of speeding up your disk without being a millionaire is an SSD cache of 32-64GB on (Windows: Intel RapidResponse, or PrimoCache if you don't have that, Mac OS X: FusionDrive, Linux: bcache or dmcache)

allquixotic

@Bob actually...

adaptec.com/en-us/solutions/hybrid-raid

Darth Android

@allquixotic I'm pretty sure all of the major RAID vendors have their own version of SSH caching on their raid cards

the question is, "Does your card support it?"

To which I've quite frequently found the answer to be "Did you spend an arm and a leg for the card?"

allquixotic

@DarthAndroid the 6405E is like $400 :P

Bob

oh crap

I just remembered that my primary HDDs are on the only two SATA3 ports

and changing that would be a pain - it would probably kill the RAID config

@allquixotic so, the answer is "yes"?

Darth Android

@Bob If you're talking about Intel RST, you can move the ports around

You can move them to a different system, even

Bob

18:18

o.O

Darth Android

as long as the RST drivers are the same version

Bob

interesting

Darth Android

It may be FakeRAID, but it's damn nice FakeRAID

Bob

I'll have to do that next time I get a bunch of drives... I'm rather behind in my full-image backups right now

last one was close to a year ago :(

Darth Android

My system makes full (non-image) backups every 15 minutes

Bob

18:19

I've got mine on daily incremental (fresh per week) for the important stuff

allquixotic

@Bob it doesn't support it unless the HDDs and SSDs are the same size -_-

Bob

a couple of drives aren't backed up at all :\

Darth Android

I was too busy to bother figuring out what was important, so I just checked the C drive

Bob

@allquixotic wut

lemme know when you get one of them 4 TB SSDs...

2

Darth Android

(Which, that's one caveat with RapidResponse, it limits it to a 64GB cache drive)

Bob

18:20

it should only cost half the GDP of the US

Blog bot

Blog - Stack Exchange: Podcast #53 – Let’s Go Rio

Alex Miller on October 28, 2013

Welcome to Stack Exchange Podcast #53 with special guest Gabe Koscky, our new Brazilian community manager, and usual suspects Jay Hanlon, Joel Spolsky, and David Fullerton. Today’s show is brought to you by the National Security Administration!

Site Milestones: We launched Astronomy, which is not the same thing as the Space Exploration site we’d previously launched. You can ask questions about gravity (the force) on Astronomy. You cannot ask questions about Gravity (the movie). Astronomy and Physics have a lot of overlap, and that’s okay! Also, you can’t say Count Dooku in Portuguese. This is an adult-only podcast. …

Jimmy Hoffa

I'm running native IDE mode, I switch to AHCI and get a "I sorry but no can WAT? kthxbye" on boot. Should installing my mobo's AHCI drivers solve this, or will it not realize it needs to change the driver it's using while boot up because it's not smart enough to swap drivers until after having loaded?

Unfortunately the OS was installed under native IDE mode

Bob

@JimmyHoffa which OS?

Jimmy Hoffa

win 8

Bob

XP came out before AHCI was finalised as a standard

Darth Android

18:21

uh

not really

5.1TB: $99,995.00

10.2TB: $124,995.00

Bob

@DarthAndroid it was a joke

Einsteins Grandson

another thing...

Darth Android

@JimmyHoffa If it was installed under IDE mode, then you have to do some registry hacking

Bob

@JimmyHoffa I think it was a simple reg change

Einsteins Grandson

there is no way to install other languages to an existing windows 7?

Darth Android

18:22

Grab VirtualBox's wiki guide for converting a physical partition to a virtual drive

Bob

6

Q: How to get AHCI support on an IDE Windows 7 installation?

I installed Windows 7 onto a SSD with the motherboard in IDE mode. Now when I try to change the motherboard to AHCI mode, Windows 7 will not successfully boot (blue screens mid-boot). How can I add AHCI support to the Windows 7 installation? If I can not add AHCI support I plan on doing a reinst...

allquixotic

@DarthAndroid srsly? rofl

Bob

10

Q: How can I enable AHCI on a SATA drive after installing Windows?

Can I enable AHCI after installing Windows 7 in a SATA drive that was on IDE mode?

windows-7 hard-drive sata pata ahci

Einsteins Grandson

so that i would be able to switch the langues of menus in windows?

Darth Android

it documents the stuff you have to do to enable the AHCI or IDE drivers in Windows

and gives you a registry script that'll do it for you

Jimmy Hoffa

18:23

@Bob do you think windows will boot in IDE mode even with that reg key flipped?

Bob

@JimmyHoffa no clue!

Darth Android

If you follow virtualbox's guide, it'll boot in both

Jimmy Hoffa

@Bob Magic!

Darth Android

I used it on my mom's system where the drive boots in AHCI from bios, or via IDE in virtualbox

so she can go native when she needs it, but has access inside windows 8 when vbox is fine

virtualbox.org/wiki/Migrate_Windows

Actually, scratch that. Go with those questions linked above

Bob

@DarthAndroid looks outdated

> Many physical PC systems (and VMware) have a AGP graphics bus which will cause agp440.sys to be installed.

haven't seen AGP in years

Braiam

18:27

@Bob if I had my camera with me, I would show you one :P

Bob

@Braiam oh, I have an old XP box (ca. '04) that has it

Darth Android

@Bob I imagine the market of people wanting to migrate an existing windows installation into a vbox drive still have AGP :P

Bob

I meant I haven't seen it on new systems.

@DarthAndroid Point taken :P

Darth Android

Oh... oh hell. Just... Surface Pro 2. 360 Controller. N64 emulator. Plane Flight.

Thank you Penny Arcade.

jokerdino

@Bob found this. Same specs but a bit less pricey.

Anyone else is free to pass opinions. @all? ^

Bob

18:37

@DarthAndroid nono, you need a proper n64 controller

Darth Android

@Bob but the USb adapters for those are like $75, if you can even find one

Bob

@jokerdino @allquixotic would probably recommend a Haswell, actually :P With Iris.

@DarthAndroid ebay.com.au/bhp/n64-pc-controller

Psycogeek

@Boris_yo so you just turn off your laptop (that did not come out of standby correct) and then turn it back on, and everything is fine right? So Most of the time it does come out of standby? and this was only standby or was it hibernation? (must be a tech joke in there somewhere , Did you try turning it off then on)

jokerdino

@Bob how new is Haswell? we barely have machines with i5 and i7. :/

Darth Android

haswell came out in june or july

I think july

jokerdino

18:40

so, we'll get haswell machines in 2017/8

Bob

@jokerdino Some June, some September

@DarthAndroid Seems like $20 for a generic USB-capable controller, or $10 for an adaptor.

Psycogeek

@Boris_yo Oh wait, you did say standby right? But that was a turn-on from some kind of off? because we wait for the boot, and hear winders logo thing? standby would be instant return with mabey a log-in?

allquixotic

@jokerdino craptop

jokerdino

i expected that from you.

what would be your ideal laptop?

Darth Android

@Bob True, seems several more manufacturers have released products in the past ~4 years

Tbh I prefer the 360 controller though

can't wait for the XB1 controller

or the steambox controller

allquixotic

18:56

@jokerdino a Surface Pro "5" with a Cannonlake CPU, 64 GB of DDR4, 2 TB SLC mSSD, 3d without glasses, built-in LTE Advanced, and... ah, who am I kidding, silicon will never be able to handle that density :|

Jimmy Hoffa

@allquixotic DDR(4?) ?

there's a 4 now?

allquixotic

DDR4 SDRAM

In computing, DDR4 SDRAM, an abbreviation for double data rate fourth generation synchronous dynamic random-access memory, is a type of synchronous dynamic random-access memory (SDRAM) with a high bandwidth ("double data rate") interface, expected to be released to the market sometime in 2013. It is one of the latest variants of dynamic random access memory (DRAM), some of which have been in use since the early 1970s, and a proposed higher-speed successor to the DDR2 and DDR3 technologies. It is not compatible with any earlier type of random access memory (RAM) due to different signal...

Darth Android

GPUs already run on DDR5, don't they?

@allquixotic No, but perhaps Graphene can hit that density?

Jimmy Hoffa

@DarthAndroid They have for a long time but that's gDDR which is different (I understand?)

allquixotic

@JimmyHoffa the number is the version of the spec, which has absolutely no direct correlation to performance

jokerdino

18:58

@allquixotic woah dude

allquixotic

if they wanted, they could declare that DDR4 chips will run at DDR1 speeds

ChatBot John Cavil

@jmort253 Welcome to Root Access chat for Super Users! I am this channel's helpful chat bot. Please don't ask if you can ask or if anyone's around; just ask your question, and if anyone's free and interested they'll help. For bot commands, type !!listcommands

jmort253

Thanks for the warm welcome, ChatBot....

allquixotic

GDDR5 is actually closely based on system DDR3, with a few tweaks that favor graphics memory

Transcript for