Root Access

NaN

9:00 PM

I searched in the BIOS with those new PC, but I don't see the same option as on the nVidia PC

Darth Android

... WHY do they list that as being included?

@NaN No, allquixotic is right

NaN

@allquixotic exact

Darth Android

the specific processor that that computer comes with is missing the hardware that drives the lower DVI ports

NaN

but why would they ship 2 graphic cards, 3DVI and 1HDMI port

allquixotic

@NaN that Intel CPU doesn't have any integrated graphics, on the AMD Radeon unit. on the Nvidia unit, the Intel CPU has integrated graphics, and I guess the BIOS supports Nvidia Optimus, to allow you to drive monitors from the AMD card whose frames are piped from the Intel iGPU's frontbuffer.

Darth Android

9:01 PM

They didn't, that's the problem

allquixotic

they didn't ship 2 graphics cards. they shipped one graphics card with 2 ports. the motherboard supports iGPU within the CPU, but that particular unit doesn't have an iGPU, so they're dead ports exactly as you said.

Darth Android

^

allquixotic

you got exactly what you ordered

Darth Android

Though that page is exceedingly misleading

NaN

right thanks guys, so imagine this cheap PC sorry in french

has an ATI, and a i3 processor not ending in "P"

allquixotic

9:03 PM

the solution is to either purchase another graphics card with 3 monitor ports, or get another same-generation AMD Radeon and SLI them together (if that even supports SLI, I don't know), or send them back to HP and get a model with 3 monitor ports on the graphics card, or replace the CPU with a non-P model and pray you have motherboard support for Virtu MVP or similar

Braiam

@allquixotic that sounds so complicated...

allquixotic

@Braiam note the "or"s -- a bunch of simple possible solutions OR'ed together

NaN

yes, it's a local vendor, I think my boss will manage to.. :)

Darth Android

@NaN That link might work, though I'd like to see a picture of the back to make sure it has the onboard ports since it doesn't list them

@allquixotic Support for VirtuMVP?

NaN

there <--------- :)

allquixotic

9:05 PM

@DarthAndroid yeah -- LucidLogix Virtu MVP is the sort of "vendor-neutral" equivalent of Nvidia Optimus

Darth Android

What does Virtu MVP or nVidia Optimus come into this at all?

allquixotic

it's a little ASIC soldered onto certain motherboards that lets you render on one GPU and use a different GPU's framebuffer as the output

@DarthAndroid he wants to use a total of 3 monitors, and neither the mobo nor the GPU he has supports 3 monitors

Braiam

how is something that isn't even a number is here???

Darth Android

If the cpu + motherboard has iGPU, then you just need to tell the BIOS to not shut the onboard ports off

the iGPU will power the onboard ports, and the dedicated card will power its ports

allquixotic

@DarthAndroid yes, but then (1) you completely wasted the purchase of a discrete graphics card, and (2) if the iGPU doesn't have 3 ports, you don't have 3 monitors

@DarthAndroid incorrect -- not without something like Nvidia Optimus or LucidLogix Virtu MVP

Darth Android

9:08 PM

@allquixotic Pretty sure my motherboard has neither of those, let me check

allquixotic

an operating system normally has the concept of a primary framebuffer, meaning that it has to pick one graphics card to use as the output... at best, you can use something like Linux and have multiple framebuffers (separate X servers for each card), but you won't be able to move your mouse between screens or drag windows across screens

on Windows you can't even do that, because there is only one primary desktop, and you can't just say "I plugged in an additional graphics card and monitor, now give me a new desktop over there"

copying rendered frames from one GPU to another is typically done using some kind of on-motherboard chipset

NaN

@DarthAndroid see the link above, I edited

allquixotic

AMD calls it Enduro, for moving frames from a discrete Radeon to an on-chip APU Radeon

Nvidia calls it Optimus, for a similar purpose

and then there's LucidLogix, which sucks, but purports to do the same but is vendor-neutral

Darth Android

Whelp, apparently my mobo has Lucid Virtu. I stand corrected.

NaN

oh lol behind there is just 1 DVI and 1 VGA

allquixotic

9:11 PM

@DarthAndroid you must be using it without even realizing it then, because ordinarily you can't just plug monitors into the motherboard and other monitors into the discrete card and have them all just work as one unified desktop environment

NaN

need to spend more $$ :)

Darth Android

@allquixotic it's not unified.

allquixotic

@DarthAndroid meaning...?

Darth Android

They each show up as a separate display

allquixotic

you can't move your mouse between them? ah

Darth Android

9:13 PM

I can move my mouse between them

It's the same as putting an ATI and nVidia card in the same computer

allquixotic

if you can move your mouse and drag windows between them, then they're all part of the same framebuffer, which is impossible without something like LucidLogix or Optimus

NaN

and thanks guys

allquixotic

software doesn't do that. that's your Virtu.

I mean, unless you have all the monitors plugged into the same card (all into the mobo, or all into the discrete GPU)

Darth Android

No, I had onboard + nvidia

It's just not listed as virtu in the bios

just a "Enable IGP Multimonitor" setting

NaN

vendors could really mention: max screen supported: N

allquixotic

9:25 PM

@DarthAndroid Enable IGP Multimonitor is Virtu MVP

Darth Android

with a description of "Keep onboard graphics enabled when dedicated card is detected"

allquixotic

the people who write BIOSes appear not to have any clue what they're talking about

Darth Android

@allquixotic Yes, but unless you actually read the Virtu MVP marketing material, most will not assume that

I thought it meant "Enable integrated graphics processor multimonitor"

Braiam

@allquixotic you quoting Linus?

allquixotic

yes, it keeps the iGPU enabled when a dedicated card is detected. but it has to do quite a bit more than just have it be enabled in order for you to have a seamless desktop

Darth Android

9:26 PM

@allquixotic Ah, I figured windows handled that.

allquixotic

nope... it's actually quite an expensive operation to do in software

Darth Android

Ah.

allquixotic

you'd have to load gigabytes per second of data into RAM (via the CPU) and copy it between GPUs using PCI-E bandwidth... ouch

the Virtu MVP chip (or Nvidia Optimus, or AMD Enduro) grabs frames through the PCI-E pipe from one GPU and passes them straight to the other GPU, also over PCI-E, without even touching the CPU

Darth Android

I didn't realize it was pulling it from the nvidia GPU

I figured it was only sending the stuff for the nVidia displays to the nVidia GPU

allquixotic

@DarthAndroid well, Virtu MVP also has the capability of rendering applications on the Nvidia GPU

Darth Android

9:28 PM

and sending the stuff for the iGPU dipslay to the iGPU

NaN

I didn't mention it, but I was running those HP pc on win8, I may try from Ubuntu for fun, but I don't expect it to do the magic

allquixotic

it can get really complicated if you have monitors plugged into both graphics cards

it's a "render-anywhere, display-anywhere" scheme

graphics card X has framebuffer F(X) for monitors X1, X2, X3, ...
graphics card Y has framebuffer F(Y) for monitors Y1, Y2, Y3...
program X is rendered (OpenGL, Direct3D, etc.) using GPU G(X) OR G(Y) and depending on where the window is on the unified framebuffer, the results are emitted to either F(X) or F(Y)

Virtu MVP can also do rendering for any given program on two GPUs

so in theory you could have a Direct3D game, in windowed mode, spanning two monitors, one monitor plugged into the iGPU and one monitor plugged into the Nvidia card, and you'd have all kinds of complex transfers flying around

Darth Android

See

the thing is

I've been cross-rendering hardware accelerated games for several motherboards and such on wrong GPU

allquixotic

o.o

the Virtu MVP software lets you customize per-application where you want it rendered

Darth Android

Prior to this motherboard,, I had a GTX580(2 displays) and a GTX 280 (1 display), and no Virtu MVP or nVidia Optimus

allquixotic

9:34 PM

their hybrid rendering mode doesn't work particularly well for me, i tend to just let the beefy discrete card do all the rendering

@DarthAndroid were they connected in SLI?

Darth Android

No.

allquixotic

you must have had something, because ordinarily that just isn't possible

Darth Android

You can't SLI different nVidia GPUs, they have to be the same

Well

The thing is that when I drug HWaccel windows across the screen boundaries

they would flash for a moment

I always assumed that was Windows switching cards where it was rendering

allquixotic

something was going on, then, but I don't quite know what... I really, really don't think that's Windows doing it in software though

unless the Nvidia drivers do it in software

Darth Android

I wouldn't be surprised if the nVidia drivers do it in software

allquixotic

9:36 PM

I've had similar scenarios with AMD and it never worked like that until I got LucidLogix

Darth Android

or do it on the cards directly over the PCIe buss

But I've never had trouble running multiple GPUs with their own monitors

allquixotic

@DarthAndroid the latest generation of AMD Radeons got rid of the physical connector for CrossFire (their equivalent of SLI), and the cards talk to each other over the PCIe bus exactly like that to share workload and framebuffer

I've never been able to do that cross-vendor though without LucidLogix or similar

maybe within the vendor spaces it works because of software

but a software solution is going to be slower than hardware by a lot, hence the flickering

Darth Android

It was just a single-frame flicker

I've run GTX210(1disp)+GTX280(2disp), GTX280(1disp)+GTX580(2disp), GTX280(2disp)+Dual GTX580 SLI (1disp), and iGPU(1disp)+GTX580(2Disp) all without issue, though the last one was the only one where I had to change things in the BIOS, and the last two are the only ones that have ever been run on a mobo with Virtu

allquixotic

amd.com/us/products/technologies/amd-eyefinity-technology/Pages/…

Darth Android

Thankfully nvidia got rid of their 2 discrete display limit, so now I can just run all four displays off of my nvidia card

allquixotic

9:41 PM

that seems to be the software solution -- Nvidia probably has an equivalent

Darth Android

@allquixotic That's not quite the same

allquixotic

oh yeah, you're right, because Eyefinity makes all the displays appear as ONE monitor

Darth Android

Eyefinity is equiv to nvidia's Surround, and those are special because they can stretch a single virtual monitor across multiple displays

allquixotic

your start button is on the leftmost monitor in the bottom left and the taskbar goes all the way across XD

Darth Android

Yeah, I want independent displays

if I want something like that, I'll run the game windowed and just stretch it across all of them

allquixotic

9:43 PM

they're independent in some ways but merged in other ways actually

Darth Android

They're merged in all the ways I care about

Windows sees it as one

allquixotic

true independent displays would not have the feature of dragging windows across them, or the mouse; you'd have to use separate display servers

true merged is eyefinity

Darth Android

Does windows even do separate display servers?

allquixotic

no :P but linux can

Darth Android

I've always been under the impression that if you need more monitors on a windows box, you just stick another graphics card in

and never seen information to the contrary

allquixotic

9:48 PM

@DarthAndroid I'm 100% positive that the only reason it works in your case in the pre-Virtu setups is that there's a special software feature in the Nvidia drivers that enables it, but it's hardly cross-vendor or part of the OS itself.... though I can't seem to find any mention of what exactly that feature is

I'm sitting at a computer without any sort of Virtu or similar, and an Nvidia NVS 540 GPU and an Intel HD4000 iGPU, and plugging in monitors into the motherboard does absolutely nothing if the system booted up to the Nvidia card originally

conversely, if you boot up to the iGPU as the primary framebuffer, the Nvidia card's monitors don't work

that's how it natively behaves, unless you have that special Nvidia sauce that joins together Nvidia cards that aren't in SLI

Darth Android

That would actually explain some of the problems I had with iGPU that caused me to buy that GT210

Which was quite an interesting buy. "Sir, I need your cheapest PCIe graphics card. "Well have you tried our new XYZ for $definitelyNotCheapest?" "Sir, I just want one that turns on a single display for web browsing. Get me a list of everything you have under $40 that meets that."

allquixotic

4

A: Nvidia GeForce SLI with AMD chip

No. Your motherboard only needs SLI support if you want to run the graphics cards in an SLI configuration (which would only allow you to drive 2 monitors, possibly 3 if you're using nVidia's Surround technology). If you just want to have two cards operating independently and driving 4 independent...

NaN

10:08 PM

guys generally you'd recommend nvidia over ati I guess?

Canadian Luke

I feel good... Piped through the Close Vote queue, edited a bunch, left some open that didn't have a good close reason... I feel good

NaN

ATI is maybe cheaper

allquixotic

0

Q: What is the mechanism that allows for dual Nvidia card Framebuffer Combining without SLI?

First, let's make sure we're all on the same page: Assume that you have purchased more monitors than you have ports on any one of your graphics cards. For instance, if your graphics card has one port, you have two monitors. If your graphics card has two ports, you have three monitors. And so on....

@NaN ATI doesn't exist anymore

it's AMD

but yeah, I tend to use AMD Graphics Product Group (GPG) (formerly known as ATI) graphics cards, because, especially lately, you get more performance for your money

NaN

yep, but you have to make sure of those settings with processor :)) (handling or not the wanted numbers of ports)

Darth Android

@allquixotic I would tweak the wording on #3: "As if you had SLI'd them" -> SLI-ing in the nVidia world strongly implies that you're going to have multiple cards rendering alternate frames or parts of the same frame

same as ATI crossfire

The ability to have merged framebuffers is not associated with SLI at all, because it's available when cards aren't SLI'd

allquixotic

10:20 PM

@DarthAndroid edited

Anonymous

@allquixotic can you plz tell me why ATI brands some of its GPUs as "XFX"

allquixotic

@PatoSáinz that isn't ~~ATI~~ AMD branding anything; that's an actual company called XFX that is an AMD board partner. they buy the AMD ASICs (the bare graphics chipset) and stick it in a chassis, put a heatsink and fan on it, maybe customize the firmware and the appearance, and sell it.

Asus does the same, as does PowerColor, and many other companies

~~ATI~~ AMD and Nvidia don't, themselves, package up the consumer-ready product and sell it directly; at least not usually

Anonymous

is it the same with Dell+Alienware?

allquixotic

@PatoSáinz more or less, yes

Anonymous

what a crazy industry

allquixotic

10:23 PM

not really

Anonymous

CONSISTENCE, CONSISTENCE, CONSISTENCE DAMNIT

allquixotic

it lets AMD and Nvidia focus on what they do best -- design ASICs and write drivers

Anonymous

yea

Darth Android

@allquixotic And in case I might have missed something, the motherboards in question are a Asus P5W DH Deluxe / Core2Duo; Asus p6t6 WS Revolusion / Core i7 920, and ASRock Z68 Extreme7 Gen3 / Core i7 3770K

Only the latter has Virtu I think

though it is also the only one where I have done cross-brand multiGPU

allquixotic

AMD designs the ASICs and writes the drivers; they send the blueprints to Taiwan Semiconductor Manufacturing Corp (TSMC) for fabrication; TSMC chunks out GPU dies and VRAMs and sends them back to AMD; AMD assembles them into boards (with display headers etc) at a factory; AMD ships the nearly-finished boards to the board partners; the board partners put the plastic chassis and cooling solution on and sell it

Anonymous

10:25 PM

^and that's why AMD hasn't shiny hyped-up products like Apple

allquixotic

@PatoSáinz if "having shiny hyped-up products like Apple" involves second-sourcing every component in the computer, putting an Apple logo on it (which, by the way, is a FRUIT produced by nature; how the fuck they trademarked that I have no idea; Steve Jobs didn't invent fruit; nature did millions of years before he was a twinkle in his dad's eye) and selling it at 300% markup, I'm glad AMD doesn't have shiny products like Apple.

seriously cavil?

!!undo

Anonymous

lol

Anonymous

@allquixotic i'm not saying having shiny shit is good

Anonymous

yea I'm glad AMD doesn't have shiny products, but I'd like to see AMD's brand unified, first, by getting rid of ATI comepletely and then by removing XFX and such bullshit from their products

allquixotic

the idea that Apple makes all their own hardware is extremely incorrect

Anonymous

10:28 PM

it doesn't

allquixotic

they openly say that their components are made by other manufacturers -- Intel, AMD, Nvidia, etc

Darth Android

@allquixotic Trademarks are not patents/copyrights :P

Anonymous

it would be overly naïve to think they do

allquixotic

their monitors are made by Samsung IIRC

Darth Android

It's the same way that the Beatles have trademarked "Apple" with respect to music

Anonymous

10:28 PM

yes

Anonymous

and their glass, by corning

allquixotic

@PatoSáinz I don't think it makes sense to remove the board integrators from the picture, because they make price competition and small feature competition fairly interesting

Anonymous

@allquixotic right

allquixotic

they can also differentiate on things like the quality and duration of the warranty

and heck, let's face it, product appearance -- because some people love staring at their GPU instead of using it

Anonymous

ugh, modders

allquixotic

10:30 PM

then there are some board integrators that ship their cards pre-overclocked to "safe" levels with improved cooling solution to make it safe

Anonymous

their fancy GPU racks and water cooling

Darth Android

@allquixotic To be fair, this looks sweet: 120hz.net/attachment.php?attachmentid=960&d=1375727836

And if positioned right, no reason you can't do both.

Anonymous

i hope all watercooler-ers's watercooling pipes disintegrate and get all their electronics wet

Darth Android

Errr

You realize it's non-conductive right?

Anonymous

@DarthAndroid ask any guy whose watercooling system has broken

Darth Android

10:32 PM

@PatoSáinz I can, I had mine crack

Anonymous

it doesn't matter the water is distilled clean if it still gets impurities

Darth Android

Leaked all over the PSU

It's not distilled

Anonymous

@DarthAndroid then you are a one percenter

Darth Android

well, not just distilled

Anonymous

what else then?

Anonymous

10:33 PM

►

Darth Android

There are nonconductive/antifungal/anticorrosive additives that most coolants come with, or as a stand alone if you do just want to add it to distilled water

allquixotic

@DarthAndroid distilled and deionized? :P

Darth Android

here ---> Anti-corrosive towards Copper, Aluminium, Brass, and Nickle; Antimocrobial to prevent algae, and non-conductive. And it GLOWS.

So, I wouldn't pour that stuff into my PSU, but I'm not going to freak out if I notice it's been leaking into my PSU

Anonymous

i think the glowy part sold it for me

allquixotic

if I had a nickle for every time someone said Nickle when they meant Nickel...

4

Darth Android

10:38 PM

Just shut down, drain the system, replace the cracked piece or replace the system with an air cooler, let the PSU dry, and then restart it.

Bob

11:10 PM

o.o

JohnMerlino

Hey all, I am reading an embedded c book, and it talks about memory. It says "The Standard 8051 can be described as an 8-bit microcontroller with a 16-bit
address space." I understand that the 8-bit microcontroller means that it can move 1 byte at a time across its data bus. But what does it mean 16-bit address space?

Jimmy Hoffa

Just because every now and then - no one can avoid watching something like this and just thinking 'whoa'

►

allquixotic

@JohnMerlino do you know what an address space is?

JohnMerlino

@allquixotic the total number of addresses in RAM?

where 1 address corresponds to 1 byte often represented in hex like address "0x00"

allquixotic

@JohnMerlino the total number of possible addresses of a memory location, not the total number of actual addresses

the total number of actual addresses is determined by the capacity of the RAM

I'm assuming that we're talking about physical addresses here, not virtual addresses

JohnMerlino

11:16 PM

yes physical address

allquixotic

virtual memory is a whole separate thing, an abstraction layer around physical memory

JohnMerlino

this is embedded c

using the 8051 microcontroller

"possible addresses"?

allquixotic

@JohnMerlino yes -- meaning that the hardware can't address more than 2^16 distinct memory locations, each memory location referring to 1 byte

but if you only have 2^8 bytes of actual physical memory, you can't just keep on accessing physical addresses beyond that, because those addresses don't map to any actual memory present in the system

so it has to know where to stop

ChatBot John Cavil

HAMMERTIME!

allquixotic

basically it's using a 2-byte unsigned integer to keep track of physical address pointers, but not every single one of those possible 2-byte unsigned integers is going to point to a valid memory location, when the system has less than 2^16 bytes of physical memory

that's the distinction between possible addresses and actual -- there are 2^16 possible ways to vary the 0s and 1s in a sequence of 16 bits, or 65536 possible ways, which correspond to the integers 0 through 65535, which are the possible indices into physical RAM address space

but if you don't have 65536 bytes of physical RAM, then if you ask the MMU to read or write memory at a memory location that is beyond the maximum capacity of RAM, you'll probably get some kind of hardware exception

Journeyman Geek

11:22 PM

@Bob I almost never run it on a desktop ;p

Bob

and 8-bit (or the bitness of the CPU) refers to the general purpose registers

JohnMerlino

@allquixotic thanks for response

allquixotic

if you were programming against physical memory addresses in embedded C and sizeof(int*) == 2 (2 bytes or 16 bits), you could technically dereference a pointer assigned to any valid 16-bit unsigned integer, but of course, you will get a hardware exception if RAM doesn't go that far

presumably the platform tells you how big RAM is and you just make sure in your code to never go past the maximum index

Bob

@allquixotic wiki.osdev.org/Detecting_Memory_%28x86%29

JohnMerlino

THat was a good explanation, I actually understand it now

Journeyman Geek

11:26 PM

morning

allquixotic

@Bob he's not using x86

but yeah that's the general idea

Journeyman Geek

meh

Bob

@allquixotic ya, and the older chips use on-chip memory anyway

well, a lot of them do

allquixotic

@Bob having no MMU is a scary thought

Bob

@JourneymanGeek did you get bitten by the werecat too?

allquixotic

11:27 PM

one wrong memory read and you crash the system

Journeyman Geek

@Bob: I'm in disguise.

3

also, iSCSI is confusing :/

allquixotic

@JourneymanGeek could be worse -- iSCSI over HTTP+TLS over IPsec over GRE on a specific VLAN with dynamic routing (RIP)

please don't try to make any sense out of that

iSCSI with UPnP discovery and automounting O_o

@Bob by Tables and @Journ HALP

2

Q: What is the mechanism that allows for dual Nvidia card Framebuffer Combining without SLI?

First, let's make sure we're all on the same page: As a bit of background information, please know that when you start up an operating system -- any operating system -- the BIOS (or UEFI GOP) tells the operating system which graphics adapter should be used as the primary framebuffer. The primary...

windows multiple-monitors gpu

well really I'm not trying to solve any particular problem, I'm just trying to learn

LearneymanGeek

Journeyman Geek

@allquixotic: naw, more the naming requirements.

oh, I'm actually pondering setting up a iscsi share ;p

allquixotic

on raspberry pi?

PLZ NO

Journeyman Geek

>_>

maaaaybbbeeee

allquixotic

11:31 PM

!!tell 12708227 no

ChatBot John Cavil

Journeyman Geek

well, not really. Probably not. Certainly not on OS/2

allquixotic

.........

I hate you

Journeyman Geek

(I'm using a spare laptop for this for now)

allquixotic

OS/2 and Raspberry Pi and iSCSI... yeah.... I think I'm leaving now

Journeyman Geek

11:32 PM

lol

naw, not the raspi ;p

I'm looking at a howtoforge howto and "The target name must be a globally unique name, the iSCSI standard defines the "iSCSI Qualified Name" as follows: iqn.yyyy-mm.<reversed domain name>[:identifier]; yyyy-mm is the date at which the domain is valid; the identifier is freely selectable. " is the bit I'm confused about.

My guess is as long as its unique and in that order I should be fine, but I'm not entirely sure

well, I can just use one of my domain names, but yanno.

Bob

iqn.2013-12.com.lupinenet:myiSCSIdrive

I guess?

Journeyman Geek

yeah

allquixotic

gah, that's annoying... it should just be something like iscsi://yourdomain:driveindex/folder/folder/folder

Journeyman Geek

but does the system actually need to have a record then?

@allquixotic: iSCSI is wierd

and properly set up can be used to boot windows ;p

allquixotic

-_-

Journeyman Geek

11:37 PM

no, I am not doing it ;p

allquixotic

oh no, Clippy just showed up in chat

Journeyman Geek

;p

allquixotic

I wonder if I can get Clippy to show up on Cavil's Firefox and have it mess up his code

!!say I'd like to trigger the regex for clippy that steals your focus

ChatBot John Cavil

I'd like to trigger the regex for clippy that steals your focus

allquixotic

!!info

ChatBot John Cavil

11:38 PM

@allquixotic I awoke on Sun, 15 Dec 2013 05:00:19 GMT (that's about 2 days ago), got invoked 30 times, learned 48 commands

Journeyman Geek

@allquixotic: once its set up, iSCSI behaves exactly like a hard drive

Bob

@allquixotic I'm pretty sure it sets values/sends via the DOM, not keypresses

also, fuck off Flash installer

I don't want your update anyway

allquixotic

true, true

Bob

stupid piece of crap wants me to close FF first

even though I don't have any Flash applet (or the plugin container) running

Clippy

Hello I need to run many VirtualBox, however I don't know what is the most important hardware that I should focus on, let's say that I'll use non-gpu (no games, no graphic software, no 3D) but only applications that perform mathematical operations on these virtualbox. I want to purchase a PC, but don't know on which hardware I should invest my money, do you have any ideas ?

Bob

11:41 PM

stupid stupid hardcoded detection

and it's always on top

grrrr

@Clippy you have excellent timing (cc @allquixotic)

Clippy

haha

Bob

@Clippy For virtualisation, first thing you want is more RAM

any RAM you assign to the VM cannot be accessed by the host OS

e.g. if you assign 2 GB, the whole 2 GB will be used regardless of the actual usage within the VM

depending on how many VMs, I'd recommend at a bare minimum 4 GB for the host + 2 GB per VM

also depends on what each VMi s doing, etc..

Clippy

Thanks, okay, and what about the GPU ? It isn't really important in my case right ?

Bob

@Clippy I learned a while ago that rule 1 of VBox is don't do anything requiring graphics acceleration in it

you'll also definitely want a CPU that supports hardware virtualisation

I'd pretty much always recommend Intel, so make sure it has VT-x (AMD has their own name for it)

almost all CPUs have that now anyway, but do check

if you're doing anything compute-heavy (lots of calculations), you might want a more powerful CPU

a VM is one of the actually quite rare cases where more cores can help

so go for at least quad-core if you can

Clippy

Thank you very much

Bob

11:46 PM

top of the range i7 won't go wrong :)

if you're running a lot of VMs, you might want to look into a server CPU

but those typically cost far more, and need more expensive motherboards and RAM too

Journeyman Geek

@Clippy: LOTS of ram, best cpu you can with as many cores as possible

Bob

at some point, I'd say just get an additional machine - don't try to squeeze too much out of a single one

Journeyman Geek

if you're only doing VMs and nothing else, consider esxi

GPU is useless here.

Clippy

Yes I'm also thinking about getting additional machines, Exsi looks interesting thank you, I'm a noob when it comes to Virtualization, but well, it looks noob friendly lol, thanks so much guys have a nice day/night

Transcript for