Also does anyone know about HTC release cycle ?

yours is a HTC?

My annoying friend wants to knwo when the next phone will arrive

@jokerdino naw Canvas HD

HTC phones are too expensive

Going to sleep, gotta wake up early tomorrow -_- we will continue our chat tomorrow ;p

@Gowtham cya

@jokerdino ROFL

no one knows :D

eh? Why are all those messages deleted now? (including mine)

@Jasper off topic banter

@jokerdino Ah, ok. I've spent quite some time in the SO php chat, and besides allowing things to go off topic a bit, they would have binned it rather than deleted it if things got too far out of hand. Of course, it's perfectly valid for there to be differences in the policies.

binning is for messages that are of some value.

i would have binned it but too lazy to find the bin

and binning invites people to the bin, IIRC

@jokerdino Not where I'm from. (and that's not an attack, it's an explanation of why I was surprised)

@Jasper no. it's perfectly fine to question how things are done

i mean, i am responsible to answer anyway

if i wanted to, i could bin it and delete it.

but then again, only mods can delete. not normal room owners. might be why they bin them in the first place

@jokerdino was just thinking that

ok guess that was a satisfactory explanation

@somequixotic (or anyone else willing to help) besides the modules running, is there any other conflict between virtualbox and vmware? I want to try using virtualbox instead now, but I don't feel like completely uninstalling vmware

@jokerdino Huh, never seen mass delete like that here before. Has that always been happening on this channel?

@DarthAndroid no. only this one time

(lol, what did I do to deserve clippy?)

@Jasper "I want to..."

ah

@DarthAndroid i been talking about my interest and didn't find it reasonable to leave it online

Ah, makes sense.

What errors do you see when your Ego malfunctions?

Lovely was to phrase things. :)

@Jasper well... they both do weird stuff with the networking stack too, creating some kind of bridge device usually

it's really inadvisable to have two virtualization solutions installed on the same box, period... unless one or more of them are entirely userspace and do nothing with the kernel, like pure software qemu

you think you have problems with vmware now, ha ha, try seeing the problems you have with vbox and vmware stuff mixed up on the same machine

:( My number 1 question is marked as a dupe!

um... that's probably OT

@CanadianLuke meh, remove the 'memory/RAM' part :P

other question has no accepted answer and user never returned to site...

might could stand a merge, but meh

@somequixotic I missed your message and just went ahead and installed it. It seems to work alright so far. It helps that we don't really use the virtual networking much (the reason for the VMs is the lack of good bookkeeping software for linux). Now I'll just hope it doesn't blow up until the next major maintenance and then we'll get it all figured out properly.

Thanks for your help and insights, though

@tapped-out: overclocking update: The GPU malfunctioned (but recovered) while I was playing Darksiders. It turns out 720/950 wasn't really stable. I've pulled back the clocks to 710/930.

@Tanner Think my question would win? Lol

hm

maybe 700/910 would be stable?

I'll see. I'll keep playing at these clocks and see if it works out. No lasting damage, though...

@Tanner Flagged for merge, see what (if anything) happens

that's good, especially since ya can't replace laptop GPUs easily

Exploring more GitHub features. It's pretty fancy.

github.com/benrichards86/Verify/releases/tag/v2.3.1

efff

the USB hub in my monitor quit :/

@tapped-out I guess I will have to lose some of that performance benefit. However, those clocks are pretty high to begin with, and it's definitely better than the stock 550/800. Given the age of this laptop, any performance gain is welcome. :)

After all, it is just a Mobility Radon HD 5650, and not something like a 5870...

it's not that old...

and plus, it's not a ticking timebomb

(like the one in mine)

This system is about three years old, and it seems to be marginal playing modern games. I haven't started Darksiders II, which is more taxing on the hardware than the original. I suppose it will hold up from the benchmarks I've seen, but it's almost time to upgrade. I'm hoping to squeeze as much performance out of the GPU as I can...

Darksiders is running fairly smoothly, hovering around 40-50 fps, but it does seem to be choppy at times. At least I'm getting a satisfactory experience, even if isn't optimal.

@tapped-out How old is your system?

NFS Hot Pursuit seems to be worse in complex scenes, occasionally getting as low as 15 fps, but the overclocking has mitigated this somewhat.

Mobility Radeon HD 5650 benchmarks: notebookcheck.net/ATI-Mobility-Radeon-HD-5650.23697.0.html

@DragonLord it's a tx2000 (2007-2008, can't remember), with one of the ill-fated NVIDIA cards

That laptop, among others, was the subject of a class-action lawsuit against NVIDIA over overheating and failure.

haha yep

i didn't buy it, it was given to me

For reference, I'm running an HP Pavilion dv6z-3000 with 2.0 GHz quad-core AMD processor.

The monitor is an average 1366x768, and the game is running at 1360x768 (the highest supported by Darksiders, losing a few vertical pixels).

Darksiders: store.steampowered.com/app/50620 - got it and Darksiders II while it was on sale this past weekend - spent less than $25 on more than $110 of content, including all DLC packs for Darksiders II.

ah, hence why you can get playable framerates out of a not-so-cutting-edge hardware; you're not driving a 1080p display

Had it been 1080p, there's no way the 5650 is getting anywhere near 30 fps. Darksiders II gets only 21.1 fps at "1920x1080 2/4 Shadows, No Ambient Occlusion, AF Low AA".

notebookcheck.net/ATI-Mobility-Radeon-HD-5650.23697.0.html

@DragonLord thats what i was thinking, having a "lower" res native screen can be helpfull when it comes to 3d games requiring much gpu processing.

@DragonLord It's a shame that DSII is such a terrible port though.

when you overclock that Mobility... it's about the same as the card in my desktop right now o.O

@tapped-out: 400 VLIW5 shaders with stock memory/core clock rates of 550/800 MHz, currently overclocked to 710/930.

I am finding that the "new" wizz bang faster desktop gpu cards, none of them have magic clock speeds that are way over , even my old GPU. but the ammount of "Shaders" or processing units things, is what really seems to get the speed going.

@Psycogeek hence why texture detail hasn't continued to increase, yet shader complexity goes up by a lot :P

@DragonLord GT430, 96 CUDA cores, 1GB GDDR3, 700/1400/800 graphics/processor/memory

they can't increase the clock rate or add more ROPs or more VRAM, etc. to make textures sharper and sharper and sharper (the perceptual benefit would be minimal after a certain amount, anyway), but they can build more shader cores and divide the screen into even more tiles (squares) for processing shaders, SSAO, etc

same issue on CPUs. clock rate has hit a physics wall due to thermal dissipation and power requirements

@somequixotic does that mean excessive texture/video memory (when buying a new desktop gpu) is not going to be as nessisary till something changes big in another area?

@Psycogeek having lots of VRAM helps if you have multiple very large monitors, like 2 or 3 monitors at 2560xwhatever

but if you are only playing the game on a single screen in fullscreen mode, your framebuffer will be reduced to the size of one screen

That's going to give similar pretty performance. Too bad memory bandwidth is so limited on the Mobility Radeon HD 5650: 1 GB of dedicated DDR2 (!) memory.

so it also depends on whether you're running "fullscreen maximized" (window is composited by DWM) or "pure fullscreen" (framebuffer is resized to just the window the game is on, the other monitors go black)

they could fit higher res textures in VRAM, sure, but they wouldn't be able to transform / manipulate / shade them fast enough to give you 60 FPS, that is why texture quality hasn't continued to explode exponentially

it's more of a processing limitation than a framebuffer capacity problem

@somequixotic Look at the $900 AMD FX-9590: 8 integer cores, 4 FPUs at 4.7 GHz with an astronomical 220W TDP--and yet it's not catching Intel.

At least AMD's trying...

The high TDP is because of the high VCore: up to 1.912 V.

with a GPU you have to think of things like shaders as essentially parallelizable, and things like higher resolution / higher FPS as essentially requiring a higher clock rate... clock rate isn't going to get much better, ever, until we break out of the digital silicon based integrated circuit paradigm of computing

that is, "serial" performance is stuck in a rut, but "parallelizable" performance can continue to scale pretty much unhindered provided that you have enough fans and your PSU has enough watts to take care of the added cores

AMD could really use a better semiconductor process. Intel is on 22nm, but AMD is still using 28nm and 32nm processes.

some parts of graphics processing are serial and some parts are parallel; GPUs are designed to do mostly the parallel parts well, and each generation does them a bit better than the last, but the serial parts won't improve any more

@DragonLord the increased process size of GPUs is due to their increased heat dissipation -- there are more transistors on a GPU and the die size is often larger, and the TDP higher, so if you take the same amount of transistors and shrink them, you're going to get even more quantum tunneling and voltage leakage and a lower thermal ceiling... and then if you try to add even more transistors the whole thing gets even harder

@DragonLord size has not change the clock rates so much, as just requiring less power to make the same switching occur. so i have not seen where smaller is faster, it is just less power to do the same switching?

GPUs will pretty much always be a ways behind CPUs in process size because the R&D needed to make GPUs smaller is a lot harder than the R&D to make CPUs smaller

My concern is the 220W TDP of the FX-9590, not the clock rates.

This really could be a more reasonable 150W or so at 22nm...

@DragonLord not sure about AMD's CPUs, but at least their GPUs are built by Taiwan Semiconductor Manufacturing Corp (TSMC), which is the same company that builds all Nvidia GPUs, as well as Qualcomm's Snapdragon SoCs for smartphones and tablets, which are one of the major ARM mobile GPU+CPU cores

@DragonLord and if they are cramming 1.9V in, they are just selling you overclocked stuff without having to overclock. overclocked and overvoltaged, while that works i dont define it as a real speed increase :-)

TSMC is known for its high volume output, its custom fab software, its low yields in the early stages of production of a new fab node, and being about half a step behind Intel's smallest available fab

AMD FX CPUs are fabricated by GlobalFoundries.

but they still make the three best GPU brands in the world at TSMC -- the two fastest, AMD and Nvidia, and the lowest power, the Qualcomm Adreno (based on technology they bought from ATI years ago)

@Psycogeek Smaller is faster because less power per performance means they can get more performance from a given amount of power. IIRC PCI-E specifications dictate a maximum power draw of 300W if a company wants to get a card certified

So being able to consume less power means the manufacturer can do more with 300W

That only really applies to the high-end cards though.

@DarthAndroid and the smaller you get, the less true that is... because of quantum tunneling and voltage leakage and heat

what we're starting to see is that each successive die shrink provides less of an improvement overall than the last generation, because they are fighting all kinds of constraints at those sizes that they didn't have to fight before

Aye.

@DarthAndroid The AMD Radeon HD 6990 had an Unlocking Switch which would bump clock rates considerably and increase the TDP from 300W to 375W, while enabling access to an unlocked, flashable VBIOS.

It's also why nvidia's dynamic clocking is interesting. It can detect when the card is running under the rated TDP and boost the clock rate

@DragonLord the HD7970 is unlocked by default, and it has dual VBIOS support; the first VBIOS is read-only and hard-wired onto the board so it can never be flashed (prevents bricking), but the second one can be overwritten by the user to flash a custom VBIOS, then you have to flip a hardware switch to use the user BIOS

it's neat -- if you accidentally flash a bad VBIOS you just flip a switch and the card is un-bricked

@somequixotic merly a failsafe bios :-)

i have one

Also, when you say unlocked... do AMD cards not allow you to adjust multiplers / vcore by default?

@somequixotic I corrected that to HD 6990, not HD 7970.

@DarthAndroid not sure about lower end, but the high end ones (always?) have

there might be some hardware limit on how far you can adjust them, but you can definitely use AMD's official Catalyst Control Center software itself to adjust the clock rate on a HD7970 without doing any kind of hacking

@somequixotic That's what I thought. Not sure what "unlocking" does then.

don't think you could really justify going beyond whatever limits they set though; that's just begging for a bricked card and a voided warranty

Oh, not on the stock cooler no

i run a completely stock system with stock clock speeds; I have "above average" air cooling in the case though so the ambient temp inside the case is really good, so the GPU doesn't get very hot even under heavy load at stock speeds

Justification for OCing a GPU is the same as for a CPU though :P

i feel that stock speeds make the system last longer, and I have no interest in laying down money for a new system any time soon, so i need it to last a long time

my next upgrade will probably be to Haswell-E, due out maybe in 2015 if we're lucky, and the next major iteration of Nvidia (yes, I've decided I'm switching away from AMD with my next purchase) after the GTX TITAN

@somequixotic way back in time, GPUs had massive leeway for overclocking by the user. a person could say that the parts were highly overrated for the job, and therefore impervious to error and fail. Now things come Pushed to the limit, and there is less space left for increases. Pushed by the people applying the chips

@Psycogeek partly due to the constraints at smaller fab sizes, and partly due to aggressive binning

@somequixotic Depends on your hardware. My Sandybridge is OC'd more than 30% of its base speed and has been for some time. It's no trivial amount of boost that you can get sometimes

conservative binning = ship a card with factory clocks significantly below the red line (the area where the card might start to fail or display artifacts)

aggressive binning = ship a card close to or right on the red line, and only bin it down to a lower card if it's significantly more unstable than the mean

the only conservatively binned chips these days are server/enterprise/workstation parts... they intentionally clock them down (and raise the price :)) because businesses want things that Just Work (TM) for years with minimal risk

@somequixotic The whole Core series from Intel is known to be really conservatively binned.

(even the consumer procs)

They're locked at low speeds to keep the TDP down

Hello, everyone!

@DarthAndroid and yet, many of them are multiplier locked

multiplier locked and conservatively binned = a waste of good chip :D

@somequixotic Yup. That's why I buy unlocked.

The chips with unlocked multipliers should easily hit 4.2-4.5GHz with air cooling

is it irony? I don't know if it qualifies under that extremely hard to understand concept (irony :)) but AMD's unlocked consumer cards are aggressively binned to drive up stock benchmark scores, while Intel's CPUs are conservatively binned to reduce support requests and warranty claims, and locked to prevent people screwing with them

@somequixotic mere few years ago, intel boards would not even allow overclock. but there was no forum for them either :-) they just always worked

I came here looking for some advice about upgrading the CPU of my Dell Precision 390 Core Duo.

I've never upgraded a CPU before.

I'm not sure where to start looking.

@BGM do you have the exact CPU model number?

The Core Duos are fairly old

@BGM what would you estimate the value of the system currently to be? how much is it worth?

you should do a cost/benefit analysis to determine whether it's more economical to buy a whole new system (with all newer parts), or upgrade the existing one and make it a little bit faster for a little while longer

^ Depending on what you're wanting to do with the system, it may make more economical sense to upgrade the core components / get a new one.

The specs say it is 6320

I already have the system provisioned, and just ordered RAM for it.

It runs at 1.86GHz right now and I would like one a bit faster.

ok so it's a "Core 2" not a "Core" -- that's not a huge deal but the original Core series were really kind of bad

No, it's a Core Duo.

right. Core 2 Duo

ark.intel.com/products/29754

Yes, Core 2 Duo.

he could drop a 85?? thing in , used off of a reliable e-bay source for little funds?

@somequixotic Yes, I think that's the one I have.

IF you only upgrade the processor, I would look at something like this: newegg.com/Product/Product.aspx?Item=N82E16819116903

That's 2.93GHz - that's an improvement - about what I imagined.

well the problem is, without upgrading your motherboard (and then probably your PSU and RAM) you wouldn't be able to go beyond the Core 2 series of CPUs

$75 is about what I imagined, too.

it's an LGA775 socket

That said, depending on what you want out of the system, I'd evaluate doing a system overhaul

Well, I just want to recycle this box. It's for my secretary, and she isn't a power user.

I think the newest CPU you can buy for that socket is a Yorkfield

@BGM Ah, then I would probably just update processor then.

Core 2 Extreme, quad core, QX9xxx model number

And I don't want to spend very much; I just want to get her some speed and make the system seem like its a bit faster.

@BGM do you really need to even upgrade the processor at all for someone who "isn't a power user"?

a little software tweak here and there would do more for performance

heck, a RAM upgrade would do more for performance than clock speed

^

Yes, I do want to upgrade because I have the understanding that the cpu upgrade would make her software run a bit faster.

Cram as much memory as the system can support into the box

@BGM well unless she's utilizing the CPU, it really wouldn't

most stuff that casual users do is I/O or RAM bottlenecked

It would help with internet browsing, no? I can see SystemExplorer report high CPU usage with Firefox.

word docs, multitasking, checking email, having a bunch of tabs open in your browser, ... this is stuff that could run on a 15 W ARM processor just fine with 8 gigs of RAM and an SSD

@BGM disable flash

@somequixotic i would, double the memory correctaly, get the higher end duel, check the PSU for its capability first. it would make it a bit sparkier.

@DarthAndroid disable? nah. get a GPU with video decoding ;p then the CPU usage goes away

Core 2 era box probably doesn't have a GPU that does video decoding

No, disabling Flash would cause workflow issues.

@somequixotic Who says it's video? Flash ads are going to chew through CPU :P

The Core Duo supports dual monitors, which she is using.

No problem with flash ads whilst using Adblock Plus.

@BGM OS is Xp?

No, I just wiped the machine and installed Windows 7.

@BGM I would be curious what is causing FX to chew through CPU then. Pages don't use CPU while they're idle. Does (s)he have a lot of extensions?

There was an issue with a program called TheSage, but I disabled it at startup.

Firefox uses more CPU nowadays with smooth scrolling and its javascript engine being multi-core

So, back to the CPU - That link that Mr. Android suggested looks good: newegg.com/Product/Product.aspx?Item=N82E16819116903

I've doubled the RAM for the new system and waiting for it to arrive.

i think their new JS engine also compiles the JS down to native (JIT compilation) on page load, which uses a lot more CPU on page load unless the JS is already cached in the native cache

So, I can just pop out the old CPU chip and plug in the new one then? I've installed them before brand new, but I've never swapped one before.

I guess that means I have to clean off the old putty from the fan, too.

@BGM Generally, yes. If you're reusing the cooler, make sure to clean it and reapply new thermal paste

@BGM well if you've got the money to blow, no one's stopping you; but just don't be surprised if there's no noticeable performance improvement when you aren't swapping out the HDD, and you're still using DDR2, and you're still using the same GPU, etc

Yes, I'm stuck with DDR2 with that machine, as I understand.

@somequixotic Why wouldn't there be any increase in performance?

@BGM probably, if that's what is in there now. I would worry about the volume of memory rather than the speed though, if it's < 4GB

No, I hadn't planned on swapping the GPU.

@BGM people upgrading often have more problems with ram changes than cpu(of the same type) changes, because of so completly different ram modules. A manufactured computer system, could have a POS power supply in there. As power supplys age the capacitors age, and it could be a concideration.

@BGM obviously you would see higher micro-benchmark scores with a new CPU, but as I already said, most workloads that casual users take on are not bottlenecked by the CPU -- meaning, the CPU is more than fast enough for the workload

@BGM He's just commenting on the fact that the vast majority of what an average user does on a computer is not taxing of the CPU, and the CPU is not the cause of most of the wait time that a user experiences.

imagine a trip you're planning, where, you get on a bus that picks you up at your street corner, then drives 10 minutes to take you to a train station where you have to wait 8 hours for the train to get where you're going

If your secretary is using a specific software package (say, for accounting) that you know is CPU-heavy, then a new CPU is the right course of action

Would I be correct to understand that if there isn't enough RAM in a system that it causes an increase in the CPU usage?

now imagine they reduce the time the bus takes to an incredible 2 minutes, because the bus driver speeds

If it's a matter of "My system just feels slow in general", then a new CPU will not have a very big effect.

you're still waiting 8 hours for the train

@somequixotic I get you.

@BGM No, not significantly. It will increase harddrive usage.

the additional RAM will reduce hard drive usage which will measurably improve responsiveness of the system once it's booted and loaded

@DarthAndroid Which can have a effect on the feelings of the person using the system. But usually only if they have been whining about the "speed".

that is a definite win

I just don't think the CPU will affect anything unless they are doing scientific computing, or gigantic calculations in MS Excel that take more than 30 seconds to run

Okay, so I think you all have sensibly talked me out of spending money on a CPU.

I remember a recent article from something like TomsHardware or Anandtech that showed that installing a Solid State Disk in an old computer is the #1 way to revitalize an old box

I did order new RAM cards - higher speed cards to the total of 4GB.

@BGM Is it Win7 32bit or 64bit?

Win 7 64.

if you can deal with the reduced capacity and cost of an SSD, I'm pretty sure your Core 2 box would feel like a modern Core i7 box (to the secretary) with an SSD replacing the hard drive

Hmmm... I've never dealt with SSDs before.

But I've already done a fresh install of Windows 7 on this here HDD.

@BGM THen yeah, 4-8GB of memory and/or an SSD will be far, far better for that system than a new processor

they look just like hard drives, mostly, except that they are very expensive in terms of cost per gigabyte of storage, and they are a huge factor in responsiveness of the OS

Okay, then I think I'll be good to go with the RAM.

I don't want to spend the money on the SSD, I think.

Great. I think I'm decided. If I have questions, I know where to come. You all are always so helpful in this room. Thank you very much.

@BGM How much RAM does the system have currently?

@BGM ram would certannly be one of the things you could work out first, and make sure it is stable and functional.

@DarthAndroid Uh, actually, I am moving her from an Optiplex to the Precision so that she has a working machine whilst I provision the new one. The optiplex has more CPU, but only 2gb of RAM. The Precision will have 4gb of ram.

your next upgrade for the secretary, once you have enough money pooled for it, could be something like a low-end Core i3 Haswell CPU (modern architecture, low latency, low power, etc), 8 GB of RAM, and a small but fast SSD, maybe 120 GB... that system would be excellent for her; she'd love it, and it wouldn't be overly costly

@somequixotic - Ah, thanks for that. I upgrade computers around here probably every three years or so.

do you have a network storage server where you work?

I'll consider an SSD next time.

@somequixotic No, but I know I need one.

@BGM Ah, yeah 4GB should be a noticable improvement.

@BGM well.. the networked storage server would definitely alleviate any storage capacity concerns about an SSD, so you could just buy a bottom basement cheap SSD at a low capacity; low capacity SSDs are extremely cheap, and very fast too

Okay, on the subject of a NAS, what would y'all recommend for a network of about 25 machines?

@BGM in win7 with ample ram, 1/2 the dang system is stuffed into ram caches.

just tell them to sling their huge files on the networked storage, and let their operating system, browser cache, etc. run on the fast SSD

@BGM What's your budget?

Budget? It depends on what I can convince my superior of.

And about how much space do you need?

it would be really useful to determine how much space is currently being used by non-system files (actual user data) first

I figure I need enough space to do full system backups for each machine.

just a ballpark: is it 1 gigabyte? 10 gigabytes? 100 gigs? a terabyte? several terabytes? more?

Right now, I don't have any failsafe system except for Cobian Backup running locally on a few machines.

I am thinking I would like several terrabytes or more.

well you could get 8 TB of usable space with four commodity 4TB disks in RAID10

you could either buy a preconfigured system or build a commodity desktop computer with your own hardware and install the software etc

doing it yourself is cheaper but would require more time and expertise

^

You push a self-built system more

but requries more work

one of the nice things you can do is use an SSD cache

My file server has 16TB of storage, but there's a 120GB SSD sitting on top of that

I like handmade systems. Thanks to @allquixotic, I've built one or two already from kits.

for 25 users, with a modern caching storage protocol like modern Samba, I'd actually say that an SSD cache would be overkill for this

just a bunch of regular old rotating HDDs would be good enough imho, plus you get excellent storage density, and low cost per GB

@somequixotic and less possible complications.

decking out the storage server with lots of RAM would help tremendously though because the RAM acts as cache if it's not being otherwise used

Aye, and depending on what you use, you can easily add a cache later if you find you need the performance

the performance wouldn't even be needed if your network is slow, too, btw :-) so either you'd have to already have pre-installed a great network, or upgrade that too

@somequixotic Random I/O would benefit from the performance, even on a 100M link

Two things I've been considering: the Synology Disk Station (I don't know anything about it, but it was recommended to me: legionhardware.com/articles_pages/…

and FreeNAS: freenas.org

FreeNAS is what you want if you're building your own

I've not messed with synology, but I've messed with the equivalent product from NetGEAR

I'm glad to hear a recommendation for it, Thank you!

another recommendation for FreeNAS

FreeNAS, that is.

Those pre-built NASes are good if you don't want to get your hands dirty

personally I would use a bare Linux installation like Fedora or something and customize it from the ground up, but if you're not a Linux head, FreeNAS is help for the helpless in that regard ;-)

I don't know Linux, though.

I'm actually giving mine to my parents because I overestimated the performance when I bought it and it wasn't serving my needs.

FreeNAS has a web UI like the pre-builts for when you just want something to work, but still has BSD under the hood if you want to tinker and tweak it

best of both worlds.

I don't do hand-holding through Linux for new users anymore unless either (a) they're a coworker and I'm being paid to help them, (b) they're family, or (c) they are paying for my services, because I get really frustrated trying to guess what's on the user's screen, and I could type it 1000x faster than telling them what to do :p so in that respect, unless you can find someone to hand-hold you through a bare install, I'd say, go with FreeNAS

I don't blame you.

Personally, I'd rather stick with Windows because I know it. I don't mind, however, muddling through Linux a bit; I already do that with my Asterisk installation.

FreeNAS is a really good project/product though because it's very stable and has all the features you could want in a web interface so you don't have to dig around on the console unless you really hose it up

Ha ha, I like a web interface, and I'm good at hosing things.

Okay, what should I run FreeNAS upon?

I don't mind wading through Linux, especially since it's free.

And... I actually have a hand-holder.

@BGM You're probably looking for a low or medium end Haswell Core i3

@BGM FreeNAS is a special purpose operating system based on BSD, which means you just burn a FreeNAS CD/DVD, walk through the installer, and you're done, then you give the server an IP address and you go to the web interface and use it

8-16GB of memory

so you would want to run it on a reasonably capable desktop computer with good RAM and approximately 4 hard drives (more if you want to significantly over-engineer it)

gigabit ethernet would be the ideal way to network it with your router or other networking equipment

So I'd want a SSD for the brains and the 4 HDDs for storage?

@BGM that'd work, but I'd say the SSD for the operating system is optional

okay.

once the NAS is booted, it isn't going to be reading from its operating system hard drive hardly at all

it just loads its programs (service daemons) into RAM and that's it

@BGM If you can get the budget approved, I'd grab 4 HDDs for storage and a small (64Gb?) SSD for OS / cache

you could make the OS hard drive just a commodity bare bones 120 GB HDD, or 500 GB, or whatever is cheap

@somequixotic Those are about the same price as a small SSD here :/

though an SSD would last longer because the NAS's operating system hard drive doesn't do any significant amount of writes, and an SSD that never does writes is essentially immortal

whereas an HDD that only does reads will eventually wear itself out

You guys know your stuff.

:P idea being, go cheap with the OS disk (whatever it turns out to be); your primary cash layout is going to be in buying the storage HDDs

even the CPU and RAM and motherboard should be cheaper than your storage disks

@BGM for the data drives, I'd recommend configuring them with ZFS or BTRFS

I'm thinking to get something that could fit into my server rack.

ZFS on BSD is quite nice; does FreeNAS support that?

ZFS will be much less painful though.

@somequixotic I believe it does

@DarthAndroid awesome

I had to delete "s (some stuff) e (some stuff) x (some stuff) y" from that post above ^ because I was so excited to hear ZFS is supported by FreeNAS that I typed a lewd word lol

lol

Yeah, dealing with ZFS was so much nicer in ubuntu than mdadm + lvm2