|
Potato Salad posted:This was the good answer too. Most applications are drive agnostic per best practice now, so it's not too much trouble during install of games (especially Steam games) to switch to G: or whatever else you want to label the game drive. I mean the other reason is that I'm gonna have 30-40gb just sitting free on the first drive and it'll bug me for the obvious OCD reasons but I can live with that.
|
#
?
Dec 26, 2016 23:41
|
|
|
|
| # ? Apr 19, 2024 10:19 |
|
|
Alchenar posted:I mean the other reason is that I'm gonna have 30-40gb just sitting free on the first drive and it'll bug me for the obvious OCD reasons but I can live with that. You should have free space on a drive anyway to keep it working well.
|
|
#
?
Dec 27, 2016 00:22
|
|
|
Alchenar posted:I mean the other reason is that I'm gonna have 30-40gb just sitting free on the first drive and it'll bug me for the obvious OCD reasons but I can live with that. Reduce the volume size in computer management? Or learn to live with it. Windows does not like getting full, and SSDs require 10-20% free space to not blow up and die.
|
|
#
?
Dec 27, 2016 07:05
|
|
Cat Army 
|
You can always just adjust the over-provisioning and forget about it.
|
|
#
?
Dec 27, 2016 08:38
|
|
|
oohhboy posted:You can always just adjust the over-provisioning and forget about it. Windows still does not handle completely full filesystems well, you should definitely keep 5-10% free on an NTFS volume regardless of overprovisioning on the partitioning or firmware level.
|
|
#
?
Dec 27, 2016 10:01
|
|
|
Pretty good sale on the Intel 750 1.2TB (both form factors) on newegg.ca - 799 CAD comes up to 588 USD. Picked one up with the U.2 connector for my Ncase M1 build with an AMD Ryzen chip (if that doesn't suck...)
|
|
#
?
Dec 27, 2016 23:57
|
|
|
nielsm posted:Windows still does not handle completely full filesystems well, you should definitely keep 5-10% free on an NTFS volume regardless of overprovisioning on the partitioning or firmware level. BS. What is this based on.
|
|
#
?
Dec 28, 2016 00:49
|
|
|
redeyes posted:BS. What is this based on. Superstition. Or well, there are various FS structures that may get shuffled around in the background. The system partition will definitely also see a bunch of movement, if the various system registry hives can't expand you're in for a world of pain. You can also get funky user profile problems if space is too low. And then there's software updates, those also want space to move around with. Yes all of those only become real problems when free space comes near single digits in MB, but why take the risk? Suddenly something unexpected happens, causing a thing to blow up and eat all the remaining free space on your almost-full C: drive, and then those risks above suddenly become real. So call it superstition, or call it peace of mind.
|
|
#
?
Dec 28, 2016 01:09
|
|
|
As the file system fills up the ability of the filesystem to contiguously allocate space starts getting limited which could lead to fragmentation of the newest data. This was probably pretty noticeable with rotational drives, but shouldn't be as bad with SSDs. I used to tell clients that they shouldn't fill up SANs past 80% for that reason, but those were highly specialized use cases.
|
|
#
?
Dec 28, 2016 01:14
|
|
|
Some applications and services recreate the files, like settings and whatever, instead of overwriting blocks. Not sure what the reason for this is (--edit: for rewriting instead of overwriting), but when the filesystem is full, these files will break and end up at zero size. Some applications will start behaving weird because they suddenly reverted to defaults, other applications will break because they can't handle reading empty files, and obviously you don't want system stuff to do that. And other applications get hosed up because they're in a loop trying fruitlessly to keep writing out any data they want but can't (event log is one of these, a full disk will drag your boot sequence along god knows how long).
|
|
#
?
Dec 28, 2016 02:08
|
|
|
Browsing through that cesspool of PCMasterRace over on Reddit out of boredom. I noticed in one of the flairs that someone claims to run six SSDs in a RAID0. ... WHY?!
|
|
#
?
Dec 29, 2016 14:34
|
|
|
Combat Pretzel posted:Browsing through that cesspool of PCMasterRace over on Reddit out of boredom. I noticed in one of the flairs that someone claims to run six SSDs in a RAID0. People that post on pcmasterrace seem to have a lot of money and just enough knowledge to be hang themselves with some stupid configs. The guy you mention is the pcmasterrace stereotype in my head. I would not go there looking for technical consults or opinions as anything you get out of there is likely to be at least some degree of 'wrong'
|
|
#
?
Dec 29, 2016 14:44
|
|
|
Mostly looking pictures there, to see what fancy cases and keyboards there are, or certain builds (watercooling and such), which seem nice. I don't get most of that subreddit. While there's people with plenty of money, shoveling it into a stove, there's also a lot of people trying to ride that "master race" theme and post things that are "meh" at best. To me, too, the PCMasterRace shtick suggests going balls out on the PC gear, but within good reason. I can at least partially understand if someone goes two SSDs in a RAID0 (or better, RAID1) to eke out the absolute maximum of performance, even then the gain is rather minimal, but at six it becomes idiotic. Same with people sticking 64GB in a computer and then just playing games with it, of which most target 8GB, maybe 16GB.
|
|
#
?
Dec 29, 2016 15:47
|
|
|
If you don't have 24 SSDs in RAID0... https://youtu.be/eULFf6F5Ri8
|
|
#
?
Dec 29, 2016 17:47
|
|
|
Too bad they didn't post a benchmark of how a single SSD fares in these "tests". Also, if you want absolute maximum read performance, you go with RAID1, anyway. If a lot of threads generate activity and for some reason there happens to be a large bias towards just one SSD, because most of the data requested only resides on there (for whatever voodoo reason), it's a bottleneck. With a RAID1 the data is everywhere and any drive can service any read request. I guess if you want absolute maximum random 4K performance (for whatever reason), it'd still apply today, given that even these NVMe SSDs only attain these magic multi-GB/s rates on linear reads. But all that depends on the RAID1 implementation, anyway. Windows' software RAID1 balances based on thread ID only. A single thread hammering tons of requests, even async, will read from the same drive regardless. Not sure what Intel RST does. And anything else in the consumer range sucks anyway. Of course, I can see those PCMR people go for SAS controllers for no reason.
|
|
#
?
Dec 29, 2016 18:03
|
|
|
Why wouldn't you just set up a RAM drive for whatever game you want to play? The money you stuck into all those SSDs would cover a stupid workstation board and Xeon.
|
|
#
?
Dec 29, 2016 19:37
|
|
|
I feel the main reason why you'd want SIX SSDs IN RAID 0 is because you can show off your SIX SSDs IN RAID 0 to the internet. e. Maybe the guy had a bunch of old 32/64gb drives and wanted to use them somehow. Which is still silly since older SSDs as a whole aren't as reliable as modern ones, but when it comes to bragging you gotta make do. e2. Looked it up for fun, a RAM drive for GTA5 would cost about $360, not including the RAM actually needed to play the game. You'd have to move to an workstation/enthusiast platform to handle >64GB RAM. Actuarial Fables fucked around with this message at 20:58 on Dec 29, 2016 |
|
#
?
Dec 29, 2016 20:46
|
|
|
Combat Pretzel posted:Too bad they didn't post a benchmark of how a single SSD fares in these "tests". To be fair, the video is from 2011. So my guess is ~350MB/sec.
|
|
#
?
Dec 29, 2016 22:11
|
|
|
They mentioned (quickly) that "one SSD gives you an amazing 220MB/s access speed". The video was made to get views, though. I agree it would be nice to see how a single SSD compares to the array with these non-standard benchmarks, but we're probably not the audience they had in mind.
|
|
#
?
Dec 29, 2016 23:28
|
|
|
Nah, I meant timing the batch open of all apps.
|
|
#
?
Dec 30, 2016 01:45
|
|
|
IOPs get exponentially worse for SSDs the more are put into a RAID configuration so lol
|
|
#
?
Dec 30, 2016 04:46
|
|
|
Combat Pretzel posted:I don't get most of that subreddit. While there's people with plenty of money, shoveling it into a stove, there's also a lot of people trying to ride that "master race" theme What did you expect? Riding the sub theme is what Reddit does.
|
|
#
?
Jan 3, 2017 23:12
|
|
|
Help, I am not very good at drives and having difficulty. I bought a new 500GB 850 EVO and want to copy my current system over to it. The computer is an i5 3570 running windows 10 off a Sandisk Extreme 240GB. The current filesystem is MBR and I would like to convert to GPT in the process. I've used Macrium Reflect and followed these instructions: http://kb.macrium.com/KnowledgebaseArticle50151.aspx followed by http://kb.macrium.com/KnowledgebaseArticle50168.aspx Unfortunately when the process is finished the new drive will only boot if the old drive is also attached. If it's detached it just reports no boot media is present. I am pretty sure the problem is in the boot configuration not being on the new disk properly, but don't understand how to go about getting it right. Thanks for any pointers you can give! e: after lots of loving around I have got it to work with lots of barely-understood bcdboot bodging. knox_harrington fucked around with this message at 03:13 on Jan 5, 2017 |
|
#
?
Jan 5, 2017 00:36
|
|
|
So hey, how do WD Blue SSD drives check out for reliability these days?Apparently dell is letting them go for 20$ 250 for 20$ model number checks out from newegg Thoughts? Apparently it will let you order up to ten of them.
|
|
#
?
Jan 5, 2017 07:45
|
|
|
TheParadigm posted:So hey, how do WD Blue SSD drives check out for reliability these days?Apparently dell is letting them go for 20$ They're the SanDisk X400 in another case. Just FYI, though, people are already all over this. They're already trying to resell on eBay before even having the drive in hand.
|
|
#
?
Jan 5, 2017 07:53
|
|
|
BIG HEADLINE posted:They're the SanDisk X400 in another case. Just FYI, though, people are already all over this. They're already trying to resell on eBay before even having the drive in hand. I take it that's a good model?
|
|
#
?
Jan 5, 2017 07:54
|
|
|
Don't waste your time. All those orders will be canceled as it was a pricing mistake.
|
|
#
?
Jan 5, 2017 07:58
|
|
|
You're suppose to leave 10-20% of the space on a SSD free right? or was that for HDDs?
|
|
#
?
Jan 5, 2017 17:09
|
|
|
OhFunny posted:You're suppose to leave 10-20% of the space on a SSD free right? Right, for SSDs
|
|
#
?
Jan 5, 2017 17:27
|
|
|
Wilford Cutlery posted:Right, for SSDs I thought manufacturers had been already accounting for that for a while?
|
|
#
?
Jan 5, 2017 22:48
|
|
|
I just got a 500gb 850 Evo (my first real SSD after the failure that was my lovely hand me down 64gb SSD from kingston a few years back), and I connected it on my Z77A-GD55 motherboard. I'm not seeing these 500MB/s write speeds (more like 250) it advertised, and the Samsung tool says I'm using AHCI but I'm using a Sata 2 port, not a Sata 3 port. I'm a little confused, because HWINFO shows that I have the drive connected on a Sata 3 port, so I guess I'll have to check that it's actually on a Sata 3 port on the motherboard (apparently it's saying one of the EHDs I have is on a sata 3 port -> USB, which I don't fully understand, I have two USB 3.0 ports in the back of this mobo, and there's a USB 3.0 port on the front of my case). If I connect it on the Sata 3 port will I really get 500MB/s read/write speeds or is that an advertised theoretical maximum? Last question: I've been reading there's no difference between using a Sata 1 cable or a Sata 3 cable in terms of what speeds the cables can push out, other than if one uses a clip on it or not (the one connected to the SSD is using a clip, some other drives are using the old slip in red cables that fall out a lot), so if that's the case, what is the maximum rating of these cables? Just how fast can they go?
|
|
#
?
Jan 6, 2017 04:10
|
|
|
I run an 850 Evo on a Z77 board and I get ~480 MB/sec, so yeah, looks like you're on a SATA II port. Some boards also have add-on SATA III controllers by Marvell or Jmicron and they kinda suck, don't use them with an SSD, make sure you use the Intel SATA III port. As for the cable question, I've used cheapie cables that fall out occasionally and good ones with the clips, and never noticed a speed difference, although I'm sure it can happen. I have had a couple go bad and throw CRC errors, though.
|
|
#
?
Jan 6, 2017 04:24
|
|
|
OhFunny posted:You're suppose to leave 10-20% of the space on a SSD free right? Overkill these days. What you do need enough space for is expanding updates for Steam games and Windows services. Some games want up to 30GB to use as swap Like Payday 2 whenever a big update drops. You can fill an SSD up, but just don't do it on a system drive, but it is not done out of a wear and tear issue. The amount of punishment the good drives like Samsung can take are so high above the stated amount no normal person in a normal situation will ever encounter it. GreenBuckanneer posted:If I connect it on the Sata 3 port will I really get 500MB/s read/write speeds or is that an advertised theoretical maximum? Here is my Samsung 840 EVO SSD hitting that advertised theoretical maximum on uncached write and generally what you should be seeing in synthetics. So there is some truth in it. code:
|
|
#
?
Jan 6, 2017 11:20
|
|
|
GreenBuckanneer posted:I just got a 500gb 850 Evo (my first real SSD after the failure that was my lovely hand me down 64gb SSD from kingston a few years back), and I connected it on my Z77A-GD55 motherboard. I have the same motherboard as you and a 250gb Evo plugged into it. Make sure you're using one of the two white SATA ports. The other ones next to the white ones are only SATA 2. The slow ones are either black or blue. I'm at work right now, so I can't take the panel off to check.
|
|
#
?
Jan 6, 2017 15:15
|
|
|
TheParadigm posted:So hey, how do WD Blue SSD drives check out for reliability these days?Apparently dell is letting them go for 20$ I ordered two, and I'm happy to report that they shipped this morning.
|
|
#
?
Jan 6, 2017 16:25
|
|
|
canyoneer posted:I ordered two, and I'm happy to report that they shipped this morning. Shoulda ordered 10 to put in RAID whatever mode for the Reddit cred. Seriously tho that's insane if you actually get them (as in they don't realize their mistake and revoking the shipping label before it actually gets shipped).
|
|
#
?
Jan 6, 2017 18:04
|
|
|
I take it that the sort of NAND in SD cards is really low endurance? Or is there any reason why there aren't any slow but high capacity SSDs for archival or RAID purposes?
|
|
#
?
Jan 6, 2017 19:09
|
|
|
Combat Pretzel posted:I take it that the sort of NAND in SD cards is really low endurance? Or is there any reason why there aren't any slow but high capacity SSDs for archival or RAID purposes? SSD's would make bad archiving tools - they will lose the data stored on them fairly quickly after powering off.
|
|
#
?
Jan 6, 2017 19:21
|
|
|
Combat Pretzel posted:I take it that the sort of NAND in SD cards is really low endurance? Or is there any reason why there aren't any slow but high capacity SSDs for archival or RAID purposes? The NAND in SD cards is almost always low-grade TLC (bad endurance), connected to a controller much less sophisticated than a SSD's controller (ie probably not as good at wear leveling and low write amplification). High capacity in a SSD means you can easily get performance as a side effect, because so much of SSD performance is about parallelism. Individual flash die (*) aren't very fast on their own. It's not entirely wrong to view a single SSD as a RAID 0 of a bunch of flash die. * - technically, planes: most NAND flash has two "planes" per die, with each plane able to process commands independent from the other. A single NAND flash "chip" (the lump of plastic you see on a printed circuit board) usually has eight or more die inside, so there's a lot of parallelism to exploit here.
|
|
#
?
Jan 6, 2017 19:27
|
|
|
|
| # ? Apr 19, 2024 10:19 |
|
|
Bob Morales posted:SSD's would make bad archiving tools - they will lose the data stored on them fairly quickly after powering off. BobHoward posted:The NAND in SD cards is almost always low-grade TLC (bad endurance), connected to a controller much less sophisticated than a SSD's controller (ie probably not as good at wear leveling and low write amplification).
|
|
#
?
Jan 6, 2017 19:38
|
|