Intel x25-V

[WC Annihilus]

So I got 3 of the Kingston drives from the Amazon deal and am trying to decide on how I want to set them up, whether to RAID0 them all or use them separately for OS/apps and games. Those RAID numbers certainly look really nice, but did you actually feel any difference? Main reasons I'm considering using them separately is cause then I don't have to deal with the RAID OS thing boot time or the array maintenance, and there are people who say it doesn't make much of a real-world difference. Though without RAID the sequential write speed is kinda gimp, and also wouldn't have to deal with figuring out how to split my apps and games among the drives. Would be running off the ICH10 on my motherboard, probably partitioned to 100GB if RAIDed to leave some space for wear-leveling.

[SteveRo]

Are you going to convert to intel first?
I need to read up on this myself.

[Soulburner]

The synthetic benchmarks are great and all, but those kinds of numbers hardly transfer to real applications. I'm talking HDTach, HDTune, etc...

You'll use sequential writes when installing apps and games or writing a large video file during encoding, but do you spend all of your time doing that?

[WC Annihilus]

Are you going to convert to intel first?
I need to read up on this myself.

You mean the firmware update workaround? Yes that's the plan. Anyways, Anand posted his review of the X25-V RAID and it looks pretty good. Might help me make a decision
http://www.anandtech.com/show/3618/i...25m-g2-for-250

[jalyst]

So I got 3 of the Kingston drives from the Amazon deal and am trying to decide on how I want to set them up, whether to RAID0 them all or use them separately for OS/apps and games. Those RAID numbers certainly look really nice, but did you actually feel any difference? Main reasons I'm considering using them separately is cause then I don't have to deal with the RAID OS thing boot time or the array maintenance, and there are people who say it doesn't make much of a real-world difference. Though without RAID the sequential write speed is kinda gimp, and also wouldn't have to deal with figuring out how to split my apps and games among the drives. Would be running off the ICH10 on my motherboard, probably partitioned to 100GB if RAIDed to leave some space for wear-leveling.

Yeah this is what I'm trying to decide....
If improvements aren't going be noticeable 85% of the time then why bother, considering all the maintenance hassle w/a RAID set-up.

Biker pointed that review out to me, yet to read, hopefully it'll make the decision easier...

good-night.

[WC Annihilus]

Got my 3 drives a couple days ago and been playing around a bit with stripe size. So far have tried 16k and 128k, next will probably be 32k. IOmeter config files taken from the first post in this thread.

16k:


128k:

[SteveRo]

Both as expected and beautifull at the same time!
When will you get a 4rth?
Show us 4!

[Budwise]

so for 2 X25-V drives in RAID 0 the ideal stripe size is around 32-64 right?

[IanB]

so for 2 X25-V drives in RAID 0 the ideal stripe size is around 32-64 right?

No. It'll depend primarily on your workload (mainly small file access versus large file streaming), but the consensus seems to be moving towards smaller stripes with Intel SSDs for typical SSD usage where fast small file access is needed. Look closer at the results posted by WC Annihilus above. The 16K stripe is clearly better than the much larger one, isn't it? GulLars also posted some very pertinent comments attached to the Anand x25v in R0 review everyone's talking about right now.

[jalyst]

so it this crap?
http://www.tweaktown.com/articles/31...uide/index.htm

this was my understanding but the above seem contradict that
http://techreport.com/discussions.x/18653

[WC Annihilus]

Both as expected and beautifull at the same time!
When will you get a 4rth?
Show us 4!

As much as I'd love to have more, I don't really need it space-wise. Plus Amazon only let me order 3 at a time. If they ever drop really low again though, who knows

No. It'll depend primarily on your workload (mainly small file access versus large file streaming), but the consensus seems to be moving towards smaller stripes with Intel SSDs for typical SSD usage where fast small file access is needed. Look closer at the results posted by WC Annihilus above. The 16K stripe is clearly better than the much larger one, isn't it? GulLars also posted some very pertinent comments attached to the Anand x25v in R0 review everyone's talking about right now.

I plan to post 32K and 64K results later as well. PCMark is running on 32K stripe right now while I'm in class actually.

[Anvil]

The Intels works great at whatever stripe size but particularly well on small stripe sizes whereas others (e.g. the Vertex) works at their best using large stripe sizes.

Ever since I got my first SSD's I've used small stripe sizes and for my kind of workload (and for most benchmarks) 16KB to 64KB have been the ones to beat.

[WC Annihilus]

More results! Not sure which to go with in the end yet. Anyways, time to test 64k

32k:

[alfaunits]

IIRC, Intel recommends 32 or 64KB stripe for endurance reasons (as its block size is between 32 and 64).
It makes perfect sense - if you split it more than that, you are essentially hitting more than the needed number of drives. But it can improve random read/write performance.

The Intels works great at whatever stripe size but particularly well on small stripe sizes whereas others (e.g. the Vertex) works at their best using large stripe sizes.

Ever since I got my first SSD's I've used small stripe sizes and for my kind of workload (and for most benchmarks) 16KB to 64KB have been the ones to beat.

[Computurd]

a big big big +1 on that. if you are running 16k's you get uber performance, but uber wear as well. it also hampers NCQ ability to do write combining, which is a big help with write amplification.

[WC Annihilus]

GullLars made a few comments about the write amplification when someone inquired about it with smaller stripes and I'm curious what other people's thoughts are on it:

@GullLars

Have a question for you regarding advice you gave another regarding stripe size for a raid setup. I frequent the other forum you do as well but would like to know what your knowledge is of write amplification is. At the other forum Ourasi wrote that he found 16-32kb best, just like you mentioned. But the others said that is not good for performance and would destroy the SSD's quicker due to write amplification of having to do multiple writes with the write block being larger than the stripe size selected. I believe the sense of what you and Ourasi said in the other forum, but how does that effect the write amplification? Can you explain more on that?

P.S. I replied to another comment you made but anands comment system didnt take it dont know why so trying again. ;-)

There is no impact on write amplification from the striping. It can be explained pretty simply.
SSDs can only write a page at a time (though some can do a partial page write), and page size is 4KB on current SSDs. As long as the stripe size is above 4KB, you don't have writes that leaves space unusable.
With a 16KB stripe size, you will write 4 pages on each SSD alternating and in sequential LBAs, so it's like writing sequential 16KB blocks on all SSDs, and as the file size becomes larger than {stripe size}*{# of SSDs} you will start increasing the Queue Depth, but it's still sequential.

Since all newer SSDs use dynamic wear leveling with LBA->physical block abstraction, you won't run into problems with trying to overwrite valid data if there are free pages.

The positive side of using a small stripe is a good boost in files/blocks between the stripe size and about 1MB. You will see this very clearly in ATTO as the read and write speeds doubles (or more) when you pass the stripe size. F.ex. 2R0 x25-M with 16KB stripe jumps from ~230 MB/s at 16KB to ~520MB/s at 32KB (QD4). This has a tangable effect on OS and app performance.

Thanks GullLars!!!

I will try out the 16k stripe. I have 3 40gb intel v SSD's on the a ICH10R. Would 16k be okay for this setup?

By the way could you explain to me why they are confusing the write amplification as being a problem with these type of stripe sizes. You and Ourasi's explanations make perfect sense to me. Some actually even believed Ourasi and had questioned themselves. Maybe a problem of these synthetic benchmarks compared to real world usage as Ourasi mentioned?

Here is from the forum..

http://www.xtremesystems.org/forums/...ht=stripe+size

Something Ourasi said

"At the moment I'm on 16kb stripe, and have been for a while now, and it is blisteringly fast in real world, not only in benches. There is not much performance difference between 16kb - 32kb - 64kb, but the smaller sizes edges ahead. As long as there is no measurable negative impact using these small stripes, I will continue using them. The X25-M is a perfect companion for those small stripes, some SSD's might not be. Intel was right about this in my opinion. But I must add: This is with my ICH9R, and I fully understand that these stripes might perform pretty bad on some controllers, and that's why I encourage people to test for them selves..."

Response to his comment

"There are like 10 times more people saying that their testing showed 128 or more to be the best. Who do you think tested properly? Things like this will never be 100% one sided. "

What bothered me was this comment later which had me ask you the question earlier..

"Most SSD's have a native block size of 32KB when erasing. Some have 64KB. This is the space that has to be trimmed or "re-zeroed" before a write. If you write 1,024KB to a RAID-0 pair with 64KB blocks, with a 32KB stripe size, it will be 32 writes, requiring 64 erases. With 128KB stripes, it will be 32 writes, 32 erases. You'll effectively see a 30-50% difference in writes. This does not affect reads quite as much, but it's still usually double-digits. Also, with double the erase cycles, you will cut the lifespan of the drive in half."

I am interested in testing some stripe sizes now.. but i think i will use something more real world. Do you think Vantage HDD would be good test for stripe sizes since it uses real world applications. I dont like the synthetic benches

Regarding the "Most SSD's have a native block size of 32KB when erasing......" quote, this is purely false.
Most SSDs have 4KB pages and 512KB erase-blocks. Anyways, as long as you have LBA->Physical block abstraction, dynamic wear leveling, and garbage collection, you can forget about erase-blocks and only think of pages.
This is true for Intels SSDs, and most newer SSD (2009 and newer).

These SSDs have "pools" of pre-erased blocks wich are written to, so you don't have to erase evertime you write. The garbage collection is responsible for cleaning dirty or partially dirty erase-blocks and combine them to pure valid blocks in new locations, and the old blocks then enter the "clean" pool.

Most SSDs are capable of writing faster than their garbage collection can clean, and therefore you get a lower "sustained" write speed than the max speed, it will however return back to max when the GC has had some time to replenish the clean pool. Some SSDs will sacrafice write amplification (by using more aggressive GC) to increase sustained sequential write.

Intel on the other hand has focused on maximizing the random write performance in a way that also minimizes write amplification, and this either means high temporary and really low sustained write, or like intel has done, fairly low sequential write that does not degrade much. (this has to do with write placement, wear leveling, and garbage collection)

This technique is what allows the x25-V to have random write equal to sequential write (or close to. 40MB/s random write, 45MB/s sequential write). x25-M could probably also get a random:seq write ratio close to 1:1, but the controller doesn't have enough computational power to deliver that high random write using intels technique.

[WC Annihilus]

And 64k results. By the looks of things, looks like it's probably between 16k and 32k for me, with 16k seeming to be just a touch better though questionable on wear.

64k:

[alfaunits]

Well, if we have 4K stripes and write 8KB of data (or 32KB stripe and 64KB data chunks), instead of erasing a single block on a single SSD, we would erase a block on each of 2 SSDs. Thus doubling the erase count.
Does anyone see some flawed logic in this?

[GullLars]

I see you have been talking about me ^^

alfaunits: WRITE != ERASE
SSDs with a clean block pool does not erase on every write, but lets garbage collection handle the erases.

Can you give a quote or cite the source of the 32-64KB erase block on intel's SSDs?
It's possible you are refering to the largest chunks the SSD will write internally to each channel, the internal stripe size so to speak...

Anihilus, try doing a PCmark Vantage HDD run on the different stripe sizes, and maybe a timed launch-script. Anvil and Ourasi here have made launch scripts using windows apps and common programs.

Anvil here has experimented with different stripe sizes on x25-V/M RAIDs of 1-4 units from ICH10R, and I would ask him for input on the difference in real life performance.

I have been using 32KB stripe on my own 2R0 Mtron Pro 7025 since august 2008. Works as a charm for OS, and i have yet to experience any trouble. Though I have the liberty to be carefree, since they are SLC and rated for 100% duty cycle in servers for 5 years, with full warranty...

[WC Annihilus]

PCMark Vantage is in the upper left-hand corner of all my screenshots Though I realize I probably should've gone to the results tab rather than just the scores. As for scripts, I'm not really familiar with scripting or I would. Mind posting some? I'm a bit less inclined to go back and deal with re-imaging each time, but I might do it if I get really bored. If write amplification definitely isn't a problem, I'll probably go with 16k stripe in the end unless you think 32k looks better (kinda new to this so not totally sure how to interpret some of the numbers, 16k appears to be overall the best to me).

[GullLars]

well, most of the numbers don't change. Your highest PCmark score, AS SSD score, sequential read, and IOmeter numbers were all at 16KB stripe.
The difference in PCmark Vantage HDD was actually almost 20K points from 16KB stripe (74 184) to 128KB (64 864), and a solid 2000 points difference in both Gaming and Productivity subscores.
The extra wear (if there is any) won't effect the lifetime you will use the RAID. With 3 x25-V's you get to write about 100GB pr day for 3 years, if i remember the numbers rougly correct. Even if it halves the lifespan, you probebly won't write more than 20-30GB pr day, and in 5-6 years you won't be using that RAID anyways, it will be obsolete.
Few users ever write over 10GB pr day to their system drive (on average), and if you will be offloading storage of media files to harddrives (and putting your download folder there), you won't have to offer it a second thought.

[jalyst]

@GullLars, if you're still about I'm also curious whether you agree with One_Hertz's point here?
http://www.xtremesystems.org/forums/...3&postcount=26
In all honesty I'm still not 100% clear with the point/s he's trying to make.

Thank-you.

[GullLars]

I partially agree with him. 2R0 x25-V is not twice as fast as an x25-M, on any meteric.
x25-M: seq R/W 250/80 MB/s, 4KB random IOPS @QD 32: 35-40K/10-15K (140-160MB/s / 40-60MB/s)
2R0 x25-V: seq R/W 360-380/80 MB/s, 4KB random IOPS @QD 32: 60K/20K (240MB/s / 80MB/s)

From a lot of benchmarks, i've seen that 3 x25-V performs simelarly to 2 x25-M 80GB. Meaning 2R0 x25-V should be roughly 2/3 faster than a single x25-M. Sequential write heavy tasks will be an exception.

[GullLars]

*doublepost*

[SteveRo]

Meaning 2R0 x25-V should be roughly 2/3 faster than a single x25-M. Sequential write heavy tasks will be an exception.

This seems right on - good summation.