6_6_6

Intel X25-M Impressions

Recommended Posts

HachavBanav,

I am sorry, that is out of question on my live system. I changed the cluster size before and i have encountered so many problems (imaging utilities not working properly, data recovery failing, NTFS compression/encryption not working, etc). It is not something I will every do again.

What matters for me is how the system is. Therefore, I would have no way of verifying it with changed cluster size.

But, when i bother shuffling my data around and removing this drive, I surely will change it as you requested. Which tests would you like me to run then?

Write-cache is enabled but no difference with enabling or disabling it. I read on some Intel document that it is not important how write-cache is set.

I kept 4KB default. However, I tested with aligned partitions and saw no tangible benefit. Changing default cluster size renders most drive utilities useless, so it is running with a default 2008 install with default cluster size.

If anyone requires any tests, i would be happy to run it for them.

I would like you to try running a test with a 64KB NTFS cluster size (It's aligned at 1MB boundary in Win2008 which is ok for me).

I think there is a real performance point out of this "page size".

Of course, it may be disappointing to have to modify some Windows defaults...and it is obviously a very very very bad idea regarding the SSD desktop market feedback.

NB: I hope you did not disabled the "write cache" !

Share this post


Link to post
Share on other sites

Isn't 16 seconds to 11 seconds in Photoshop (whatever you did?) i BIG improvement?? What else can you do that improves the times as significant as that?

If Intel came out with a 5 Ghz procesor tomorrow that wouldn't improve the times as much.

Also 80 to 41 seconds boot time is impressive.

Also, how to quantify the experience from the Windows user experience when things in the start-menu pops up faster?

I doubt you'd see a 20-30% improvement switching to a newer drive.

All SSDs with Jmicron controllers is absolute crap when it comes to random writes, they are slower that floppies or storing your data on a webpage in Ulanbatar...

Edited by Laglorden

Share this post


Link to post
Share on other sites
Isn't 16 seconds to 11 seconds in Photoshop (whatever you did?) i BIG improvement?? What else can you do that improves the times as significant as that?

No it is not. It is something like 5 sec on subsequent runs on Seagate drive since it is pulled off of memory. If I am running Photoshop 10 times, 9 of this is 5 secs. Initial is 16 secs. So this translates into half sec improvement for me which is jackall nothing.

Also 80 to 41 seconds boot time is impressive.

It means fvck-all nothing when you boot your system once in 3 months. If I would be bothered, or if it was important, I could probably be bringing that down to 41 secs on Seagate by using xPerf and optimizing boot items.

Also, how to quantify the experience from the Windows user experience when things in the start-menu pops up faster?

I don't have anything popping up faster. If you want things to pop up faster set ShowMenuDelay to 1ms for your start-menu.

Of course there is improvement, but for most of the things that matter involving file operations, there is no change at all or it is worse. And there is no bloody space. Do you know how much time I had to spend to make my PGP disk data and OS partition to fit into that crap? It should have been worlds of difference to compensate that, but unfortunately not.

Like I said, this is only good for a laptop user or average goon who keeps his system on for an hour a day... Or someone who is upgrading from a 10 year old drive... Or DB/Web servers... Or perhaps gamers.

Edited by 6_6_6

Share this post


Link to post
Share on other sites

For your own, or for your Oracle?

I don't think it will make a difference for your own if it is MLC or SLC. They have similiar benchmark data.

Thanks for your effort. I guess, I'll go for Intels SLC version.

Share this post


Link to post
Share on other sites
I changed the cluster size before and i have encountered so many problems (imaging utilities not working properly, data recovery failing, NTFS compression/encryption not working, etc). It is not something I will every do again.

The point is that many many X25-* benchmarks noticed that the best performing (mix of average and random read/write io) choice should be using a page size of 16KB 32KB or 64KB...which makes sense if you know that the internal parallelism must be optimized to equally balanced regarding this kind of size.

About having non-working or broken features (I won't care of NTFS compression if I was looking for speed...and quickzip is still alive) when using a non 4KB NTFS cluster size, may be you could blame this guy who did choose this f...g non-professional tools :)

Share this post


Link to post
Share on other sites

Strange that your subjective perception of the X25-M SSD was not a significant improvement over your mechanical drive.

I have been playing with SSDs for about a year now on three different systems. In each case the subjective increase in performance was outstanding.

My main desktop machine (XP Pro) has two X25-E SSDs in RAID 0 config on an Adaptec 5805 controller. Formerly I was running two WD VelociRaptor 300GB drives in a RAID 1 config on this same controller (certainly no slouch). The performance is nothing short of phenomenal. I didn't even bother messing with alternate cluster sizes or partition alignment. I'm using the default 4KB cluster size.

I have two laptops where I replaced their 7200rpm drives with SSDs. They are both Samsung SLC drives. The performance increase in both cases was quite stunning.

It's a shame that you didn't experience the same boost in performance that most of us see with SSDs.

I don't think it will make a difference for your own if it is MLC or SLC. They have similiar benchmark data.

MLC drives have notably slower write performance.

Edited by drwtsn32

Share this post


Link to post
Share on other sites

So, you installed a SSD in your system. A system that you don't regularily run any 'real' disk-intensive applications on. And a system that you specifically configured (8GB RAM, no swapfile) to avoid using the disk subsystem. And the result was the SSD didn't seem like an improvement.

Was the SSD supposed to overclock your CPU 30%? Double your memory bandwidth? SLI your videocard? On a PC tailored to not use the disk... you bought a faster disk.

Right.

I'm off to buy a new inkjet printer: I hear it will double my download speed!... :)

Share this post


Link to post
Share on other sites
So, you installed a SSD in your system... <snip>

Oops, now that I've come home and read my post again, it sounds more snarky than funny... not what I was aiming for. My apologies: long day at work...

Bit

Share this post


Link to post
Share on other sites

I echo my early sentiment.

If you have a workhorse PC/Laptop that you spend >= 8 hours a day staring at, you will find that an X25M makes a fair bit of difference. Unfortunately (with the possibility of the newest OCZ Vertex), the Intel drives are the only ones that don't currently suck. The other (current) Jmicron based MLCs on the market are complete and total stinker. I'd recommend X25s for every PC, but they're too expensive to justify for a PC that doesn't help you earn a paycheck.

F

Share this post


Link to post
Share on other sites

6 6 6, allow me to disagree.

1)although i use 6 OCZ Core V1's on a Adaptec 5805 controller since june 2008, i must say that this setup works great in a Raid0 config. Back then no Intel or Vertex what so ever.

2) Thanks to the card cache, no stuttering. Anyway not only Intel X25-M or OCZ Vertex (SLC coming soon) are great, but as i posted somewhere else the PB22-J from Samsung (MLC) also is.

3) Take 4 SSD's of those mentionned above in a Raid0 setup with a Raid card and BLOW your Velocity's away.

Jeff

Share this post


Link to post
Share on other sites

I think, I'll wait until the E-Series is a little more affordable...even with the last firmware upgrade of the M. My ARC-1220 in the Oracle server is looking forward to them

(3 or 4 should be ok for the old IOP333?) to put the log files and rollback segments/undo tablespace on this array.

Share this post


Link to post
Share on other sites
1)although i use 6 OCZ Core V1's on a Adaptec 5805 controller since june 2008, i must say that this setup works great in a Raid0 config. Back then no Intel or Vertex what so ever.

Raid controllers (caching controllers for that matter) actually do a fair job of mitigating the deficiencies of the early Jmicron based SSDs. Specifically, the random write performance is greatly improved if you have enough cache.

F

Share this post


Link to post
Share on other sites

6_6_6, you've written at length about proper NCQ implementations. Probably most users don't have NCQ properly enabled on a drive which will use it correctly. This makes this computer vomit whenever they do more than one thing at once (like unzipping a big file and opening a game at the same time.) If an SSD performs better they would see a big improvement. (I don't have a drive which properly supports NCQ, so I can't vouch for this but it's my supposition).

Share this post


Link to post
Share on other sites

Installed my X25-M and I have to say for me it is a huge improvement over the WD Velociraptor.

My main application I use is Outlook and the difference it makes in doing searches in Outlook is tremendous, easily 5x-10x quicker, so am very happy, as I no longer have to walk away form the computer and do something else when doing "search field and message body" searches

Basically I'm very happy, but saying that I don't think I would have got it if it wasn't a freebee, it's still too $$$ for desktop environments

Edited by czr

Share this post


Link to post
Share on other sites

Malician, I must say that NCQ was a bigger improvement for me like you said. I was also emphasizing that 100s of times better random access timings did not translate into immense gains on desktop arena.

Where are you people all getting these freebies from?

I also must say this drive will need to get a secure erase. Its performance degraded terribly.

Share this post


Link to post
Share on other sites

But my ATTO screen says like this too yet it means nothing to me.

Are you running this as your boot drive or testing your oracle? If your boot drive, how do you feel compared to your previous system?

Couldn't wait :blush: and got an E-Series drive.

I guess, the ATTO screen says it all:

Share this post


Link to post
Share on other sites

Just deployed a vertex SSD to a customer in an architectural firm the other day. The SSD made SAV v9 die and crash the computer. After removing that it was stable (have to look into other AV applications..)

Customer is happy so far, Revit and Autocad load significantly faster, his most pleasant moment was making a PDF using adobe acrobat writer and just having it come straight up on the screen instead of the usual 10 seconds of disc churning before it appears. he uses his computer usually 10+ hours a day, so far he is more productive... it's probably money well spent in his case. All the other componants are so cheap now you can buy a really high perfoming PC with an SSD for $1500 (AUD so probably 1000 USD).

I'll migrate his email over soon and ask him about outlook performance

Share this post


Link to post
Share on other sites
The SSD made SAV v9 die and crash the computer. After removing that it was stable (have to look into other AV applications..)

You couldn't pay me enough to run Symantec's piece of crap. Give AVG a try; I've got two Vertex drives deployed at work, AVG is on both systems, I haven't had any issues. And despite the apparent assertions of some people in this thread, they have made a significant, tangible positive difference to system speed and responsiveness. If they prove to be reliable, every machine I build from here on out for work will be getting either indilinx-based MLC or good SLC SSDs.

Share this post


Link to post
Share on other sites
You couldn't pay me enough to run Symantec's piece of crap. Give AVG a try

While I wholeheartedly agree on the Symantec part (substitute with Macafee if you want) I'm not 100% positive with AVG. I run it - it's free after all - but I've had several false positives and although I've reported this, AVG doesn't seem to care. Also the updating system kind of sucks - it takes way too much resources while it runs. But it's free, so I use it. I'm more of a fan of Trend Micro products though.

Share this post


Link to post
Share on other sites

http://www.av-comparatives.org/ if you want some detailed A/V testing. Evidently both Symantec and McAfee have both recently totally redesigned their virus scanning engines as they're both pretty damned good now in current releases, compared to the garbage they were before.

Avira, AVG, Avast, Kapersky, NOD32...

Share this post


Link to post
Share on other sites

I'm definitely going to check that site. My gripe with SyMacafee was not the engines but the bloat of the software and horribel management tools (for the corporate versions). A comparison of those corp AV's would be nice. No Trend Micro to be seen though - strange.

Share this post


Link to post
Share on other sites
The SSD made SAV v9 die and crash the computer. After removing that it was stable (have to look into other AV applications..)

You couldn't pay me enough to run Symantec's piece of crap. Give AVG a try; I've got two Vertex drives deployed at work, AVG is on both systems, I haven't had any issues. And despite the apparent assertions of some people in this thread, they have made a significant, tangible positive difference to system speed and responsiveness. If they prove to be reliable, every machine I build from here on out for work will be getting either indilinx-based MLC or good SLC SSDs.

I should note that this was the corp edition so it just sits in the background and is generally unoffensive and does its job this is NOT the consumer version which I also dislike. personally I think AVG is way too ugly (these are design people..) What would you recomend that has had at least a minimum amount of effort spent on it's interface design?

Share this post


Link to post
Share on other sites

My favourite for years now is Avira. Many people tell about many false positives but that's not what I experienced. Avira is also very light on ressources and has a fantastic detection rate.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now