The new SSD Thread (Benchmarks, Brands, News and Advice)

If you buy a new drive there is a good chance that it will come with the latest firmware.

 

Considering that X25-M stock is pretty much at an all time low right now, it is possible any drives that are available will have the new firmware. I've got some tips that Intel is pretty much out of drives until the beginning of November, so maybe if you get a drive now it will have the old firmware on it.

 

So? Who knows? Maybe the 3rd gen will scrap the basic arch. But anyways the my focal point was their overall mentality towards it.

Here in Canada the price is not anymore expensive.

 

has anyone tried the intel ssd optimizer?

download link frm intel-hp:
http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&DwnldID=18182

is it any good?

oh and if i want to update the firmware of my postville, do i need to send it back or can i do it on my own? and will i loose my data?

thanks

 

Seriously - dude, don't make another post until you've reread the last ten pages. You glossed over the MAIN THING WE'VE BEEN TALKING ABOUT.

 

haha sorry to jump backward here, but does anyone else here think it's completely bias that Anand failed to mention how severely he underestimated the samsung drives? it seems to be near the top of the pack for all of his latest tests and after he said it is worse than a spinner in the last article, he failed to mention he may have been a little harsh... not to mention the fact that Samsung has said they will be releasing a user upgradable firmware.

although his articles are great and very informative, i think anand just likes to flex his muscles in the industry a little bit too much...

 

Yeah, I noticed it... even in his post that had "whats wrong with Samsung" he clearly benched a Summit with old firmware right at the time that the new firmware had come out. I offer as proof within his own article he benched a Corsair rebadged Samsung drive with the new firmware VBM18C1Q then removed the Corsair benchmarks sometime later that showed considerable differences in results between the Corsair and the Summit which are identical drives with Samsung chips and Samsung controllers.

If you have a Samsung drive with the older 1801 firmware performance will degrade and the only way to restore it is to do a full wipe with HDDErase. All of the Samsung drives that were shipping at the time of his most scathing article of Samsung were shipping with the new firmware that restores performance automatically when the drive is idle.

Clearly the new Samsung firmware should have rendered Anand's criticism of the Samsung drives moot yet "knowledgeable" posters are still quoting that article as if it were gospel!

 

Yes, and Samsung, like today, didn't have any issues 2 years ago.

I owned 1st generation Samsung SLC with 57/38 MB/s read/write more then a year ago

I noticed that too, but forget to mention it.

That's one more reason to stay with Samsung.

 

Hey guys,

hindsight is always 20/20 and please don't think that this post (or any of my other posts) is pro Anand. Of course, today you are (both) right to some degree. However, this does not negate Anand's statements at the time he made them.

You must understand (and I'm sure you all do) that at the time of his writing, all his statements were true (as I've stated before). With time marching on, those statements can and will change, but that doesn't negate what was reported originally. All that means is that we must be in continual research mode if we want to keep a handle on the very quickly changing landscape of SSD's in late 2009.

When I (openly) side with Intel and post my reasons (biases) for my decision (do a search of all the posts I've made, if you're interested), then it is the readers responsibility to decide if they are aligned with me or not (and basically, 'take' my 'advice'). The information that I post though, is not only for like minded people as myself. The information stands on its own.

What we each do with that information is what makes us progress from where we were yesterday. If we decide to nitpick each and every point - our lives are consumed by minor details (needlessly). If we decide to look at the overview - we can enjoy each new data point for what it is and quickly implement it into our unique situations - to our betterment.

I read/post here to learn (and relearn) about SSD's.

All information is relevant to me (even if it isn't about Intel ), because I am comfortable in my knowledge about my needs/wants to quickly place any new piece of information at the attention level it deserves.

I don't see it as 'bias' that Anand didn't point out his 'underestimation' of Samsung drives.

He went and made a new test which threw almost all previous 'benchmarks' out of the window.

This is exactly what I asked for a month or two ago!

If he were as biased as some here claim, then he wouldn't have included these new tests that showed the Samsung as superior (in the 'Heavy Trace'). Instead, new techniques, new results, new conclusions - seems like pretty unbiased reporting to me.

As a matter of fact, in the 'Light Trace', we can still see Intel dominating and Samsung sitting in last (SSD's) place. Is this usage scenario where Anand, before his Light/Heavy trace tests, felt that Intel was superior (and rightly or wrongly) extrapolated that to mean that Intel would also dominate in heavier usage too. This isn't too far fetched when we look at the numbers and see Intel getting 'better' the harder it was pushed (in IOPS at higher and higher queue depths).

Anyway, my point is that we shouldn't rejoice about failure (nor domination) by any one product or company, but rather, chose the product/company that offers us specifically, the best fit to our values, at that particular moment in time.

The same goes to nitpicking over an article that is now almost two months old.

For the record, I originally didn't (and still don't) quote that August 30 article as 'gospel', rather, I asked from the members here for proof otherwise. Nobody here could provide that. So I was left to rely on the benchmarks available at that time (... same as Anand, hmmm...).

Still looking forward to sharing any new information about SSD's with all the members here - objectively.

 

Just an update;

I had posted last week that I would be installing Win 7 x64 onto an Intel 160GB G2 (desktop) and also upgrading a Samsung 128GB Vista to Win 7 x64 (notebook) for a client yesterday.

With the Intel G2 firmware glitch, this has been postponed till Saturday.

My client agreed to wait for at least a couple of days just in case any problems showed up (and it didn't even take that long, ugh!) to update the firmware on his G2. This approach was wise (problems started appearing on the 'net almost immediately on Monday night).

(Normally, he doesn't even consider things like firmware updates - but the drive is brand new and still sitting in it's shipping box (unopened), waiting to be installed).

I see that many are 'laughing' at Intel right now and honestly don't see why.

We are at the bottom of the upwards slope of the bell curve with these SSD's and these issues are almost to be expected. These products are far from 'mature'. Nor is any single company immune to these failures.

What we risk (dead SSD) does not outweigh what we can gain (Trim, xx% increase in write speed, etc.), but we are always taking a chance that we will be without a system SSD for awhile (until it gets RMA'd).

While the totally safe option (no firmware available) is also currently forced upon its customers by Samsung drives, it is also a no-win situation for those looking to be on the bleeding edge using that companies products.

The best approach is to consider if you want to be on that edge (and take appropriate responsibility for your decision) or if you'd rather wait and be content with what you have.

Having the above choice is ultimately the best for everybody.

If Saturday is really the 'go' day for my client, I will post a further update then.

Take care.

 

Again i disagree. anand's benchmarks are not bias. His articles, however, are.

 

Thank God mine never encountered any issues. Hopefully Intel will release a new firmware for us that already flashed since I'm not going to risk having a drive that could potential brick.

 

Mormegil83,

I will certainly not ignore you (I've learned too much from your posts in this thread).

I also don't 'have' to post here. (I feel though, that I'm obligated to do so).

I've stated this before and I'll repeat myself here; I've read every single post on this thread at least once and most of what I've learned and believe about SSD's is because of the information coming from here.

I feel obligated to reply to some posts because of the help that I received (and continue to receive) freely too.

As to my having no real 'direct' experience with SSD's - that still currently stands undisputed. However, I also believe in the saying that a smart man learns from his mistakes (experiences) but a wise man learns from others (mistakes/experiences).

I base my comments not only on real 'experts' experiences on the world wide web, but also collaborate those opinions with laymen's experiences as provided on forums such as this.

When I was younger I had the time to be 'smart' (i.e. time to make my own mistakes, or what I currently see as reinventing the wheel), whereas now, I do my best to be 'wise'. But if I have mistakenly or intentionally provided information that is false, please correct me.

I have my biases as everyone else does, but I also try to make those biases known - especially when I am offering 'advice' in a public forum.

I am not searching for a title such as 'expert' or otherwise, but hope with my posts to not simply give an answer, but rather also give the reader an idea, some information, a keyword or simply another point of view that they can further pursue to either collaborate what I presented, or, bring back more information that makes the discussion progress to the benefit of all.

On a more selfish note, I also post because I know that when one can 'teach' the material they are talking about, that is when they truly 'know' it themselves. As I've mentioned in a previous post (paraphrasing); we don't read an article or two about this SSD stuff and then stop (because we learned all there is to know), rather, we are all continuously learning as this topic is in constant flux at this point in time.

There is another reason why I post replies; to not only provide an answer for somebody's question, but to also verify if the answer I (thought) I knew, is still valid currently (if it's not, I'm sure the whole board would let me know very quickly!).

Many of my clients have told me that I am a great summarizer, especially of complex ideas or relationships - maybe this is where I may come across as an 'expert' - sorry!

Do others feel like you do (and want me to simply ask questions instead of giving back some of what I've learned about SSD's over the last few months)?

I understand your concern that I may be influencing people inadvertently - note that I do not benefit if anyone here is influenced one way or the other - we are here to learn/share/discuss our ideas/facts/opinions of SSD's and hopefully we can all distinguish between the real information vs. the mere opinions that are offered here - particularly as it relates to our specific needs.

I choose to believe that we are all able to make this distinction and more information (for example, a few posts by myself) is always preferred to less.

 

Sorry Mormegil83,

but I have to take you to task.

Where do Anand's articles not reflect what the benchmarks show, at the time he wrote them?

If your position (as I understand it) was even remotely true, how come there are only a handful that defend your position?

Please, this is a fair question... are the rest of us really that blind?

Show us the light!

I'm already glad I personally haven't bought any SSD as these new 'Light/Heavy' traces are the exact type of 'tests' (don't want to call them benchmarks) that I have been waiting for.

Should I be mad that he didn't do these tests earlier? Or should I be glad I waited? Or, should I simply be sad that I'm missing out on the huge performance jump SSD's offer (especially over notebook HD's).

In my case, none of the above, because I make a decision and take responsibility for it.

Even if hindsight one day makes me regret my waiting in this particular case, I will hardly be blaming any of the sources of information that I based my (original) decision on.

There is nobody without biases - learn to sift through them to get to the information pertinent for your use - and then apply that knowledge appropriately.

 

as i said, my intention was not to make you stop posting your opinions and research. It is just my thought that your novel long "summeries" are based on merely opinions you've gathered from others' opinions. and in my opinion it is tough to base so much advice on just that.

 

just the fact that he made no mention of the samsung controller performing well shows his bias, after slamming it so hard. the least he could do is say "Samsung actually did pretty well..."

 

IntelUser,

I agree with your post above, but want to point out that even Intel has 'experience' with SSD's for almost/(over?) 20 years. If Samsung is also at around the two decade mark (I really don't know their history), then it is pretty difficult to say who has more 'experience' time-wise.

More importantly, past experience isn't any guarantee either, but simply one additional indicator of future performance.

Not that one or two years (even over two decades) isn't significant - sometimes all you need is one day for a company to leapfrog over all others (like Intel did in 2008), but time by itself is no measure of what the future will hold for any particular company, nor what the company can reasonably be expected to not stumble on.

 

Fair enough,
but may I correct you in that I post my summaries on 'facts' based on the direct experiences presented by 'experts' (which can always be checked and re-checked, independently), backed up by my personal experience (if applicable, in similar fields), and finally backed up by the majority of 'opinions' on forums (biased I know, because this is only forums that I read, not all of them) world wide.

I do not usually offer my posts as 'advice', but rather 'this is the information that I have that relates to the question posted'.

If there is other information that I should also be taking into consideration (for that particular question I tried to answer), then I hope that a person would share it with the rest of us too.

Conversely, if somebody disagrees with what I wrote, I am not against changing my position in the light of new facts. Provided those facts are presented irrefutably.

My favourite saying is 'I hate it when I'm right'. Not because I'm trying to be funny, but because it means I didn't learn anything new.

I am pretty sure of myself and my facts (sure, call me stubborn ), but I'd rather be proven wrong (and have learned something new along the way) than simply repeat what is considered 'common' knowledge (or at least considered 'common' to me).

 

Hmmm, something seems odd about this. My driver says:

Intel(R) ICH8M 3 port Serial ATA Storage Controller - 2828

The driver version is 8.2.0.1011.

Now, you are saying that this driver is for an NVidia chipset? If so, then why wouldn't the driver be made by NVidia?

You think that I should uninstall this Intel driver?

Do others agree with this?

Should I install a more recent Intel driver, say:
Intel(R) Matrix Storage Manager, Ver:8.9.0.1023 Date:7/17/2009

I notice here that doing so can increase performance:
http://communities.intel.com/message/70866;jsessionid=929BE119C3339F3B8D3F30389C08F60E.node5COMS

but it can also make the drive unbootable (see bottom of page at that link).

 

Mormegil83,

I don't mean to egg you on, really, but this is how I read the article:

Light Trace: Samsung second place finish (by brand) and 20% slower than #1 Intel.

Heavy Trace: Samsung second place finish (period) and 25% slower than #1 Vertex.

Gaming Trace:Samsung second place finish (by controller/brand) and 10% slower than #1 Intel.

I can agree that he may have glossed over 2nd place finisher Samsung in his comments, but really what I take away from this article is that the two players worth considering are still Intel and Indilinx based SSD's. Just like he hinted at in his previous SSD article.

I've already mentioned that I'm glad I haven't purchased an SSD yet. Not because my previous favourite was Intel and now it has 'lost' at what I would want it to excel at (Heavy Trace usage), but rather, I'm glad because I want the best combination of the performance that Intel and Indilinx based drives offer currently in one SSD (best in Light Trace (which I'm considering calling the 'snap' that the drive offers) and best in the Heavy Trace (which I'm calling my 'default working mode' and primary reason for me to buy an SSD).

Notice that Samsung didn't even enter into this equation for me?

Just like for Anand, I suppose - he's not looking for 'most improved' he's looking for the best SSD currently available - just like I am (and I'm sure most of his readers are too).

 

I have a need to comment this text:

I'm not laughing to Intel at all, just want to say that Samsung isn't that bad as Anand stated before.

Even I was considering selling my SSD and try Intel, but thees FW problems delayed my plan.

 

can someone post the URL for the comparison chart of all the various SSDs? I've tried searching here for it and can't find it for the life of me. Thanks

 

Do you mean this one?

http://www.storagesearch.com/ssd-top10.html

 

Yeah honestly, maybe i posted just because we think differently and lately I have not been able to participate in the thread as much as I would like, and you usually have a different view than me and I am not on enough to give my opinions as well, and get frustrated seeing conversation pass by that I would have liked to be involved in. Whatever... Agree to Disagree

 

no it was a nice chart with all the various manufacturers SSD's with controller type etc.

 

Hmmm...

Don't recall that one, but would hate to see how that reviewer looked like when he was finished compiling all that data into one chart!

Still would like to see it though...

 

imho, that chart is not even worth looking at as a consumer lol

 

Mormegil83,

the difference is that I'm still deciding when to put down real dollars and budget about a week's worth of time (even if it takes me 'only' three days straight - no sleep) to get whatever SSD that seems 'worthy' enough of my time and money to a working state on my computers so that I can continue to produce the work I need to stay afloat.

Whereas you're already enjoying the SSD experience - for a while already.

That's why I 'have' to be here as often as I can; you can actually enjoy life now and worry about the little things like not participating in conversations that interest you.

Okay, bad joke, but we'll agree to disagree as long as you keep trying to keep me in line and I keep trying to do the same with you.

 

I have: 1 and 2

First one is more accurate.

 

Whoa! We agree on something!!!!

(Really, I've clicked on a link so many times with that 'chart' coming up that although technically it is what it says - it is also pretty useless to make a buying decision from).

 

it was the 2nd, thanks

 

I have to take issue with this (please don't write a novel reply - I'm subscribed to 38 threads on this forum as it is), because I just don't understand how Intel, a company I assumed would have access to W7 to test on its machines (not the RC, but the final release of W7 before it was delivered to the retail sector). I don't know how an error like this wasn't caught before the release. It just seems like to many of us Intel is pushing the race-to-release a bit close on this one. I mean, bricking people's drives (ones that carry the OS, and possibly any other of a person's personal files) is serious business. Especially the first day you try it out. That is messed up.

 

Okay; short version (for me).

You've got to be kidding, right?

With all the combinations of customer specific hardware, software and O/S interaction - not to mention M/B BIOS and other hardware bios/firmware/software in the mix, it is actually amazing that it worked for most people. You really don't believe that any one company (even one as huge as Intel) is able to test all the possible hardware interactions? Let alone all the software/hardware/bios/firmware/OS interactions too.

Now, bricking an O/S drive (and potentially the customer's data) is serious business - hopefully nobody actually lost their data (backups).

But, Intel did the best they could do; they pulled the firmware updater off their site immediately and are searching for the cause of those bricked drives as we speak.

How would you suggest they improve (or any other company in the same situation)?

I too wish for a perfect world with my hopes and dreams making me strive for that perfection... while my actions are more solidly (and cautiously) based on the imperfect one I live in.

 

You recall this minor hiccup too?

Just bashing Intel for once, to compensate for all the Intel nut-hugging going on in the past months.

 

While I do have a confidence that anything that goes wrong with intel, they will fix it. Unlike a company like Jmicron. Firmware f-ups like this are the reason ppl said just go with intel, they don't have screwy crap going on, and they are fast as all hell! Well...

Given same price and capacity, I would still have chosen intel. That not being the case, I have been more than pleased with my latest purchases. Even my Sammy slc has been stellar, just lacks in the sequentials which are most important for gaming...

 

(this is a repost - hoping to solicit a response...)

Hmmm, something seems odd about this. My driver says:
Intel(R) ICH8M 3 port Serial ATA Storage Controller - 2828
The driver version is 8.2.0.1011.

Now, you are saying that this driver is for an NVidia chipset? If so, then why wouldn't the driver be made by NVidia?

You think that I should uninstall this Intel driver?

Do others agree with this?

Should I install a more recent Intel driver, say:
Intel(R) Matrix Storage Manager, Ver:8.9.0.1023 Date:7/17/2009

I notice here that doing so can increase performance:
http://communities.intel.com/message...F60E.node5COMS

but it can also make the drive unbootable (see bottom of page at that link).

 

I said, "IF" it has an Nvidia driver...which it clearly DOES NOT but there is still a bottleneck no?

 

Well, is there a bottleneck? I don't know. I do see that my numbers are lower than you posted.

 

I don't think it is bottlenecked hooterbif, but I also have no idea what is causing your less than spec ATTO benchmark either--maybe you can give some more details about your system specs, what operating system you are using etc.

 

Dell D830, XP Pro, 4 GB RAM, Intel 160GB G2 SSD

 

XP Pro is your problem

 

OK, thanks. So, in order to get the performance that you showed, would I need to change my operating system to Vista? Win7? Also, would i need to choose a 64bit operating system?

 

Did you align the blocks for the XP installation?

 

I'm gonna shut up now cuz I haven't used XP Pro in like 3 years...there are some knowledgeable users here that are doing fine with it that can help tho.

 

No, I didn't do the align procedure.

 

TomyB. has a thread on how to change alignment in XP without data loss.

 

XP SP3? (Is your G2 firmware updated, or 'original')?

If your partition was aligned, then you may also want to check your BIOS that you are running in AHCI mode too (instead of IDE or Compatible).

Just watch out that you can't simply change the mode (Windows will not boot up afterwards).

You may have simply missed it (or are simply ignoring me ), but I also suggested a few pages back that you might want to try updating to the Intel Matrix Storage Manager (IMSM) which will update your drivers to the latest possible for your chipset (and it shouldn't run the installer if your chipset doesn't support it).

Does this help at all?

 

My partition is not aligned, but I will check out TomyB's post. My BIOS offers AHCI or ATA. Right now, it is set to ATA. My G2 firmware is updated, and I run TRIM every night.

I was not ignoring you, but I didn't see your post. I just tried to run the IMSM, and it said that my computer did not meet the minimum requirements.

 

I can't find this TomyB character, but here's a link I emailed to myself a while ago: http://www.ocztechnologyforum.com/forum/showthread.php?t=48309

 

Okay, with IMSM out of the picture the only thing it can be is that you are not set to AHCI mode.

I would recommend a full re-install of XP with your BIOS set to AHCI (you may have to download the drivers and install them at the 'Press F6 to install additional device drivers' prompt).

I know I have read some posts where you can change the setting from ATA to AHCI with a 'live' system, lets see (googling)...

Aha!

http://forums.techarena.in/windows-xp-support/936428.htm

Hope this is the cure for your benchmark blues!

(Yah, I know, sometimes benchmarks are useful).