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

What firmware do you have on your 2nd gen Samsung drive? Clearly Anand was benching a Summit with VBM1801 because he acknowledged that the Corsair drive is the same Samsung drive as Summit and it had VBM18C1Q, he even showed a screenshot of it. He also "had" benchmarks of the Corsair at one time in that article and it was in the middle and toward the top in many of the testing screenshots. He has subsequently removed the Corsair comparison from the article, I wonder why?

 

because you know better, of course!! and you know the truth how the world tries to make samsung look bad and all. oh the conspirancy.

 

Not the whole world ...but you and Anand clearly have an axe to grind with Samsung--just go back and look at the rhetoric both you and Anand use with regard to Samsung drives in your posts.

 

I agree with most of what you say davepermen, and enjoy reading your posts, you are a very knowledgeable person--Anand too as a matter of fact. I just take exception with you and Anands criticisms of the Samsung drives. While Anand has backed off somewhat and his recent benchmarks are showing the Samsung drives about where they should be, you have not. You are still using rhetoric such as "sucks".

The drives Anand benched with VBM1801 do suck, just read the posts of the 1801 FW Summit owners over at OCZ forums to find out how much. Those drives have no PRF, they degrade and stay degraded unless you image the drive, do a full HDDErase and put the image back on. The drives that were shipping at the time of Anands sharpest criticism of Samsung had the newer FW yet he chose to bench a Summit with the old sucky FW.

Never once have I denied that Intel drives are the best, heck, if I had the money to spend on my computer right now I would buy one. I don't but I did buy the Summit Newegg shellshocker for $129 because I was interested in SSD's and I am not disappointed.

 

I always stated it simple: Samsung is not worth the money.

and I stay with this argument. Two reasons:

a) In the bottleneck case of the devices (random reads, espencially random writes) they are MUCH worse than similar priced other ssds (the intels are 10% higher priced right now right here in switzerland, but deliver 4x or 400% more performance in the bottleneck cases). so paying 10% more, but getting 400% more is very simple to understand why i get biased to the intel, then. (indilinx aren't cheap here why ever).

b) i have physical experience with both (and more ssds, like ultra old samsungs, some jmicron crap, usb sticks, mtrons, first and second gen mlc samsungs, and intels), and they result in lower perceived performance. slower boot, slower app starts, longer "idle times" when you click something. as said, not as snappy.

so i stay with my point: they're NOT WORTH THE MONEY THEY COST. they feel and behave like a good hdd, but cost much more. others that cost much more feel and behave like something from another world (or century, or what ever).

what's wrong with my retoric? i always said they work fine, are not like jmicron crap that is beyond usable. but they are NOT WORTH THEIR MONEY.

how hard is that to accept?

and if i requote anandtech again (*sight*)

he states the same.

 

and btw, samsung stated, themselves, that the degrated state is a feature. but i won't find that quote from them anymore as it's now around 1.5 years back on some random internet page.. (and if it was on one of their press releases they sure pulled it by now )

 

I sure would like to try one of those Intel drives out if it's that much better than my Summit. I thought my SSD was good, never a hiccup, stuff just pops right up when I click, hell, it took 8 minutes to install Windows 7 Ultimate x64. Wow I bet that Intel could do it in 2

 

just to add few things
yes, garbage collection is link/compile/debug term
I know it from java programing, it is for deleting variable values that are no longer fetched and can no longer be used, so GC simple erase that data and more memory is available after erasing

everyone use garbage collection, but, I would name it self refresh or something like that

g1 maybe had it since day one, but don't forget, it works fine only if fw is up to date

my sammy refreshes itself like a charm, also, tomy b 16 gb slc did the same thing very good, so, I think I can say that samsung is doing a good self refresh thing since a day one

 

not yet, as i never got lower than 8min install, at least from a usb stick. but i even got 8min installations on some hdds

maybe with usb3?

but it looks like i have to reset my mtrons finally. they do have some hickups by now (around one year of daily usage now, including testing and restoring os', copying and editing hd movies, music production, gaming, etc). where is my firmware update tool again? it resets them so quickly

 

Samsung is going to issue a FW flasher towards the end of the month that supports TRIM but I'm probably going to stay with VBM18C1Q. Tony over at OCZ has been testing the new FW and is saying VBM18C1Q is better. My drive restores itself when idle.

 

well, my opinion is that ssd that could easily magically shrink to 8mb, or could brick itself if I update fw under se7en or if I set bios pass for drive is also not worth my money

also, anand were wrong stating that there is no sammy fw upgrade...
end, for sake, WHY O WHY do they call summit as a samsung drive...
it is OCZ drive
samsung drives are samsung drives, summit iz ocz drive...

 

Because they are Samsung drives with Samsung controllers--they are just labeled by OCZ and Corsair etc. It's what they call, "rebadging"

90 something % of SSD chips are made by Samsung, they are differentiated by the controller. Most OCZ drives use Samsung chips but Indillinx controller. AFAIK, the Summit is the only OCZ drive with a Samsung controller.

 

so much wrong in that post it's quite ridiculous.

btw, you know that any device can fail. but if it happens only randomly, only for a minority, then it's called bad luck, and it can happen to the best most reliable things. (that's why one always needs some form of situation dependent backup). but the sammies are not up on par with other ssds in performance, physically not possible (firmware upgrades can only do so much), and thus are continuing to always, for every one of it, not being worth the money.

and about the naming chaos you introduce, as said. so much wrong in this post..

 

Yes all the OCZ drives are rebadged. They don't make anything themselves except maybe the supporting apps (at least the last time I checked they don't). Did you know that the OCZ Vertex and the Partiot Torqx are the exact same drive. The Samsung PB22-J is the Corsair P256. I'm sure others can come up with a large list of drives that are the exact hardware just sold by different companies.

 

I know that summit and corsair are same as sammy, but, if you are benching ocz drive, than it is ocz drive, also, if you are benching corsair drive, it is corsair...
if you wanna bench sammy, then get damn REAL sammy and bench it, not some repacked versions...I know that it "should" be same drives, but, it kind a bugs me

jeah, well, it seams that only intel had bad luck so far
like 3 or 4 times

 

Yeah, you can't take it for granted that the rebadged drives are the same without comparing the firmware. I found out the rebadged Intel drives by PQI, Kingston and others have firmwares that are missing features that come with Intel branded drives. It could be that other manufacturers do the same. Some of the rebadged Intel drives have JMicron controllers--ouch!

I know that some of Dell's Samsung ssd's have Samsung controllers and should be the same as Summit and Corsair mlc drives but they have slightly different firmware codes. I don't know if they are better, worse or the same. Samsung, Intel and Indillinx all keep a tight lid on info when it comes to firmwares and I can't really blame them. Supposedly the Indillinx controller was invented by Samsung engineers.

 

Dude no intel rebadges have Jmicron contollers! LOL yes Kingston has rebadged intel drives and yes they ALSO have a V series of drives using the old $hitty jmicon controllers, but u for sure can't call them rebadged Intels just cuz the company also sells intels....

Now I know why Dave hates sammies so much more than I do. They cost more for him than others. In america the sammies are actually less expensive than indilinx for a while now...

 

well, sammy doesn't sell them (or back then didn't) direcly in stored, only let them get repacked by other companies. that's their preferred way to sell stuff: sell it in big batches to companies, let them put a name on it, and then run away if it sucks and the company which put the name on it can't do anything to help the customers

hm, no, heard of enough failed drives on all sort of brands. still, hdds die much more often.

 

i don't hate them. i just say they're not worth the money. indilinx based stuff here isn't worth the money, too. much more expensive than intels actually.

still, when you buy indilinx, you get a massively faster drive, same as when you buy intel. on a sammy, you get a 'a bit faster' drive. spending more than 150$ on a samsung 128gb is too much for what you get imho.

on the other hand, i was finally able to put my sammy for a test into my work laptop, and it's such a blessing. then again, that drive in there results in a 30minute start (incl. outlook and all).. so anything is a blessing

 

if sammy only sells drives as repack, how come my drive has a nice big samsung logo on it ???
what are you talking about man ???

and no, I've NEVER read that someone has bricked sammy, on last 1000 posts, and I've read all of that, but intels, there were more than one example, even jmicron crap, they did suck at performance level, but, they don't lose data or shrink to 8mb, so, no, intel IS the only bricking problematic drive....
intels are fast, but, that's nice, but, how log they will work ???

 

yeah, you're right, a drive having Intel flash with JMicron controller is no more a rebadged Intel drive than a Samsung flash chip with JMicron controller would be a rebadged Samsung.

 

you got it from dell i guess? because dell is the rebrand, then. as far as i can see, they start to show up non-branded in shops, but they existed for years (older versions) but you could never go out and buy them in a shop. i even phoned with samsung about it, and they "don't want to sell directly, only to oems and rebranders".

why? much less support hazzle.


and about your big issues with intel drives. as far as i remember, there where around 2 - 3 different people in this very thread complaining, and from google, i can detect around 5 people around the world. compare that to 10000 to 100000 sold drives, that's around .. maybe .. 0.05% or so? compared to what? about 100% of the jmicron customers not able to 'just install an os' on their drives? you just don't get it that some complains, no matter how big they might be in a forum, don't mean anything about product quality. not a tiny thing.

and i know we had topics about issues with the samsungs they recieved from dell, not in this thread, but on this forum. 1 or 2 at least.

there is a problem when every 10th person has an issue, not if one in 10000 has one. then, the problem gets easily solved: RMA, next drive, done.

the drives get classified by the controller they use. so far, kingston and intel themselves sell drives with intel controller. ocz and others and dell and apple sell samsung controller based drives. all sort of vendors sell indilinx based drives. i don't care about who sells them, i only care about what's in there.

 

deveperman, usually it's nice read Your posts, but sometimes You and anandtech are wrong. You saying that Samsung doesn't sell SSDs and anandtech saying that 4 KB random writes on new SSD are 5 MS/s.

Samsung 32 GB SLC, Samsung 64 GB SLC, Samsung 64 GB MLC, Samsung 128 GB MLC, Samsung 256 GB MLC

Geeks.com had 64 SLC for something like 200 $ or lower.

And why now this 0.05 % of Intel users doesn't matter, but when You were arguing something about backups then some low percentage of users was very important?

I agree with Your opinion on backups and stuff like that, but don't agree with You that this 0.05 % Intel users doesn't matter.
Would it be nice if You are in those 0.05 %?

I don't expect any answer, just feel a need to let it out.

 

as i said, afaik they are in stores, but you're right about the SLC which where available since long. still, they don't show up (here) in real stores, unlike the intels which you can buy in physical stores.
and samsung stated once they don't like selling them directly themselves.

still, thanks for the correction.


and about the intel thing, well, it's just an overblown tiny issue for some users, that gets overblown like any tiny thing in the internet. omg it failed somewhere LETS ALL CRY AND POINT FINGERS. typical crying behaviour on the net, you should know that, and know to not react like this is the truth.

other than that, yeah, backups are the holy grail in having a great live (not only in computer sense btw). some days ago, i had a power out, killed my os boot files. restore, done, all going on like before. and anything can fail (seen enough while clubbing and performing with my laptop on what can fail besides software and hardware.. i got drinks spilled into it, etc).

and anandtech is right about the sammies, BUT THEY TESTED IT DIFFERENTLY THAN OTHERS. they tested it with several random writes IN PARALLEL, unlike f.e. crystaldiskmark. this is a VERY important difference, that matters very much (and why do we know it's important? because that's where jmicron failed massively. if one write hung for a quarter of a second, there was NO other write possible as it can't run in parallel. and the full os got blocked).

samsung is bad at parallel random writes. at least in their mlc implementations.

you expected an answer, you know me.

oh, and btw, i wanna be crybaby now, too.

i'm DAVEPERMEN, NOT DAVEPERMAN!!!

(just feel a need to let that out again )

 

Agreed SSD are reffered to by there contoller. Slower flash and diff firmware have some affect but as long as u got a good controller u have a good drive

 

Ok, quick question.

I'm able to set both drives to IDE mode, instead of RAID or AHCI, and see both drives as two different drives....

Will other FW flashers work on the drive? I just want to update it so I have GC....even if it means I have to re-install windows and everything.

(I'm assuming I can, but the FW flashers out there are only for Samsung SLC and not MLC drives....unless someone has obtain a MLC flasher...)

 

sgogeta4,

I saw the laptop being used, but I still don't know how you had eBoostr running.

I will also state that even if eBoostr tries to mimic ReadyBoost, it surpasses it so much that I would say that ReadyBoost is trying to mimic eBoostr (and ReadyBoost fails miserably).

For an example of what information I would like to read about your eBoostr setup, here is mine:

Sony VAIO notebook 8GB RAM, Scorpio Blue 500GB HD, eBoostr Beta 4 (latest build 538) using an Allocated size of 15,268 MB - Cache Fill of 95% - Random read speed of 27,195 KB/s). The eBoostr cache is located on a 'dog slow' Lexar 16GB ExpressCard 34 SSD, which, BTW, takes over an hour for it to fully write the cache initially - that's how slow this SSD is for writes, but it does read (randomly) at about 27 MB/s - much faster than the Scorpio Blue drive can deliver data to the CPU.

Maybe I went into too much detail, but you have definitely not provided enough info.

As to using a RAM disk, for a subset of what eBoostr caches/speeds up, I agree you would not see an improvement - whereas with eBoostr I see an overall system improvement - just like everyone reports with an SSD.

 

MegaMan X,

Quick answer; No MLC flasher will currently work. Samsung supporters say 'Stay Tuned'.

Longer answer; although I have read even on this forum that a Dell firmware upgrade for an SSD will work on a non-Dell computer/SSD, I am too leery of such attempts to recommend them. Especially from Dell.

Each manufacturer has the option to put in microcode to ensure that their software/firmware doesn't work properly except in their own systems. I would not be taking a chance that a Dell firmware (meant to work around issues that Dell systems have) would work properly for a non-Dell system (even if the SSD we were flashing were exactly the same one meant for the Dell firmware upgrade).

Call me chicken, but I would rather have a slower, but working computer - than a computer that is bricked or almost worse; randomly 'hiccups' for apparently no reason (except that it's running an SSD with firmware that is trying to correct/minimize a problem on hardware that would never exhibit that problem - because it's not a Dell.

(Please substitute 'Dell' above for any manufacturer that offers firmware upgrades for Samsung drives like 'Lenovo' and 'HP' etc.).

 

Just an update on getting XP Mode working on my Sony VAIO (Thanks again MegaMan X on that BIOS!).

When I originally enabled VT in the BIOS, the computer was running as good as it ever was, but when I would enter XP Mode, the mouse would stick/jump the XP Mode computer would seem to grind to a halt and generally it felt as if there was a virus on the computer when XP Mode was running.

A few dozen (hundred?) webpages read later, I downloaded and installed VMWare 7.0. WOW!!! This is how a VM is supposed to work.

Although XP Mode is free for Win 7 Pro and Ultimate, I think that most 'real' enthusiasts would not use that slow (mostly because of 16 bit Video drivers!) and irritating implementation of getting XP to run virtually.

Just a little heads up for anyone else thinking they 'need' XP Mode in their Win 7 installs. Better to use VMware instead - (about 4 or 5 times faster for me).

 

@ davepermen: those problem Intel have with SSD are big thing. First time they withdraw whole G2 lineup, then new FW, what's next.

And booth issues ended with loosing data for users.

 

davepermen,

you are one of the reasons that I joined this group and your posts are also one of the reasons that I also believe Intel to be superior to Samsungs.

The benchmarks you post above from Anand just confirmed what you were saying many weeks/months before his article was published.

People have questioned me about my 'theoretical' knowledge about SSD's and about my not having any direct experience with them. You are one of the people that I have read and trust because you have played with many SSD's and in the end, you choose Intel.

After last weekend, I too have some direct experience with SSD's and all I can say is that it continues to confirm what Anand, yourself and all the other sources of information that I have gathered for myself say; Intel is the single best SSD drive out today.

I will expand on my SSD experience in my next post.

Glad you decided to stay on the forums for a while more!

 

Iometer 4k random write 60gb OCZ Summit 50% full.

 

Those 4K numbers are not bad, still much better than a traditional hard drive. Yes, other SSDs do better...but right now the Summit can be found for half the price of the alternatives. Keep that in mind.

 

512k read


bootup


512k write

 

flasher itself will tell you is your drive ok to flash or not

 

You know what, at least Samsung has not released firmware - TWICE - that has bricked drives from its customers.

This is supposed to be a thread for sharing data, benchmarks, and advice. Nothing in the thread description says flame bait or troll. If you have a claim to make about a drive's performance, we expect you to back it up with data.

 

Okay, Samsung 128 GB SSD firmware VBM19D1Q (no updates available on Dell site (this is a Dell notebook). 8GB RAM, Vista 64 bit Ultimate. Intel Quad Core Extreme 3.06 GHz CPU.

The above notebook is less than a month and a half old. The Samsung SSD stutters, the right-click menu takes minutes to show up (sometimes... other times it is normal/instant). The owner did not know about leaving the computer on, but logged off so that it could do 'garbage collection'.

This client called me to upgrade this notebook to Windows 7 x64 Ultimate. When I saw the state of the SSD, I was a little more than worried. I did a few comparisons to my Core 2 Duo 2.26 GHz, 8GB, Scorpio Blue 500GB, eBoostr enhanced VAIO and the VAIO blew the Samsung SSD out of the water.

The VAIO was easily 5 to 8 times faster than the DELL for starting programs. The DELL owner saw the VAIO move and I could tell he was thinking 'why the () did I spend so much on the DELL'!

If this was my (new) machine, I would have returned it so fast, Michael Dell's head would still be spinning. Even though Vista reported WEI as 5.9 for everything, this was not even running at a WEI of 2 in 'real world use'.

Never the less, I did an Upgrade install of Windows 7 Ultimate x64 and almost 2 hours later (I guess that's fast compared to the reports that a Win 7 upgrade can take 8 hours or more), Win 7 booted up for the first time on this notebook. I checked Defrag and turned it off. Then I ran WEI on Win 7.

As reported before to MegaMan X, the Samsung now rates at 5.4 under Win 7 x64. This really had me confused because now (with Win 7) the little Sony/eBoostr combo couldn't touch it. Excel up before you could let go of the mouse button, Word the same - Now this was an SSD!

Except... the computer froze (totally) and we had to reboot it. (It was totally unresponsive so I pulled the power plug, started Windows 7, logged in and ReBooted so that it was a 'proper' boot up). The reboot was taking about 3 minutes.

(The 'freezing' was corrected by uninstalling Roxio and installing the new version).

I repeated this two more times and each time was the same. With more than one program using/needing the SSD, the computer would revert more and more to mechanical HD levels (or worse). With 3 programs or more starting concurrently, my VAIO/eBoostr setup would continually be quicker.

The owners response? This is night and day different (better) than how it ran under Vista - don't touch it!

Hmmm... So there you have it. The inconsistency of a Samsung drive. 'Magical', one minute and a 'dog' the next. I disabled sleep/hibernation timers and logged off the computer to do it's 'garbage collecting' (this has gone on for the last three days (over 8+ hrs each time) and the owner is still experiencing 'stuttering' with this SSD - but he is simply willing to live with it).

Now, for this same client, I also did a clean install of Windows 7 x64 Ultimate on a desktop AMD Opteron, 8GB, Intel 160 GB G2 SSD. Once installed, WEI showed as 7.5 for the G2.

The CPU on this computer 'sucks' compared with my experiences with Intel CPU's. However, the SSD (with the original G2 firmware - not the TRIM capable one) was insane.

Compared to my 1xVRaptor/3xRaptor based 8GB 9450 QuadCore CPU, the Intel SSD extracted, installed and did the updates (over 1GB downloaded) for CS4 in the time it took for my VRaptor to extract the CS4 files. Or to put it another way (Intel G2 vs. VRaptor); extract 5 min vs. 30 min, install 12 min vs. 30+, upgrade 13 min vs. 45+ min.

Compared to my VAIO/eBoostr notebook, the Intel G2 drive was sometimes quicker, but also sometimes slower than the Scorpio Blue 500GB / eBoostr combination. What? Yes, you read that right - opening programs (or the 'snap' as I refer to it) eBoostr sometimes could beat an Intel G2 SSD with an AMD Opteron CPU.

Were these wins 'meaningful'? I would say no, not overall. What it does is confirm for me that currently SSD's are not offering anything near what I would call 'value'. Not even the Intel SSD's. At least not with the 'tests' or 'real world' use I was able to put the drives through.

If you haven't maximized your RAM, if you aren't running the best/fastest CPU and if you aren't using the best current O/S (Windows 7 x64), then an SSD may or may not be the best system speed up. (My eBoostr experience shows that).

This does not mean I won't try one (G2) if I get the chance on my own system so that I could do my own tests (I'm not saying my VAIO Core 2 Duo was above the Opteron AMD CPU that the Intel was installed in, I just know how much less 'snap' I felt compared to my Intel CPU/eBoostr combo).

Comparing my clients computers to each other, the Dell notebook with the Samsung SSD and Intel Quad Core Extreme CPU would open up an AutoCAD render in about 2 minutes whereas the desktop Opteron AMD CPU with the Intel G2 would take closer to 6 minutes. This operation doesn't depend on the SSD (the file was fairly small; about 25KB), but rather the speed/efficiency of the CPU.

The Samsung felt fast under Windows 7 unless you needed it to do more than two or three things at one time. The Intel G2 felt slow (because of the CPU) but would get faster the more work you made it do.

This is not a comparison between the Samsung and the Intel drives - this is just real world experience of these drives in use in typical scenarios.

I guess that I'm buying eBoostr 4.0 at the end of this month (or whenever the beta period is over) because it is giving me an immediate and obvious boost to my performance with the way I use my notebook computer.

Samsungs are a 'never-buy' item from what I have experienced (about 8 hours total) with this top of the line Dell notebook over the course of the last month.

Intel drives need further investigation (on a 'modern' or Intel platform) by me, but they certainly delivered what they claim; consistent performance.

Finally, davepermen has been proven to provide 'solid' information (at least to my satisfaction) on Samsungs vs. Intels SSD's - all I would recommend to you davepermen is to try Windows 7 - the difference from Vista to Win 7 on the Samsung is mindblowing, on the Intel, it may be a little more subtle, but I'm sure the improvement would be welcomed anyway.

I'm not turned off of SSD's, but I am certainly comfortable that waiting for G3 drives (or even G4 drives, if the G3's prove a letdown) is the best option at this point in time (for me).

People keep saying that 'for the price' XXX drive is okay. No it's not. My time to install, reinstall, maintain, etc. is worth more than the cost of the drive itself. What is the point of getting 'good enough' performance - at a premium - when the same effort and a little more money can get you 'best'?

I really wish I had an Intel G2 to try along with eBoostr (like I have politely asked others here to test for us), or at the very least be able to directly compare it on the exact same system.

Right now, eBoostr is giving me an estimated overall 75 to 85% of the benefits of an Intel SSD for less than 1/10th the price. When I spend $500+ on an SSD and close to 3x that in my time I expect more than a possible 15-20% increase in productivity - even from Intel.

I still don't understand how people are so caught up in Benchmarks when those benchmarks don't even hint how diverse the performance of 'similar' SSD's are.

Yes, ***any*** SSD will be an improvement over a mechanical hard drive. Keep in mind that major improvements can be found between SSD's too - just that those improvements would not be easily found in simple benchmarks.

Even more interesting is what is promised with Braidwood in the near future or even eBoostr today. Or, even better, an optimized combination of Braidwood/eBoostr technology alongside SSD's.

For my desktop, I am now considering two 80GB G3 Intel SSD's in RAID0 to replace my boot/program drive VRaptor. Why?

Because I have sufficient RAM and a fast enough CPU to be able to fully use the benefits an SSD offers (which I unfortunately can't say for my notebook).

My SSD experience has reaffirmed what I've always known; any single component does not stand alone - it is the 'system' in total that matters, where performance is concerned.

Hope some have found my findings informative.

 

That firmware does not contain the string '18C' so no garbage collection on it. Commander Wolf knows a bit more about this particular drive I think, but if it is not using the new Samsung controller yes it would have stuttering issues.

Hmmm...fully use the benefits of the SSD. Unless your notebook is limited to SATA 150 speeds, you are getting the full use out of it.

So how is this related to the SSD? It is not.

Again, check the firmware and the controller. That sounds like the old Samsung drive right now.

If you go over to OCZ Technology Forums, you'll be able to read up on the Summit. Yes, it has its disadvantages - some still come with the non-GC firmware and have to be sent into OCZ for an upgrade, no TRIM support yet through Samsung will release a user-flashable updater soon to enable TRIM. But it has its advantages - no stuttering according to reports on that forum and others, much cheaper than an X25-M or OCZ Vertex/Agility, and if you have a GC enabled firmware you are good to go.

 

I have samsung 256G with VBM19D1Q, but I havent had any stuttering problem. Greg is right, this firmware doesnt have GC feature which was also confirmed by Samsungsemiconductor.

 

Greg Ross,

Thanks for the comments - if there is no GC on this drive, then why does the drive light stay on non-stop when you log off the computer (but the computer stays on) - even 8 hours or more later? Remember this was delivered just 6 weeks ago on a $6,000+ notebook (Dell).

The notebook may be Sata2, but because the VAIO's CPU is running at 2.26 Ghz, then an SSD would be a waste for this specific notebook for 'real work' (I can produce up to 40GB of RAW NEF image files in a single shoot).

As to how a 3+ minute reboot is related to SSD's? This is normal for you? The AMD Opteron desktop with the Intel G2 shut down in 2 to 5 seconds and rebooted in less than a minute. The Quad Core Extreme notebook with the Samsung drive took over three minutes to shut down and reboot. Even after 3 days (24+ hrs) of 'GC' it is still doing the same thing and when it is booted up again, it still 'stutters' by simply right clicking on a desktop icon or using multiple programs (three or more).

I can see that there are some options (send the drive to Samsung/Dell to get the newest/best firmware), but these options are to me 'ridiculous'.

I can also understand that the cost of these drives is in line to the competition on pure 'specs' and 'benchmarks', but if someone gave me this drive I wouldn't run it as my work system - it is simply below the performance of a mechanical hard drive - at least in terms of consistency.

Again, I believe this drive has 'GC' from the fact that the drive is going non-stop once it is logged off.

If you are right and Dell shipped less than two months ago an SSD on a $6,000+ notebook that is this bad, then it simply proves what Anand said about 'keeping your fingers crossed and praying' to get the latest firmware version on a Samsung SSD.

 

ronan_zj, I see that you also have the 256GB version (my client has the 128GB one) and you also have the Intel SSD too.

Oh! You also have the same model notebook as my client!

Edit: Sorry, you have the E6400, he has the M6400.

Is the Intel the system drive, or the Samsung?

Also, has Dell released any firmware upgrades for the Samsung (I couldn't find any) or have you otherwise updated the firmware on your Samsung?

 

I do not know. While I admit a drive that is in a notebook that is 6 weeks old should have GC on it, the firmware string tells me otherwise.

I have absolutely no idea what you are saying here. Are you saying that you need a 3GHz CPU to take advantage of a 3Gbps SATA II link?

The crash was not related to the SSD, and it sounds like the program or OS was related to the crash. Again, I'm fairly certain there is no GC on that drive. Do a secure erase and see if performance improves.

Agreed. Right now Samsung is working on a firmware update that we can flash ourselves. Most drives (read most) had better be shipping with GC enabled now.

Even if not, estimates point to an end of November delivery for TRIM firmware from Samsung. Any non-GC OCZ Summit user over at their forums flashes their drive back to factory state every month or two, just to old them over until the new firmware comes out.

I agree that is stupid, but at least they do not have firmware issuse like:
Intel) BIOS passwords caused an earlier firmware to brick the drive
Intel) Firmware update to enable TRIM bricks drive
Intel) If above mentioned firmware update did not brick drive, if you have W7 and TRIM enabled the firmware may brick the drive after *any* reboot.
OCZ Vertex/Agility) Early firmware updates bricked drives
OCZ V/A) Drives randomly stop appearing to the computer

Granted most of these problems have been fixed, but the Summit has not had any firmware related issues that I am aware of.

Once again, I do suspect the firmware is not updated on that drive...or if it is a Dell OEM drive it might even have the older controller entirely. Stupid I know.

I'm going to emphasize a few things here:
1) Buying the SSD straight from the OEM is dumb, considering drives from OCZ and Intel perform the same *or better* for the same *or less* money.
2) The only Samsung drive that I know of that is worth its salt is the OCZ Summit. From what I understand, if you buy a Summit today you should have GC. We do not have any evidence (other than your Dell OEM drive that I think does not have GC) that OEM Samsung drives are good or not - AnandTech is certainly correct in that regard for OEM drives.

 

If you are trying to compare directly to Anand's results then you will need to configure IOMeter exactly the same as Anand. The length of the test, area of the disk to be used, queue depth, and so on can make MASSIVE differences to the final test results.

OCZ used this fact to their advantage by creating an exceptionally light-weight IOMeter test for people to use when benching their old JMicron-based drives.

 

The firmware on my two Samsung 256 GB SSD is VBM15D1Q and I'm not experiencing any stuttering.

 

Here is the thread from OCZ forums Iometer setup and you can check Tony's 4k random writes link for inconsistencies.

 

Again, thanks for the response.

What I am trying to say is that in my situation, spending $600+ on an SSD to streamline the data stream from the hard drive to the CPU (which is all an SSD does) when that CPU can only do it at 2.26 Ghz, is not cost effective. Especially when I am achieving around 80% of the SSD's performance with a $40 product like eBoostr.

The 'crash' as you call it certainly was not related to the SSD - this is simply a Windows 7 compatibility issue.

What is an SSD issue is that even after the cause of the crash was found and solved (removed old Roxio, install new Roxio), the computer is Still taking 3 minutes to reboot - most of that time is spent on shutting down. This is certainly attributed to the SSD as the Intel was in the single digit second range for a shutdown.

Greg, we are on different sides of the following issue; but I would rather have the option to flash a drive (or not) and have it die straight out and need to be replaced (I know how to backup/protect my data), than not be able to run it at an acceptable performance threshold (any mechanical drive).

I am agreeing more with davepermen here with thinking that yes, the bricking was happening, but nowhere to the extent that people think (in other words, a very, very low percentage). Just that those that have had bricked drives are very vocal about it (as well they should be).

As to losing data - I have much less sympathy for those that reported that because any drive can fail, let alone when you are updating the firmware on one.

What I take away from your posts is that the firmware on my clients SSD is what is causing these issues for me (yes, me... he thinks that it's running fine on Win 7 compared to what he was suffering with Vista - even when he has his Intel G2 based desktop to compare it to).

I will continue to be on the lookout for a firmware updater for him and if anything changes, I'll certainly post it here for all to know.

Thank you for giving additional information as it applies to my post on this subject. I hope that with the G3 (from all manufacturers) and later SSD revisions these type of issues will be moot.

I still stand by my statement that in my (notebook) usage case, any current SSD would not provide a cost effective upgrade compared to my Scorpio Blue 500 GB drive and eBoostr combination on my VAIO.

This conclusion is contrary to what I read previously (many times) that 'an SSD is the biggest single performance boost you can do to your system'.

This will also change when/if Braidwood is introduced and natively (to Vista/Win 7) offers the performance boost I'm currently seeing with eBoostr.

To state this another way. I have more than doubled the performance of my notebook (when new) with Windows 7, 8GB RAM ($400) and an additional $40 on eBoostr (when it is available to be purchased). Don't forget that I am using a 16GB SSDD ExpressCard as eBoostr's cache, but I have bought that many years ago (no current cost to me).

Even with the Intel G2 SSD, I would have spent over $600, still be using 4GB RAM and have had at best a 20% better performance 'experience' (if we totally and unrealistically discount the effect that half the RAM would have on my photo editing needs).

So to be 'worse' (4Gb vs. 8GB RAM), for $200 more and to only gain at most 20% performance - I'll state this again - SSD's are in their infancy still - I'll consider them again when their many 'pro's' are not countered by such deal breaking 'cons'. The biggest of which is price currently. But the most deceptive is 'benchmarks'.

I'll also repeat that unless you have maximized the other parts of your system (RAM, CPU, Video - whichever affects your computer usage the most), an SSD will only let you hit your particular performance brickwall faster (and only Slightly faster, if you are able to run eBoostr on that system properly, like I can with my VAIO).

This was proven to me with the Intel G2 SSD installed on my clients AMD Opteron CPU based desktop system. Yes, the SSD posted 'insane' comparisons to my VRaptor/Raptor desktop for HD based tasks (unRARing, installing, etc.) but even to a firmware crippled Samsung SSD based notebook, the notebook hands the Opteron it's A$$ when used how my client needs it used (running AutoCAD) - this 'use' has nothing to do with hard drive speed - this 'use' is all about CPU horsepower and efficiency - which definitely held back the Intel SSD too, I'm sure.

System Balance. That should be our goal.

Which is what my first computer 'mentor' told me almost 30 years ago and it is still true today.

 

jedisolo,

how full are your Samsungs?

How (for what) do you use your computer?

From your sig, I conclude (please correct me if needed) that your usage is very light. This is based on the 3GB RAM, 32bit O/S and same CPU that I have in my VAIO.

Please note that I don't consider gaming 'usage', not because it doesn't tax the system, but because I have no personal basis of reference (never gamed).

Have you tweaked your Win 7 Install to achieve a 'stutter' free system? Or is everything at 'default'?

Thanks for any and all information you can provide.

 

I have ran the Iometer 4k random write test 7 times along with the rest of the tests once and the 4k writes are degraded yet my WEI disk score has only dropped to 6.8. I will idle my system overnight and run the 4k test again in the morning. I will run PCMark Vantage HDD suite first.

 

I also have P128 VMB1801Q Corsair drive and my HDD is as good as the day i bought it in June/July.
I do though normally keep my HD's mostly empty anyway. I have always tended to do this even with a mechanical HD. I must admit i am more mindful to do with my SSD because of what i have read about performance degradation over time. My HD has about 97GB HD space left.
I'm not a power user. I am though enjoying about a 24s boot up in W7 to to my desktop from a cold boot. The very low power consumption is very welcome as well.

 

Rachel,

I can't remember if I mentioned it in my last few posts, but my clients 128 GB Samsung drive has about 55 GB free. Surely this is 'enough' free space no?

Can you be more specific of 'not a power user'? Could a netbook suffice for your needs?

Also, it seems like you have a default install of Win 7 with no additional programs (suites) installed? Is this right?