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

As far as I know (from this thread ), as long as you don't install Intel Matrix Storage Driver TRIM will work, and we are waiting for Intel to update it to allow TRIM support.

 

1 minutes? I dont think so

 

more like 3 seconds, yes..

normally.

no, don't mess with the os. (at least win7 100%, vista don't mess much with it). both of them handle an ssd very well by default. don't bother tweaking it.

 

its more than that tho, in my case, ssd tool will fill up my hd, then release the space, usually takk me 20 mins

 

each time? it did it for me slower once.. from then on, seconds at most, every time.

 

rightnow, ssdtoolbox seems doesnt work. everytime when it loads at 8%, it stops, the C drive has been filled up ,but SSD toolbox just stuck at 8% for drive C. no problem for my D and E partition tho. werid.

 

1 partition, 3 seconds..

 

Are you sure that's trim and not the data integrity check?

 

Last years top SSD controllers were made by Intel, Indillinx and Samsung in that order. Who will be the big three this year? No doubt Intel will be in the top 3 but what about Sandforce and Micron?

 

I think Samsung will always be in there as they are the only manufacturer that has been in the game since day one without any major issues. They seem to have carved a strong steady path and even their non-TRIM drives have preformance that can't be beat by many...

From their, we saw OCZ, Corsair and others reband with the Samsung guts.

Intel, IMHO, has its name and it stand like that of Harley Davidson, Microsoft, and others where name alone allows you a great deal of success. TRIM was a HUGE sales agenda for intel.

Sandforce....hmmm...name of the day which is gaining popularity amongst manufacturers.

I am still looking for an SSD that is top....128/256Gb definitely and, for the money...top dog seems to be the new A-Data S596 at $300 for the 128Gb with the new Controller and 128mb cache.

 

you're still into those samsung thingies? ...

 

Mmmmm...no.... Im close to grabbing this A-Data now yet would like some more info on that OWC 100gb. I am as anti-Intel as I am Mac...oh and Harleys eheheheh (Does honesty get me points here?)

I foresee a purchase in the next few weeks...

Gotta admit though..the Sammys go on and on and on and on and on.... Energizer Sammy.

 

well, what do you do in an ssd thread, then, when you're anti intel.. that's like being live at the superbowl, hating sports.. (and ads).

 

Where was Intel when I started a few years back? Credit to them that they have jumped in feet first and with some success but the thread encompasses just about every other ssd manufacturer as it does Intel. We all need to concede that Intel is an overpriced commodity to start. Unlike others who are still trying to find there place and are selling their ssds at a reasonable price point, Intel gets our dollars simply for their name, just as Harley or Mac do.

Myself, I don't find their drives superior in any way and keep seeing the others jump up...its really still like greyhounds chasing that rabbit around the track now isn't it? Today I see the best numbers from OWC, tomorrow maybe Samsung...and then some other unknown surprises me.

The thing I want most....is that visible difference and an unmatchable start time and I will concede that the visible difference doesn't always equate to the best performance in the end. I am like that though and, lets face it folks, so are the majority of new ssd owners. They are sick of minute plus computer starting times, waiting seconds for the control panel to pop up, and they want the feel of the way it was as I described it in an article once. They want to feel like the computer knows what they are going to do or what button they are about to press just before they do.

And thats how an SSD should feel don't ya think?

On a side note....alot of that, and the ssd, is why I am still such a 'No Page File' fanatic. This setup is ideal for shutting down your prefetch, superfetch, defrag and pagefile... I am some 3 years into running my system with 4Gb and no page file and I will go to my grave stating that there is a visible difference when we force our ram to do its job. We are also doing the ssd good by decreasing the read/writes caused by pagefile. I have yet to lose data even once or get some unusual error.

I will stand by this as I did when everyone laughed as I pushed 64 Bit on the site...and with Dell. I believe truly that Dell used this site in its decision to move 64bit along. We were here with 64 Bit when everyone was saying its a dying fad because software vendors would never follow suit. Dell kept contact and allowed an open door, even going so far as to provide tools just to see where things would end up. It was there that I made a few great and knowledgeable contacts who have since had to find other jobs thanks to Dells outsourcing a few years back.

Anyway...kinda went off base...back to my search.

Oh and.... I want to thank everyone on this thread before you decapitate me for the pagefile stance once again. I don't think we realize the number of readers, consumers and professionals that rely on this thread to keep the industry up to date eheheh. I would go so far as to say that outside articles have drawn on the information presented here as much as the authors have through their own testing....

I only wish the site would see that we were the leader there once and consumers are still looking for that in the ssd industry. We had made a great many contacts from companies such as BitMicro, STEC, Samsung, Sandisk, Ridata, Mtron and Memoright...all lost now...too bad.

Heck...I was coddled into a job offer within the SSD industry in a beautiful part of the US that would have fared well that my wife still doesn't let me live down ...not for a second. Apparently I don't know how to look a gift horse in the mouth so she says; I just stick my head right in and let it clamp down.

EDIT: I just read that entire babble....putting the coffee and chocolate candy away for now eheheh.

 

haha way to show i hit some nerve

all in all, your stuff is all right. but you fail to acknowledge how much the intel delivers all your wishes. forget the past, les. it doesn't matter who was first back then (still loving my mtrons, dancing around ANY sammie). it matters who has constistent great performance RIGHT NOW, and for a fair price.

and in this section, since now OVER A YEAR, intel delivers always at the top end. maybe in some cases, it might not be best. but it's always one of the best choices you can make since day one of the first intel. and that is now LONG TIME BACK.

so, no matter how you hate them, you miss much by not trying them. i hate certain companies, certain products, etc. no reason to not know them, explore them, and try them. know your enemies, dude


all in all, you brought great knowledge back then. by your dismissal of the intels, you showed degression of yourself instead of progression. (degression?? i like that word )


edit: oh, and your pagefile bla is fully placebo, at least in win7, and documented as such. but do what you want. just never preach that false knowledge and chance of crashes (not for you, but for those who read it and try it) to others. it hurts them. still fighting the fight against moronic tweaks all over the world, solving bugs that would never have occurred in the first place. and i will on anyone's feet trying to fight that good fight.. no matter how wellknown and praised he might be. including you, a father for me

 

I've been running no page file on my systems since around 2005 and I've yet to run into a single problem either, I'm on the Les boat here; I want my RAM to do its job and not let Windows decide that data shouldn't be stored in and then recalled through one of the fastest interconnects in my PC.

Superfetch and indexing I have no problem leaving on, but pagefile I don't have any reason to enable when everything runs as it should on my particular system and always has.

 

I could be wrong but I am going to jump out on a limb here. I will start by saying my love for Intel falls short only in the fact that we pay the premium price for their name. You get a 60Gb Intel for what you would get a 128Gb for normally...and thats it.

Praise Intel in technological advance. Where would we be without them??

I follow this thread although I don't jump in as often but I don't see where you are going with the 'Intel is God' analogy. I don't see any of their performance results that cannot be equalled or surpassed with TRIM and I know they don't stand beside the Sammy 256Gb for both read and write when you take TRIM our of the picture. I remember posting my results not too long ago when questioning TRIM and asking why, if TRIM is so valuable, so many drives arent suffering real world and noticeable performance dips as claimed.

Having said that, my new ssd will support TRIM and I do believe in the logical ideal that it professes.

You can't hit a nerve with me. I know it seems like it sometimes because I am one of those people that is writing what he thinks before he considers how it sounds to others...I am pretty upfront for the most part. You may not believe me but I stand by you, for the most part with respect to your tweaking views but...we all love to do it don't we. The public still cries for it.

Save me a second start time and I will thank you forever!!!! ehehehe

My system is tweaked to heck....love doing it. I am working on a new Win 7 Tweaking Guide...decided to do it after so much prodding... I don't claim to be well known, definitely not praised and I am absolutely sure my knowledge of technology doesnt come even remotely close to most on this thread but...

Gotta love it all in the end!!!

Oh... I JUST noticed that your ssd is an Intel, you know the brand you pay premium for the name ehehehe. Maybe I hit a nerve??? Sorry if I did but it wasn't intentional, just me writing before i consider the end result again.

 

Disabling the pagefile is the last thing you want to do with an SSD. The strong suit of any SSD worth it's salt is fast reads so that is precisely where you want the page file. I've tried disabling the page file, and I have to admit that I crippled my system. Re-enabling the page file resulted in a much faster, responsive system.

Having a pagefile on an SSD is one the top 3 benefits of having an SSD, why would you want to cheat yourself by disabling it? The only benefit of disabling the pagefile is conservation of disk space -- the next gen SSD's led by Intel will be 320gb and larger. Conservation of space will soon cease to be an issue to be reckoned with as far as tweaking(crippling) operating systems to maximize space

 

well, you fail to understand how it works and believes it's BETTER to have it disabled. while it's proven to have NO gain (in vista, win7.. winxp it HAD gain).

interestingly, that premium (which is debatable due to the performance and reliability and support gains) is very low around here (switzerland).

well, they won in about any benchmarks except sequencial writes since day one, and never really got taken over by any big margin anymore.

they, in short, set the bar, and still do.

and samsung absolutely FAIL to deliver compared to them. having them here, i can tell from experience (the slc ones where nice, though.. but nothing compared to the mtrons)

yah, tweakers.. freakers..

if you love it, fine. being good, no.

Oh... I JUST noticed that your ssd is an Intel, you know the brand you pay premium for the name ehehehe. Maybe I hit a nerve??? Sorry if I did but it wasn't intentional, just me writing before i consider the end result again.[/QUOTE]
well, you hit a nerve, but not by attacking my babies. but by attacking common knowledge and fair judgement, and instead judged by blind hate against some company for no reason.
i hate such behaviour. i really HATE it very much. verrrry much

 

The page file argument will ALWAYS continue to dumbfound me as to why many, who have so much knowledge, don't see the history and needlessness of it. Its actually listed in a few threads which speak of ssds but the argument is simple.

Virtual memory, or your page file, was created to make up for where our physical RAM falls short. It was created in the days of more powerful OS's and the maximum of 512/1Gb ram available. Its always been a horrible drag on a system just by its very mechanics which I will get to.

The unfortunate part of virtual memory is it does not REALLY give your physical RAM the priority it should have. If my 4GB physical RAM got used first and no pagefile, I would be happy, but it does not occur that way.

Next, page file was really good in a 32bit OS that couldn't recognize all the available resources.

Nowadays the typical system is running 4Gb which was more than adequate with Vista, but is especially so in Windows 7.

Do we wonder why there is a selection in Win 7 for 'No Page File' It allows us to shut it down when we really have no use for it.

Now, by very existance, physical RAM is faster than any pagefile... I don't really believe anyone can argue that which, in itself proves the entire performance theory.

It is ALOT moreso obvious on hard disks because they are so much slower and I cant believe we don't remember the days of XP and intensive applications where pagefile would kill us as it accessed it.

Now, by logic, and for those unaware, pagefile is a portion of your hard disk that pretends to be ram to store information and its supposed to be because the RAM is full or not available. Systems haven't been built like this, however, and like to leave physical RAM available 'just in case'.

So...the system then takes the info and moves it to the pagefile which is on the hard disk. It is exactly the same as reading or writing from the disk because that exactly what it is. All the while, in this process, the hard drive is also fighting for its own proprietary rights in reading and writing. Its really kind of like watching two guys try and drive a car at once. Even though MS has perfected it, we are not so far from the days when you could here the system chugging down because of it.

Anyway.... I know we all know this but don't understand how there is any logic that changes the reality in its mechanics. The only answer always seems to be 'show me the performance increase'

In any case...loved responding and this does tie into this thread because the 'no pagefile' thought process is being introduced into ssds because of our concern of lifespan read/writes which are cut down with the pagefile.

Off to a nice steak, ice wine and good chat with friends. Later all and cant wait until I get back for the yelling at me eheheh.

Oh and back to the Intel drive...how many advertisements do we have to see that says Intel has max write of 99mb/s where my lil ole Sammy 256Gb was around 145 last checked (posted here somewhere).

 

A SSD is limited by the SATA controller, which on most systems is a theoretical maximum of 300 MB/s. The speed at which a processor can extract data from my RAM is 13,107 MB/s.

Why would I want my operating system to have to pull data through an interface only capable of 300 MB/s when it can pull data through an interface capable of feeding it from a source that is 43.69 times faster than my SSD?

 

disabling it gives 0 performance or any other gain on vista, win7. proven and discussed even by microsoft. it only gets into action if an app requests it, or if you're low on ram. IF you're low on ram, the pagefile will resque you from app/os crashes. that's it's purpose by today.

disabling it disables that saveguard which you might never need. disabling it has 0 benefit.

so why, again, should you EVER do it. except for the geekgasm of "i know that, i do".

and how many real life tests and users and experiences do YOU need to understand that it's absolutely UNIMPORTANT how fast that max write speed is.

 

I agree with Les on the page file issue, but davepermen on the SSD sequential write issue.

My bottom line point on the page file: I've never run a program that needed it, and I'd rather my OS store data in a location with bandwidth of 12.8 GB/s than 300 MB/s (Windows will use your page file whether you need it or not).

My point on the sequential writes: Very few people need 145 MB/s sequential writes when the vast majority of your operating system activity drops below, say, the 16K level, which is where the Intel drives are the undisputed leader at this point.

 

and this is what you don't get.

and i repeat it in bold, again and again:

WINDOWS VISTA AND WINDOWS 7 DON'T PAGE OUT WHEN THEY HAVE ENOUGH RAM. THERE IS NO LOSS IN PERFORMANCE EVER AS LONG AS YOU HAVE ENOUGH RAM.

the only reason the page file IS there, is for the moment where you DONT have enough. there, it's slow against crashing. i prefer slow. and you would, too.

 

http://www.pcworld.fr/comparatif/materiel/stockage/comparatif-ssd-2009/434611/

JMF612 doesn't look bad at all, until it gets spanked by random write tests.

Oh, and on a related note, I started getting 2-10 second stutters (Well, at that length it was a full on freeze) on my desktop with Vertex 30GB, and CDM performance had dropped severely. This persisted even after running Wiper. After I updated the firmware to 1.5 (From 1.3) the stuttering stopped. Strange.

 

newegg dropped the prices on the intel drives,whats going on

 

They dropped them to what everyone else is offering them at lol...

 

Well the 80 is 219$ and the 160 is 429$ .I haven't seen those prices elsewhere ...

 

Oh right I guess I'm not too excited cause CA residents have to pay taxes if purchasing from Newegg. I got my Intel 160GB for $450 with no taxes after BCB.

 

i got my 160G for 410 after cbc

 

Where did you buy your SSD for that price? Thanks in advance for the info.

 

Got my 160GB Intel G2 for $426.99 minus $54.90 Bing Cashback from Newegg (total paid is $372.09). Thanks to the poster on here who posted that deal back in Nov/Dec!

 

Now that is sick! I wish Newegg didn't charge tax in CA

 

from circitycity/tigerdirect with BCB around thanksgiving

 

That pre Thanksgiving sale that had the 160 GB Intel for less than MSRP. I forgot about BCB, but I still don't care. $415 or whatever it was happened to be more than good enough

 

OEM 160GB G2 for $429.00 right now on newegg, not too bad. There is only 2% bing cashback for it though.

I am tempted but I think waiting will a while longer will pay off, and I do not need it right now its just a want.

 

Pagefile enable vs. disable
Intel SSD vs OCZ SSD

All these arguments make me think plateau is being reached in computer hardware development.

Me on the pagefile issue: I'm usually the type to just leave it alone. I'm pretty sure most SSD manufacturers know that users leave it alone and design their algorithms in such a way. I think as long as you have 4GB RAM or more though, the advantages and disadvantages become purely individual.

Intel SSD vs OCZ SSD: A slow transfer SSD might still have fast response times, but I still think sequential speeds will matter. Like those people that prefer to get high-end for "quality" even though it may not be the case, I'll get an Intel on the high-end.

But for low price SSD? If I was going to spend money in a system where cost becomes an issue like a Netbook, I'd go for a similarly priced, but 50% larger OCZ SSD instead. It's kinda entertaining how merely a year ago dave was very doubtful on the Intel SSDs, but now he swears by them. If someone were going to get a first generation OCZ Core SSD based on the JM602B controller, I will smack them in the head. But a Indilinx based Vertex? That's not a bad choice.

The difference between 13MB/s random writes and 30MB/s on the Intel drives are so theoretical that I think some people argue purely because they have nothing else to do.

I've been running some benchmarks on the Core i5 661 system, and I lost about 10% performance in frames with most games after free space went from 25GB to 15GB. The larger capacity OCZ drives will further make up the difference when talking about 40GB X25-V vs 60GB Vertex.

OCZ Vertex is kinda more balanced over X25-V with miserable write speeds made worse by lower capacity.

The X25-M with 10K random write IOPS with 50% free space can drop to around 1500 with 7% free space.

 

this. forgot to mentoin it but it's very important (and the reason one shouldn't tweak, and the reason such tweak guides are NEVER really helpful, les).
an os is by default used in it's standard installation setup. that's what the hundreds of millions of users base their systems on. that's what EVERY hw and sw developer works for, tests with.
and my argument that there is no loss in HAVING a pagefile, versus there is a possible loss disabling it holds true, too

interestingly, even the x25-v with TERRIBLE write speeds happened to perform about the same as a samsung announced with 3x the write speed. an that is the low end budget version. so while they do matter, they don't really often matter in real life usage.

that's true. then again, here the prices are about equal in $/gb.

no we do discuss them because it's important for people like f.e. les to get rid of their false belief in some numbers to be important.

depends on the costs, and the storage requirements. but it definitely matters. but the write speeds interestingly don't matter for os and app running. but the additional space might be great, sure.

like, uhm, about any ssd. even the so hailed (by les and others) samsung. they completely stop performing when filled. i see some 10 iops in such cases on them. talking about walking away for 10 minutes till a click gets registered.

what ever you fill completely will be very hard to fill further. this is common sense. true for ssds. true for hdds. true for your living room...

 

Is there any info on this OWC drive that keeps getting mentioned? Looks interesting. As an aside, I am starting my core i5 build this week. And have already moved to Win 7 64 Bit.

 

Has anyone really reviewed/tested this??

Is it the same unit that Anand tested?

 

Skal vi snakke Norsk??

Don't know, but was there any English in that thread??

Okay, okay - you guys are used to reading these results...

 

http://translate.google.com/transla...on.no/index.php?showtopic=1149834&sl=no&tl=en

 

Technology makes it too easy nowadays!

 

3.1 swiss franks / gb for the intel 160gb gen2
3.5 swiss franks / gb for the ocz agility
3.8 swiss franks / gb for the ocz vertex


just in case, we can look at the cost in public.

with money conversion and tax reduction we get to
2.65985606 U.S. dollars / gb for the intel.
3.26046872 U.S. dollars / gb for the vertex.

but as we all know, intel sells their disks at a much lower price (2dollars / gb i think?) rest is the seller.

 

Hibernate is not ideal for SSDs, based on the idea that the more times you write, the less the life of the SSD, regardless of their long long life cycle - it is 'bad' none the less, as is turning on/off your light bulb.

With this strict line of thought, is it 'bad' then to sleep an SSD? I suppose sleeping only writes to the RAM, but is the SSD being used during sleep? Presumably, there are writes and deletes on an SSD into and out of sleep?

 

with this though, it is BAD TO USE AN SSD. because, you know, one day, it might die..

why do you life actually? because, one day, you might die, too.. (actually, you will, i'm quite sure about that).

what have we calculated? how many millions or billions of times do you have to hibernate to actually kill your ssd? in short: it does NOT EVER MATTER. just use it.

 

And does hibernation really matter? Only if you use it 10 times or more per day - but even then, Win7 only write "actively used RAM" to the HDD/SSD.

Wear levelling should ensure equal writes

 

Can you link me to this calculation. I am new to this. I'd like to see the math.

Thats quite easy to do. People close the lid when they walk out of the office, or take their notebooks for lunch of coffee. I can see people opening and closing a notebook, either sleep or hibernate, many times a day. I myself do this at least 6 times a day on a normal routine and what not.

 

For short "breaks" you're possibly better off using "hybrid sleep" where the laptop goes into standby and will only hibernate if the battery gets low - much quicker.

But even then - yes, you will add a lot of writes to your SSD, but honestly, does it really matter?

 

no, as i said i don't know where anymore.

but in short:

hibernation of a 4gb ram based os will take at most (win7 most of the time much less, just the ram you actually have in use, without any cached data. in my case f.e. right now 1.7gb) 4gb of writes onto your ssd, per hibernation.

now you have 10000 cycles per cell, and you have 160gb worth of cells (on the 160gb intel ssd that is).


this means, to fill the ssd ONCE, you have to write at least 160/4 times to it, or 40 times. that's 40 hibernations.

now you have to do this 10000 times to kill it. so we're at 400000 times.

now on win7, your average memory footprint will be less than 4gb (see my example, 1.7gb right now)..

so you're definitely >500thousand times you can write. sorry, no millions

but then again, if you do 10 per day, then you're at around 140 years till it dies.