Thinkpads Forum

The problem is every time I post in this topic, I have a different person trying to analyze my response to someone else and bringing something else up. So I will clarify this post and note it as a response to zhenya.

In my case, the performance hit isn't extreme because as long as any important data I plan on using a lot is on the outer edge is within the (Lets say this for now, because I don't have the specific transfer rate curves) 80-100mb/s zone, then all my frequently accessed, important data will read pretty fast. That outer edge can be 10% to 25% of the disk (Depending on what your disk's transfer rates are on the outer edges.) 10-25% of a 74GB Raptor is a tiny space, 10-25% of a 320GB RE3 isn't. The RE3 does have a higher sequential read/write rate than the Raptor but practically I have found the Raptor outperforming the RE3, perhaps because of the way the disk is formatted at the disk level (Servo/Sector. Not familiar with the specific name.) so that the disk has much better random burst read/write. Now I can argue that 1TB monsters with 500GB platters will perform at astounding levels but because of various factors like SNR (HDD head technology, how close the head flies, etc) these things with high aerial density may or may not perform as well as single platter disks. In my case, I still can use all 320GB of a 2.5", 3.5", or even 500GB HDD while the 10-25% outer edge will be the fastest performing. In that area, all the data I need (OS, commonly run programs, swapfile, etc) will run fast while everything at the inner parts of the disk will be the tradeoff (<40mb/s.) I still can utilize the entire disk. At 10-25% of a 2.5" 7.2k HDD, that is between 32-80GB of data with performance close to a SSD for under 100 dollars, but I still have the remainder of my 320GB to utilize. I can't beat the performance to price ratio there. Of course a SSD will obviously be faster with all 128/160/*GB of everything on it.

My system is very easy in upkeep. The T410 is powered on, as well as my T43 and desktop. When I go take a shower, I launch my defraggers on all disks and hit the proper option (Consolidate or Volatility) with the suboptions enabled and let it do its thing for 5-10 minutes or so. When I get back my system is optimized. If I am doing low intensity activity on the system (IE: Making a post to the forum right now) I may just launch the defragger, let it run while I complete a post. Practically this isn’t hard to maintain as long as you follow good habits. For the average user who may or may not care about upkeep, they may pay the price premium of the base system plus 200 dollars to have 128GB that is constantly fast for a specific lifespan.

I will try to find the information on data retention if I can cite it somehow, most of the data I have gotten on that comes from engineers who are very critical to many of the non Intel, Samsung, WD’s Latest thing SSDs that do not have proper implementation of various things. This includes proper wear leveling algos. On data retention, my sources tell me that Intel, Micron and WD actually do data retention tests with their NAND media by placing them into an oven at a specific temperature, heating them and pulling them out for accelerated testing. Of these three, Micron does not make SSDs but sells the media.

Perhaps the benchmarks combined with a little bit of “I just spent 200 dollars” may have a placebo effect, but practically I have seen people bring in their SSD equipped HP Elitebooks, Macbook Air, gaming desktops and they are still inferior performers to my systems. Now this may or may not be buying el-cheapo SSDs or having really bloated setups. I encounter these types because I happen to live in a relatively “well endowed” part of California with young adults who have a lot of money to throw around. I am however not one of those types.

This specific response about knowledge was directed at Wiz who claims he “works” with storage a lot, therefore making him “immune” to having to even talk to engineers. My engineer friends told me not to bother with him if he thinks of himself so highly, as anything I will say will be obviously “inferior” to anything he says in his mentality.

A lot of SSDs are underperforming, OCZ I know my engineer friends have “issues” with due to some of their… “Ideas” on implementing features that SSDs need. Some of Intel’s, Samsung’s releases and WD’s latest N1x thing seem to hold merit with them. There is however, a bone to pick with many other SSDs on the market, I can say the same about a lot of consumer level drives released by the HDD market as well. I have heard many, many horrible things about specific harddrives just from a simple picture under the cover, from how turbulent the airflow in one drive is and how the head is going to resonate to hell and drop the read rates down.

On data retention, some SSDs and hardware they’re attached to may have a “refresh” capability when they are powered on. I am specifically talking about data retention where I fill them and leave them in an archival environment for a year and come back for data. Or possibly if the SSD or hardware does not “refresh” the disk periodically, there will be data loss. Some may designate specific sectors to the delete query then rewrite to it automatically to ensure data retention. I know harddrives have a refreshing “mechanism” where they will occasionally rewrite data to specific sectors. I am not sure if this is implemented on the OS/FS side or the HDD hardware side, but this behavior definitely exists.

For most users, the bottleneck is not the hardware but them. I forgot where this quote is from, but “Users are at best, pieces of benevolent hardware that cannot do anything to the system. At worst, they are malevolent devices that try to destroy the system.” For the casual user that never defragments, cleans their system, then probably a 200 dollar “user competence tax” would be best for them. Otherwise if they just do some simple things to keep their system maintained, they will be able to retain a much higher performance to cost ratio. As nasty as this sounds, it is really truthful in the sense that some simple steps would allow a lot of people’s systems with 5.4k RPM 2.5” HDDs outperform 35-50% of the SSD market. Might just be over generalization but the SSD market is not filled with glowing examples like the earlier SSDs I mentioned.

I will retain my stance of waiting for major maturation on SSDs within 3-5 years before I would use them unless there is some major breakthrough that makes the rewrite cycles on cells for 10mil under high temperature combined with the most advanced, proper implementation of wear leveling, nil speed degradation and the capability to warn me incase of impending death from excess usage. And that rewrite cycle better not be rated “for the first sector on the first chip.” Until then, I will stay with harddrives for their cost and durability; I haven’t suffered any losses with the harddrives that I have purchased with performance and durability in mind. However some of my “Fry’s special” Seagates-Maxtors… Let me just stop here by saying I am happy I have three-four copies of the files I had on those drives. A little paranoia can help sometimes. That and keeping all my current drives at under 30C.

Thanks for giving me a really good post to reply to zhenya, this was much better for me to respond to than other posts on SSDs.

Navck wrote:This specific response about knowledge was directed at Wiz who claims he “works” with storage a lot, therefore making him “immune” to having to even talk to engineers. My engineer friends told me not to bother with him if he thinks of himself so highly, as anything I will say will be obviously “inferior” to anything he says in his mentality.

Since you didn't understand the first time, i'll repeat: "I don't need to "consult an engineer before i say anything about SSDs". I work with networking and storage so probably have more experience with SSD and HDD than the average user. It doesn't mean i always know best and is the expert, but i can speak without the help of a engineer."

Basically what i said here is that i can talk about SSD without getting advice from an engineer before i speak as a response to you where you basically told me i'm not capable of talking about SSD. As my job is about storage of course i talk to engineers almost every day....that is pretty obvious for most people based on what i do and since they are my colleagues. I am a engineer and talk to engineers. At the end i made it clear that the purpose was not to tell everyone "i know best". You decided to interpret that in your own unique way though. You certainly have you own way of interpret what you read. When i said i work with storage and networking you changed that to "work with storage a lot" and decided that i said i was "immune to having to talk to engineers". Why do you have to overstate everything? You find it really hard to accept when someone don't buy your fairytales.

Still the fact reamains that the majority find SSD to be faster and the durability to be good enough to compete with HDD's. You overstate a lot of things where you claim the placement of the files and defrag do miracles for you which you cannot prove and everyone know that defrag does not have such a magic effect as you claim. Also you claim that you can make you 7200rpm disk much faster than a 15k SAS HDD with your magic. A lot of your statements is against every fact. You might need to consult some new engineers if they feed you with all of this false info. The magic placement of files you talk about sounds great in theory, but actually when you boot your computer the files are not read in the same order everytime. Also the head would have to jump around pretty much either way since you would have to read/write to the registry, mft, swapfile, logfiles, temp files while loading drivers, dll's, program files etc. You make it sound like you did something magic to find a placment of the files so everything is read in a single spin. Ultimatedefrag cannot do such magic. This magic fileplacement is impossible. Of course with fewer files on the disk stored closer together the movement of the head will decrease, but it still won't do a miracle with the performance. With all of this effort the SSD is still faster.

All of your theories about how they create a miracle for the performance does not have such an effect as you claim. SSD is simply faster no matter how much you refuse to accept this fact. In this case i compare a single quality HDD to a single quality SSD which is mean't for the same use. I didn't decide to pick the worst product available when talking about the performance of SSD or HDD for that matter. Of course the benefit might differ between system like the SATA/150 vs SATA/300 discussion, but in that case the HDD might not be the bottleneck in the first place and then a faster HDD or SSD won't have that much effect.

Of course with new technology like this (new compared to HDD) there will be a bigger difference between the worst and best products available and quality might vary a lot. But why should you recommend people to not buy a SSD for that reason? Why not recommend them to buy the correct SSD instead of telling everyone some magic stories that make HDD's into a SSD with regards to performance. Also you been talking a lot about the "data retention" and "durability" and didn't make any exception like this was the case for all SSD's. Now you finally admit that some companies create SSD with quality. Of course you can find bad product in all categories, but no one ever denied that.

You can wait 10 years to buy a SSD if you want. It's your choice, but you shouldn't try to scare everyone else away from SSD based on false info or info that is correct only for the worst products available while there are many other products availalbe with much higher quality. For those that accept the price the Intel X25 G2 MLC is a very good alternative to a HDD for a laptop with higher IO no matter defrag, magic fileplacement or 1 million theories that doesn't change the simple fact. It's also a quality product that will certainly last longer than "max a year". Why do you even make such false statements? There are others as well so Intel is not the only option. You and your engineers should really do some better research if these engineers you talk about really exits.

In several of your posts you are pretty patronizing against those that decided to buy a SSD. You make it sound like you are superior and they are basically buying a SSD just to be hip or because they lack knowledge. Why do you find it so important to scare people away from buying a SSD if they can afford it while almost everyone seems to think SSD is a good choice except for the price? Since most people seems to agree that Intel and several other have quality SSD's that perform better, have you ever considered that you and your "engineer" might be wrong and not everyone else?

Without being drawn into the technical discussion here (because I could care less..show me results), I can tell you that I just installed an X25-M G2 160GB SSD in my T410 and the speed over the stock 320GB 7200rpm HD is like night and day. It is much faster booting and loading apps. Firefox for example pops right up almost instantaneously...try that on a conventional hard drive.

Wiz, let me explain why I have an issue with people who buy SSDs. They happen to behave in a similar way to you. I don't mean this personally but I have been practically hounded on by the SSD crowd locally who claim their "superior performance" over "stupid, dated technology." Now from an engineering standpoint that would be just a stupid comment (Good luck owning anything more advanced than a harddrive, resistive element shaped airfoil on the head to adjust for fly height! The levels of engineering required to something fly at scale speed equivalents of Mach 20 airflow and adjust it's airfoil shape to control fly height, imagine that.) and happens to be directed toward "owners of stupid, dated technology that is obviously inferior to the SSD."

On the engineering thing, their words hold much more value over yours from my standpoint because they practically WORK with these devices or their associated technologies ON A DAILY BASIS. To them you're just another customer who has specific demands, they do the hard work to make these things happen. When they work practically from that level, every little issue that exists they have to work to solve and in SSDs there are so many more problems than you can believe.

And I can prove that but I doubt that any of the previous-HDD-but-now-SSD owners within my area or some on the internet would like to tell the world they never defragmented when I could tell. Sustained HDD activity with random seek to open a small 1MB document? Just imagine how much performance loss is going on when 60% of the drive is fragmented and 1MB files are in 500 pieces. When you talk about my "magic placement" then you need to talk to Microsoft's idea of Bootvis and related to circular spinning media, Accelerated Knoppix. These things place files so that require minimal movement in reading because the files are very closely arranged, practically to the level with harddrives the head is moving minimally to seek.

And for data rentention and durablity, I have said this again and again. Speak to an engineer, I have already shown you issues with SSDs and short of getting the internal specsheets for the NAND media they use that shows how the 100k claims are really 1k combined with how certain manufacturers only say those claims are true for the "1st" or "on the first chip" NAND cells are able to survive those rewrites, then I cannot show you more. Perhaps the data recovery industry with contacts in the SSD industry will be able to help me here.

Nothing I have done is "theory", it is practice and you are trying to claim that I am a fool. You are the one that is throwing out things are "special" and "fool" combined with calling my practice "theory" that is unprovable. If you even knew what happens inside SSDs and HDDs, then everything I have said holds. Show me how my files constantly move around the harddrive enough so that I see performance degradation.

See first post for last parargaph.

And I will still hold my ground, unless WD's N1x is able to demo for me that it has overcome the major flaws of a SSD and is low in price, I am not buying SSDs. They are still immature, go read the post and see how that guy's Intel SSD had issues. You might claim his system is improperly setup and that he is a fool as well just like the others in that topic. But from what I can see, he has a better understanding than the fanclub trying to assassinate his character in that

Edit: Credentials please.

Navck wrote:Wiz, let me explain why I have an issue with people who buy SSDs. They happen to behave in a similar way to you. I don't mean this personally but I have been practically hounded on by the SSD crowd locally who claim their "superior performance" over "stupid, dated technology." Now from an engineering standpoint that would be just a stupid comment (Good luck owning anything more advanced than a harddrive, resistive element shaped airfoil on the head to adjust for fly height! The levels of engineering required to something fly at scale speed equivalents of Mach 20 airflow and adjust it's airfoil shape to control fly height, imagine that.) and happens to be directed toward "owners of stupid, dated technology that is obviously inferior to the SSD."

I'm not sure how i gave you that impression. There is nothing stupid about being a owner of a HDD and i never said that either. In that case i would be among the stupid since i use HDD in my desktop as i said. Neither did i say that it's a stupid technology. I been using a HDD for a long time and will continue to do so for pretty long. I find it very hard to belive the HDD will be replaced by the SSD in the near future so the HDD will disappear. You put all the SSD owners in the same pot here like saying that all people of a kind are stupid because you met a couple of those that hounded you in the past. My personality didn't change when i got my SSD, the SSD is simply faster. I never really felt that SSD owners are different people. If you read my posts i never said that HDD have anything to do with stupid. I just claim SSD is faster in general and especially for laptops. As i also said the results might differ between systems, but the general opinion seems to be that SSD will give you a performance boost. The spec, benchmark and most reviews confirm that statement.

Navck wrote:On the engineering thing, their words hold much more value over yours from my standpoint because they practically WORK with these devices or their associated technologies ON A DAILY BASIS. To them you're just another customer who has specific demands, they do the hard work to make these things happen. When they work practically from that level, every little issue that exists they have to work to solve and in SSDs there are so many more problems than you can believe.

For all you know i might be working for EMC, Hitachi or some other big company that work with storage, but whether they have more or less knowledge than me i don't know and is really beside the point. I'm not here to prove that i know better than anyone else. Also for the record i know about problem with SSD's, but as you said "real use" experience is what matters and too many people claim that the SSD work great. You make too much out of this problems since it's not even close too the big problem as you claim. A lot of people actually proven that already by using a SSD without all of the issues you talk about. You make it sound like "your SSD will break and it will happen soon". Of course some have issues as well....no technology is free of errors and new technology then the chance of failures are usually higher. If you go into a lab where they develop technology and they tell you about the issues they been facing that would most likely scare away anyone. That doesn't mean the product you find at the store contain all of these issues.

Navck wrote:And I can prove that but I doubt that any of the previous-HDD-but-now-SSD owners within my area or some on the internet would like to tell the world they never defragmented when I could tell. Sustained HDD activity with random seek to open a small 1MB document? Just imagine how much performance loss is going on when 60% of the drive is fragmented and 1MB files are in 500 pieces. When you talk about my "magic placement" then you need to talk to Microsoft's idea of Bootvis and related to circular spinning media, Accelerated Knoppix. These things place files so that require minimal movement in reading because the files are very closely arranged, practically to the level with harddrives the head is moving minimally to seek.

You give me an example here where you talk about extreme fragmentation way beyond normal. At least im my country your example is extremely seldom to find. 1mb files in 500 pieces....that is actually pretty good if you manage to do that. You can use the worst defrag application available to prevent such a fragmented HDD or most likely not get that much fragmentaton without any defrag program at all. When i compare a HDD to a SSD i don't talk about such a fragmented HDD. I compare with a HDD that is kept in pretty good shape and the SSD is faster. I never said the file placement is irrelevant, but the huge improvement you talk about i never seen or experienced on my own. Never read any report of fact that confirm the huge enhancement you see either. If i compared a HDD as fragmented as you describe to the SSD the performance boost would be even higher switching to a SSD. I compare a HDD kept in well shape with a SSD and find the SSD to be pretty much faster. "Microsoft's idea of Bootvis" doesn't explain the huge enhancement you see which has been my point from the start.

Navck wrote:And for data rentention and durablity, I have said this again and again. Speak to an engineer, I have already shown you issues with SSDs and short of getting the internal specsheets for the NAND media they use that shows how the 100k claims are really 1k combined with how certain manufacturers only say those claims are true for the "1st" or "on the first chip" NAND cells are able to survive those rewrites, then I cannot show you more. Perhaps the data recovery industry with contacts in the SSD industry will be able to help me here.

And i said again and again that you make way to much out of this problem. Too many people proven you wrong already. Every spec for new SSD's claim that the problem is not even close to the big problem you describe. We have to be a bit realistic and not make the problem 100 times worse than it actually is. I speak to engineers almost every day as i already said, but you must have missed that part.

Navck wrote:Nothing I have done is "theory", it is practice and you are trying to claim that I am a fool. You are the one that is throwing out things are "special" and "fool" combined with calling my practice "theory" that is unprovable. If you even knew what happens inside SSDs and HDDs, then everything I have said holds. Show me how my files constantly move around the harddrive enough so that I see performance degradation.

I never said you are a fool, but without any real report to prove your statements how should i or anyone else just buy your story and figure the world is wrong because you know a enigneer that said so. No one is a fool here, but you are not the only one talking to competent people. You are neither the only one that know what you talk about and have knowledge about HDD and SSD. You are the one that actually act superior and claim that you and your engineer know better. By "theory" i mean you have not provided any fact, reviews, test results that confirm your statements especially about the fileplacement, defrag and lack of speed improvement of the SSD. Then i'm talking about the performance boost that make the SSD superfluous and not a 5% improvement.

When almost every benchmark and spec prove you wrong and too many people by now claim that the SSD gave them a performance boost then you need something more than to claim they forgot to defragment the HDD. A defragmented HDD and file placement help, but not that much so it will still not keep up with the IO in a decent SSD. You keep saying i don't understand how the HDD work and continue to explain how a defragmented HDD and file placement will make the HDD a lot faster...what about some test results to confirm this performance boost. I never denied it will make the IO faster....i just said not as much as you claim. I can find a lot of people that confirm the SSD gave them a performance boost and spec that confirm the speed of a SSD. I could also easily find benchmarks to prove those statements. Could you show some reports that will show the huge benefit of a defragmented HDD with the correct file placement compared to a fragmented HDD with the wrong placement including some real test results?

I'm sure you still believe that a SSD is no faster and too immature technology which is basically your opinion and you as everyone else is allowed to have an opinion. You do have to understand it's hard for people to belive that you and your engineer know better than what they see with their own eyes without any real test results to confirm such a statement? While people see a different result than you do you still tell them it's wrong, but except from giving some info about your system you provide no real test results, facts, reviews or reports to confirm your statements. The results i see after i switched to a SSD is better performance. So if someone tell me i'm wrong i need something more than i know an engineer and he say so or you forgot to defragement the HDD or the file placement is wrong. That is not why my SSD is soo much faster than my old 7200rpm HDD.

Let's drop this endless discussion and provide some real test results, reviews and facts. That way it's much more than just opinions and something that people find easier to accept and believe. Even if we disagree and you don't like SSD owners i hope you understand why a lot of people find your statements hard to believe. I'm not saying you are a fool, but have no idea who you are and what kind of knowledge you got. I need something more than your opinion (and i belive that goes for everyone) to believe what you are saying when i cannot find any real fact to support all of your statements. I can find a lot that disprove your statements though. With regards to the "data retention" i agree with your statements except i think it's pretty much an overstatement and the problem is not close to as bad as you describe it. For some systems where there is a lot of write operations then the SSD might be a bad option because of the shorter lifespan, but then it's way beyond the write operations for the average laptop user. So for most people it's a "no issue".

Provide me your credentials to prove your claim. If you refuse to then I am unable to accept what you say with any level of belief. You are saying you are an engineer at a company that works with storage and speaks with other engineers? Provide it.

http://parts.jpl.nasa.gov/docs/NEPP07/N ... esting.pdf
This gives a good overview of the problem at the level of the media, but this is not necessarily equally applied across all flash media found in SSDs.

http://uflip.inria.fr/~uFLIP/abstract.php
http://uflip.inria.fr/~uFLIP/results/in ... =OCZ_120GB
Nor are their "seek" times as good as they claim. The more chips in a SSD (Likewise: The more platters on a harddrive) the more overhead is experienced by a SSD. This is one of the reasons that SSDs will suffer in "seek" performance compared to a harddrive sometimes (I know, what I say sounds like a blantant lie doesn't it?) combined with horrible controllers (jmicron) or how the firmware handles it. Many drives on the market and many users do not use the "ideal" SSDs.
http://www-db.cs.wisc.edu/cidr/cidr2009/Paper_102.pdf

When I am able to access some specsheets, I will post them. There are binning procedures that are used to separate the media. An example of this would be some of the cheaper flash sticks have the "rejects" down from Compact Flash cards. From this standpoint, SSD makers are free to choose any of their media grade. Now some of the examples I provided are only a fractional percentage of the SSD products you can find on the market (Intel and Samsung, select drives. WD's N1x is just in release.) out of many others that are used. The distribution there is not uniform and I am still holding my ground that many SSDs on the market are flawed more so than they should be.

PS: No country is exempt from the less than competent from the casual computer user who does not understand what a defrag operation is and feels his or her SSD is superior to HDDs universally.

I will bring more in the next reply.

Navck wrote:Provide me your credentials to prove your claim. If you refuse to then I am unable to accept what you say with any level of belief. You are saying you are an engineer at a company that works with storage and speaks with other engineers? Provide it.

Why don't you learn to read what i say before you just assume something. What i said is "For all you know i might be working for EMC, Hitachi or some other big company that work with storage". Take one more look at the beginnging of that sentence and maybe you understand this time. Basically that means you have no idea where i work or what i do so don't presume anything. What i did say is that i work for a Microsoft partner. Based on your claim that goes against what people know from realiable source you are the one that need to prove you have knowledge here though.
Of course i won't give away my personal data here to prove anything. I'm sure you wouldn't do either for you or you engineer. So let's not start to ask each other for personal info. Also i don't care what you believe or not. I'm not the one that have to prove anything here.

Navck wrote:http://parts.jpl.nasa.gov/docs/NEPP07/N ... esting.pdf
This gives a good overview of the problem at the level of the media, but this is not necessarily equally applied across all flash media found in SSDs.

Yes this is a report that is very old considering the development of SSD's during the years since this research so i agree it's not relevant. Also this to test the reliability under special conditions and for special use. So like it say at the end "COTS flash devices targeted for NASA missions be individually screened and characterized before being accepted for use"....well i agree to that statement. How many that buy a SSD plan to go one a NASA mission? In this case we are talking about using flash for people with special requirements. Even if some product might not be good enough for those going to space doesn't really tell much about how the product will perform for a person sitting at home or at the office with his laptop. Let's talk about the general here and not those 0.0000001% of people that have special requirements.

Navck wrote:http://uflip.inria.fr/~uFLIP/abstract.php
http://uflip.inria.fr/~uFLIP/results/in ... =OCZ_120GB
Nor are their "seek" times as good as they claim. The more chips in a SSD (Likewise: The more platters on a harddrive) the more overhead is experienced by a SSD. This is one of the reasons that SSDs will suffer in "seek" performance compared to a harddrive sometimes (I know, what I say sounds like a blantant lie doesn't it?) combined with horrible controllers (jmicron) or how the firmware handles it. Many drives on the market and many users do not use the "ideal" SSDs.
http://www-db.cs.wisc.edu/cidr/cidr2009/Paper_102.pdf

I'm sure you googled to find these. They basically give away some theories about flash and no real results. The first example in your post is at least interesting to read even if not that relevant to the general public. These three basically say nothing. What model of SSD is actually tested here....a OCZ 120gb of some kind.

Navck wrote:PS: No country is exempt from the less than competent from the casual computer user who does not understand what a defrag operation is and feels his or her SSD is superior to HDDs universally.

No one said that, but you always bring the worst possible examples and not talking about the average user. As i also said the SSD perform better then a HDD that is *NOT* even close to being as fragmented as your examples which is way beyond normal. Also like i said i have never seen the magic result form a defrag program that you claim to have seen. The 1mb files with 500 fragments is way beyond the normal in my country and find it hard to belive we understand the defrag part that much better then where you live.

I can give you a test of a product that is actually a bit relevant today:
http://www.tomshardware.com/reviews/Int ... ,2012.html

As you can see the SSD perform better even if not in all tests. I never said the SSD is faster at every aspect, but it clearly show that this is faster than a HDD for general use. This test is a lot more relevant for the average user compared to what NASA find good enough for their missions. Especially read operations are fast and that's what really give a performance boost with faster reboot, faster lanuch of application and so on and the system feel faster and respond faster. I also have to add this is the older model and not the latest that is suposed to perform better and handle data retention better. You should also find a lot of benchmarks here if you dare to check them out.

I will assume you are another user on the internet who may or may not be employed for a non technology related company then. If you "claim" you happen to work with storage and talk to engineers, then back yourself up or don't bring that to the table.

If you're ignoring a fairly RECENT report... That actually shows the wear levelling mechanism and how it functions. It also shows that it is not a solution for "premature failure" from "10 year+" claims.

Uh, that isn't googled. Some of this comes from someone's master thesis citations... Perhaps they have a good reason for citing those. That 120GB OCZ is a Vertex, one of the more "recommended" SSDs on the market. You know, one of those that people think are better performing and have long lifespan moreso than the el-cheapo things from eBay? If you bothered to look on the results page instead of throwing my link aside then you would of caught that.

I am talking about the average user. Perhaps students, their parents, professors, K-12 teachers are not "average" enough for you. I see this day to day from Win95 (Still used) all the way to Vista machines, laptop or desktops, personal or buisness. Do you think these machines are "average" enough?

Reviews today are often paid for or sponsered. I would rather have you bring scientific papers to the table showing how they actually are durable as you think they are. Did I mention data retention? See NASA paper again.

Yes you assume a lot of things and who really cares. I understand how you try to change the subject since all you have done so far is to claim the opposite of what is common knowledge and fail to provide any useful proof. It's totally ridiculous to think i would share any personal info with you. That would be the same as me asking for personal info about you and your "engineer" so you can prove he really exist which i doubt. I'm sure you would love to post your personal info here as well. Let's stick to the subject and not make this into each of us prove our knowledge and use that as some kind of excuse.

You still don't get it with regards to the wear level. No on ignore it, but get real here for a second. You provide a outdated document that says a old technology 2gb flash might not be good enough for a NASA mission. Even if you missed it there been some development during those years and unless someone plan a travel to space or attend to a NASA mission why would they care what NASA might not find good enough. The requirements they have is irrelevant for most people. If you are going to provide some info then try to google for something useful and something that is relevant for most people. If you actually are talking about the average user then why did you provide the NASA document in the first place? No one buy a 160gb SSD to use for 10 years+ and neither did i ever claim anything about 10 year+.

And since you seems to live in the past take a look at this article:
http://www.financialexpress.com/news/mt ... sa/312142/

Maybe your NASA document is a bit outdated? Looks like even NASA found the SSD to be a good option back in 2008. Below is taken from the article:
"Mtron's SSD makes it possible for us to store a large amount of data in the safest way against the Antarctica's extreme condition. We are very satisfied with Mtron SSD's excellent performance, stability, and the cost"

Wow....if only NASA had you and your "engineers" expertise they wouldn't make that mistake, right? Or maybe NASA have great engineers that did a lot more research than you and your "engineer" will ever be capable of doing? Seems like NASA found the SSD products to be good enough and they probably picked the more expensive storage for a good reason instead of saving a lot of money and buy HDD's. Could it be they picked SSD because it's slower, shorter lifespan and more expensive like you claim? Or maybe because it's much faster and reliable enough for their usage and that compensate for the more expensive products? You might have to consider getting a new "engineer"....or is the "engineer" something you made up to backup your fairytales.

What is your point with that benchmark you provided. It basically doesn't say much or prove you point...whatever that is.

To throw away the review i gave you with a silly statement like to claim it's paid by Intel is just ridiculous. In that case i could do the same and no results will be valid. If you bothered to take a closer look you would have seen a lot of benchmarks as well....you think all of them are fixed as well because ALL the SSD manufacturers pay them money? And the HDD manufacturers are too cheap to pay so they suffer in the benchmark? Then maybe every SSD owner is paid as well to say the SSD gave them a performance boost? To state something like that without any indication whatsoever it's true will be speculation only and totally ridiculous. I could give you many more benchmarks from different sources, but you would find a silly excuse anyway why it's not valid so it's useless.

Scientific papers doesn't really test products that will tell how fast they perform or the quality of the product. They test technology and not products. You have to really understand them to figure out if and how that should be relevant for the product you want to buy. You have so far not understood that part since you refer to documents with old info and info that is irrelevant in this context. A document that explain the weakness of NAND doesn't mean it's too weak to be a good alternative to HDD....this is the part you don't understand. There are documents that describe the weakness of HDD as well...i guess you have not read those yet because then you would be out of options for storage. You might as well read about the Intel SSD and find out about the technology they use to minimize the "data retention" problem.

Yes you did mention "data retention"....about 200 times. The NASA document tell nothing useful in this context as i already said. What i have said 200 times now that you just ignore is that your point about "data retention" is irrelevant. You claim the SSD won't last more than max 1 year which is a fales statement for the average user. It will last much longer so why would anyone care about a problem they won't see in several years for a device that will be replaced before that happen.

Please feel free to reply and have the last word where you can repeat what you said 10 times already, but i won't reply to your posts anymore since it's useless. You can continue to say that i'm not an engineer, know nothing about SSD, HDD, google for old documents and what your "engineer" said.....oh and don't forget to mention "data retention" once more. I couldn't care less since you basically say "i'm not wrong it's everyone else that is wrong".

For those that can afford a SSD and consider buying one for a laptop i would recommend the Intel X25 series since it will give a performance boost and the "data retention" is a "no issue" unless you have special requirements where a SSD might not be the best option. Defrag and file placement give a minor improvement, but it's not magic like some claim. Don't get scared away by a few bitter people that for some reason hate SSD and the SSD owners.

Wiz, you're completely derailing anything I have to say that holds importance while you waving are credentials without proof. This is all while you claim that me, engineers who have worked in storage and foresenics people are less important than your claims, backed by reviews. Reviews that are sponsered by the people who send SSDs in.

That paper is from 2008, NAND does not "magically" change flaws in the technology. You treat it as if NAND completely made every flaw it has INNATELY disappear.

I will say again, prove yourself and quit the attacks on my character.

Let me give you an edited quote ("Colorful language"). Perhaps a professional who works with flash based media forensics will have something on wear leveling?

If I have a ________ OS that takes a page and then writes 10k of single bytes in single updates, it is STILL 10k of writes, were leveling only assures that the single block is moved about within the storage 10,000 times.
The correct way is to consolidate the 10k to a single write and do it once.
Yep wear leveling has 'hidden' the failure, but there is STILL 9,999 un-needed re-allocations.
Such matters can only be handled by the OS , or the controller with a ________ load of memory (what about power outages)

Perhaps you should realize these people actually work with flash media at the controller level for a living? Or how about manufacturing processes and how NAND chips are binned?

Harryc wrote:Without being drawn into the technical discussion here (because I could care less..show me results), I can tell you that I just installed an X25-M G2 160GB SSD in my T410 and the speed over the stock 320GB 7200rpm HD is like night and day. It is much faster booting and loading apps. Firefox for example pops right up almost instantaneously...try that on a conventional hard drive.

I 100% agree. Where is my 7200 drive now ? its in the ultrabay as data storage... so all the program file sits on the ssd.

I think the best part that I rate over faster booting/apps loading is not getting pauses from Active Protection when I use the pc while moving around. I hated that soo much before... no shock protection needed with ssd

Navck wrote:The hardcore SSD supporters believe my system should inherently in every manner possible, be crushed by SSDs in performance. However this is not true in that I am getting levels of performance on par or superior (!) with their systems even if they kill services.

Some specific claims to look at: "SSDs are instanteous seek / Sequential read, write, readwrite does not matter."

1. Not necessarily "instanteous", but closer to numbers like .25ms. You can get harddrives to do sub 1ms seek times by having data arranged into clusters (Certain defrag software arranging folders and files within so that they lie within the same tracks.)
2. Yes it does, you try dealing with 4GB files.

The main points of getting a SSD is
1. You can do all sorts of things physically (Tossing it, striking the palmrest with your fist..) it that a harddrive definitely does not like. At this point I would question what your Thinkpad feels too.
2. If you never defragmented your harddrive or bother to go do any major maintance things to your system (Ex. Delete unnecessary temp files, have defrag software quickly fill empty space on outer tracks with frequently used data...) and want a total "improvement", definitely there.
3. This may seem rude, but to brag about how you have the latest and greatest... (But not necessarily true.)

The downsides of a SSD are
1. Immature technology, no matter what people say, SSDs are still being developed. They still have relatively short lifespan and all sorts of issues with data retention over a year or two. The people who say otherwise have drank the koolaid and hate anyone who has proper reasoning.
2. Expensive per GB. This will eventually get better but not compared to the rate of harddrives (Especially when they do all sorts of things (Imagine when your data is written onto a small dozen cluster of crystals instead of hundreds, density greatly increases!)
3. Does not scale in performance like harddrive (Harddrive density leads to read speed while coupled with spin rate...)
4. Wear leveling is only a way to make them last 5 years instead of 1, don't buy what the manufacturers say.

That said, if Seagate tells you that you'll likely see an improvement from buying their SATA III (Only useful for burst read/write benchmarks! You still aren't saturating SATA I...) harddrives with massive caches (Bigger benchmark numbers, minimal real life improvement!), which is as much of a marketing claim as the SSD manufacturers saying their drives won't die for 20 years of hard and intense usage.
Wait 2-3 years for SSDs to improve then buy one, you'll save yourself 300 dollars now and get a higher capacity drive that will survive closer to 10 years with possibly better properties (Like data retention and lower power usage) than you will now. Harddrives still rule data storage right now due to their age and amount of money spent in development. SSDs still need a few years to catch up to a practical level of usability.

Some of your claims are totally false. Although you are correct that the longer your wait the better and cheaper SSDs will be. They are practical today, and offer vast improvements over hard drives except for capacity.

1.) Most SSDs range in "access time" (not seek time, SSDs do not have a seek time) of 0.1ms to 0.3ms. Your hard drive, no matter how you defrag it cannot possibly come close to that consistantly. On a hard drive, access time = seek time + rotational latency. Rotational latency is the amount of time the disc takes to rotate so that the sector you want is under the read head. There is NO way to improve this on a constant angular velocity hard drive, defragging will not help. A 7200rpm hard drive can take up to 60/7200 = 8.33ms to rotate to the correct location. The average is half of that: 4.16ms and this is what you will usually expect to occur whenever you perform a seek. Even a 15000rpm hard drive has a 4ms maximum rotational latency and a 2ms average rotational latency. Yes, half the time you will be below the average, but the other half the time can't be discounted. After the rotational delay you also have seek time. Here your defragmenting will provide improvements, but you will never get anywhere near what you are suggesting. Take a look at the Cheetah 15k.7 which is amongst the fastest hard drives available. Here are the specs http://www.seagate.com/www/en-us/produc ... ifications 2.0ms average latency is consistant with 15,000rpm. The random read seek time is 3.4ms. This means the average read access time of this drive will be 5.4ms. While it will be possible to improve this slightly if you do not use the entire drive, and very possible to reduce this performance, this will always be your average. On an SSD that has been properly aligned and supports TRIM and garbage collection, your average is usually always between 0.1ms and 0.3ms. This is a huge difference. An SSD also does not require defragmentation (in fact defragmentation will perform no benefit and can damage the drive).

I replaced 2 15k RPM SCSI hard drive with one SSD (on my desktop) and the performance difference absolutely shocked me. I use a hard drive for data because SSDs are still not cost effective when it comes to capacity (although very high speed hard drives are not particularly cost effective either).

In the end, like everything, each technology has its own benefits and its own purpose.
SSDs are much faster, cooler and silent
HDDs are much cheaper with much larger capacity
The power requirements are very comparable, some SSDs do better, some don't.

In a desktop the easy solution is to use a small SSD or an array of small SSDs for the operating system and applications and use a hard drive for mass storage. Laptops are a tougher choice unless you're willing to give up a bay for a hard drive or use a usb hard drive, but I would personally sacrifice the capacity in favor of the performance gain, silence and cooler temperatures.

Wow quite a heated discussion. Navick seems like you have a chip on your shoulder against SSD's. Have you tried running a system with a good one yet?

I spent 2 solid weeks doing benchmark and testing with our corporate image on stock 2.5" 5400 to 7200RPM HDDs from T60/T61 notebooks and a couple SSDs including the X-25M. The intel was superior to all other SSDs at the time and is still a top choice now.

My testing found a 5400 RPM HDD took 6 min. from Windows boot to Outlook open and functional, which is all our users care about at our company. Don't even ask how long it took to get Outlook, Word, Excel and IE open at the same time at boot (8 min.). A 7200 RPM HDD took 5 min. The Intel X-25 took less than 3 plus all other apps also benefited from the performance increase.

We were so happy with the ROI on the SSDs that instead of spending millions to replace over half of our notebook population, which are older T60/X60/T61/X61 with the T400s (128GB SSD of course), we decided to replace all notebook HDDs with the Intel X-25 80GB G2, so far we have deployed over 800 of these upgrades and our users could not be happier. No hardware failures yet and in my opinion if you can afford it a GOOD SSD is the best upgrade any notebook user could make.

SSDs do not run cooler than harddrives, I hope you realize those 100+ million transistors do not magic away the heat. Perhaps your usage patterns with SSDs do not heat them up as much.

Good luck making the controller of a SSD magically fast, they do not magically read entire files in .3ms. Those controllers are keeping track of several other processes (As I have said earlier, delete, wear leveling, proprietary processes just to list a few) all while handling data back and forth to several NAND chips inside the SSD. File systems exist for a reason and harddrives are completely fine when they read a portion of that data on the first pass.
300GB VeloRaptors are still cheaper per GB if you have a massive RAID array of those than a SSD.
And SSDs are not silent either, good luck making oscillators magically silent (Hey whats this about Intel CPU buzz? Check the guy on the Lenovo forums with a SSD making noise too)

For me, the 7.2k RPM Hitachi HDD in the T410 boots in 25 seconds and I can load IE8+FF3.6 as well as Word within the next 15 seconds. I don't know what is going on those corporate preloads but I have been able to boot most of my systems in under a minute.

The problem is that SSDs innately have many issues at the MEDIA level combined with the controller inside them having to "suppress" those issues. If you read about that, then you'll understand the problems that SSDs have resolve before they can become a serious competitor against harddrives.

For me, the 7.2k RPM Hitachi HDD in the T410 boots in 25 seconds and I can load IE8+FF3.6 as well as Word within the next 15 seconds

You really didn't want to pull the speed card as an argument while you are running a conventional hard drive did you?. I am sitting in front of a T410 with an Intel X-25 SSD in it, and I can tell you it would put those numbers to shame.

I'm not getting into this discussion again as i already said since it's pretty useless, but there is one important thing to consider when comparing the boot time. Navck said the following in a previous post: "I stripped out some unnecessary services like parental control (services.msc, look up things you don't need), changed some processes to be manual or delayed load". Of course this will not increase the speed of the HDD/SSD, but it will decrease the number of processes started or delay the start and therefore make the system boot faster. So to make a fair comparison you should make the same tweaks or run Windows with the default settings on both systems. I could disable a lot of services as well to make my computer boot even faster, but it says nothing about the speed of the SSD.

I just put A Intel X25m in my T60. I am Running 3 gb RAM AND WIN 7 PRO. The SSD is amazing!!! My computer boots and runs at least 3 times as fast. Outlook opens almost instantly. I use my laptop all day at work so the $240 i got the 80 GB drive for makes it easily worth it. I will probably buy more of these in the near future. Yes it would suck if it fails in a year but hey you could say that about anything new, also it comes with a 3 year warranty from Intel so i am just not that worried about it. I work in IT and have done a lot of desktop and laptop upgrades, the SSD is probably the biggest speed increase i have seen.

There has been alot of numbers and "facts" in this tread, but i think that installing a SSD in a laptop would (regardless of benchmarks, configurations and removed services) have a boost in everyday performance - and i think that is what we seek, faster boot, quicker desktop operations and so on. Still saving money for an Intel SSD

For me with SSDs, some slower opening apps like iTunes and Photoshop so seem quicker, but most things are not significantly faster to warrant giving up that much space and spending the extra money. At least not at this point. Perhaps when they come closer in price and size to platter based drives, it'll be worth the extra money.

I have made a screencast showing how 25 applications load on 7200RPM vs. SSD. A bit unfair using this kind of test, but this truly shows what SSD is capable of!

http://system-log.com/?p=803

system-log wrote:I have made a screencast showing how 25 applications load on 7200RPM vs. SSD. A bit unfair using this kind of test, but this truly shows what SSD is capable of!

http://system-log.com/?p=803

Pretty slick, demo Sure mirrors the kind of results I've seen on my lowly Z61p when I shoved in an SSD...

Care to elaborate a bit on how you wrote the script that opened/managed the apps?

I don't know much about doing that, although I've poked around a bit, investigating macro generators. Haven't had time to get into it much.

@AMATX

This is my .bat file :

@echo off
start "" "c:\programfiler\cpuid\pc wizard 2010\pc wizard.exe"
start "" "c:\programfiler\foxit software\foxit reader\foxit reader.exe"
start "" "c:\programfiler\gimp-2.0\bin\gimp-2.6.exe"
start "" "c:\programfiler\google\picasa3\picasa3.exe"
start "" "c:\programfiler\inkscape\inkscape.exe"
start "" "c:\programfiler\internet explorer\iexplore.exe"
start "" "c:\programfiler\itunes\itunes.exe"
start "" "c:\programfiler\movie maker\moviemk.exe"
start "" "c:\programfiler\mozilla firefox\firefox.exe"
start "" "c:\programfiler\openoffice.org 3\program\scalc.exe"
start "" "C:\Programfiler\OpenOffice.org 3\program\swriter.exe"
start "" "C:\Programfiler\OpenOffice.org 3\program\sdraw.exe"
start "" "C:\Programfiler\OpenOffice.org 3\program\simpress.exe"
start "" "C:\Programfiler\OpenOffice.org 3\program\smath.exe"
start "" "C:\Programfiler\opera\opera.exe"
start "" "C:\Programfiler\paint.net\paintdotnet.exe"
start "" "C:\Programfiler\quicktime\quicktimeplayer.exe"
start "" "C:\Programfiler\videolan\vlc\vlc.exe"
start "" "C:\Programfiler\winrar\winrar.exe"
start "" "%SystemRoot%\system32\mshearts.exe"
start "" "%SystemRoot%\system32\sol.exe"
start "" "%SystemRoot%\system32\winmine.exe"
start "" "C:\Programfiler\abiword\bin\abiword.exe"
start "" "C:\Programfiler\adobe\Reader 9.0\Reader\acrord32.exe"
start "" "C:\Programfiler\cdburnerxp\cdbxpp.exe"
start "" "C:\Programfiler\maxthon2\maxthon.exe"
start "" "C:\Programfiler\skype\phone\skype.exe"
start "" "C:\totalcmd\totalcmd.exe"

I have read at SSD benching forums that running a test like this is not valid as a benchmark. If you run the most heavy loading applications first in the script (in paralell mode), the SSD will shine even more because of the internal paralell architecture - go figure But as a demo of how a user will experience his og hers desktop, this script will work great

Hey, thx for the quick response. Had no idea you were just doing a series of loads & go from a batch file. Does the trick, though

May not be valid as a benchmark, but sure makes a hell of a good demo...

system-log wrote:There has been alot of numbers and "facts" in this tread, but i think that installing a SSD in a laptop would (regardless of benchmarks, configurations and removed services) have a boost in everyday performance - and i think that is what we seek, faster boot, quicker desktop operations and so on. Still saving money for an Intel SSD

Exactly.

I'm a software developer, and (finally) got a chance to use a Thinkpad equipped with an Intel SSD (2nd generation), and the difference between that and a typical 7200 RPM drive (which is what I have in my T60) was both EXTREMELY noticeable, and absolutely desirable. That said, SSDs are more expensive per MB, so obviously they're not suitable for movie/music collections (I have a 2TB and a 1TB WD drives installed in my media center for that).

Anyone who is attempting to make the argument that a well-made SSD with a good controller (such as Intel's X-series G2 drives) is slower than a mechanical hard-disk (2.5" form-factor, since that's what fits in most notebooks), is not a typical user and has not had a typical user-experience.

Most of a typical notebook user's time (if you're a developer, writer, or user of Micrsoft Office, whatever, anything not involving huge GB-sized files) and disk-I/O is dealing with small files. The random seek times on Intel's new SSDs are faster than any mechanical drive available for Thinkpads - period. As long as you're using Windows 7 (which has far better support for SSDs than older versions of Windows), you will absolutely see faster boot times, faster application launch times, etc. Seeing large spreadsheets open up in literally a second, compared to the obvious wait-time on a similar machine with a HDD, was eye-opening to me.

Watching my friend do a build in Visual Studio 2008 in about 2 seconds, when my almost-identical notebook with an HDD takes 8 - 10 seconds was enough to convince me.

Whatever the theoretical possibilities of a perfectly-tuned, perfectly-defragmented HDD with data only at the edges, etc, etc, in real-world use, most of us don't have these "ideal" systems. For the vast majority of users, good SSDs are faster for almost every scenario (with the important caveat being, don't buy a cheapo SSD, because you won't see the same improvement).

^ Well said; matches my experiences exactly.

I even have a ramdisk on my main box(W700), and find it interesting that while it's faster than SSD(as I'd expect), the difference is no longer near as amazing as a ramdisk vs. hard drive. Just shows how fast the SSD platform is. Yes, there may be some write & long running I/O issues, but for 99% of anything I do, with SSD, if you blink, you miss it

I also have a couple cheap SSDs in two Z61p WinXP boxes, and while not quite as blindingly fast as the W700, it sure is nice booting up quickly and having that almost instant response cranking something up.