Nice thought there. I would certainly try that out.

well, I really don't halt/reboot my boxes very often (when it happens most of the times is because I built the latest kernel on them) so I may agree with you that usefulness of this is relative and, like explained in its faq, is only for rotating devices: if you get an SSD (also a small 60 Gb one, like I have in my home desktop) you'll see a real difference in disk access (and e4rat won't speed it more).
but maybe dual-booters just reboot more often and can see this as an improvement...

BTW, the point of this was just testing if e4rat works as advertised, and I think we can say it does

1 members found this post helpful.

Oh yes ! I nearly forgot about that, even if I'm a "dual booter". Because I boot on windows anyway all the time because of my "business job" and the tools I use, and fortunately, VM Player allows me to share physical partitions with the native system, which I do (and abuse ). So at the end I rarely boot Slackware natively, and most of the time it's my VM running.

But yes, I remember when I had to boot more often because I wanted to be "native", and ok, there I can see the gain.

Thanks for the hint !

Just jumping in here to emphasize some info and a cautionary point:

First, you'll get a *MUCH* bigger boost from replacing your (boot) HDD with a SSD than you'll ever get from some defrag; plus, by definition, SSD's don't need defragging (provided you set them up correctly in the first place).

Second, to emphasize the point, you should *NEVER* defrag a SSD (or any other FLASH media) due to cell erasure wear.

2 members found this post helpful.

While e4rat is an interesting project, it's about 10 years too late. The industry is already in the final transition from rotating media to flash...

Modern solid-state drives are doing defragmentation internally, it's a side effect of having wear leveling. You can help them by using the Data Set Management TRIM functionality.

The additional wear on a hard disk head assembly by using defragmentation tools is worse. That fastest way to defragment a partition is to backup it completely, erase it and restore the backup. That adds just below one write-cycle (depending on used capacity) and SSDs have thousands of them. Defragmentation is actually less write intensive than that, because it moves less data.

So does defragmentation wear a drive? Yes, both HDD and SSD.

Is it useless on SSDs? Yes.

Does it harm or trash SSDs? No, not really. They're way to reliable for that.

Hmmm - that would entirely depend on your definition of "final transition". If you mean "in the next 10 years", then you might be right. But I can assure you that SSD's for large system databases are just not cost-effective yet. Sure, they're making their way in areas such as high-speed logging, index lookups, cacheing etc, but for massive data storage - nope. Even now, a complete and matching replacement of the single 1 TB hard drive in my laptop would cost 3 to 5 times the price of the rotational storage drive. That I cannot afford right now.

By 2020 magnetic hard drives will be as obscure as floppy drives today. If you didn't notice: Half of the hard-drive vendors already left the market and the remaining ones put a big effort into catching the SSD train.

This thread is about client boot speed optimization. The 2013 client computer OS drive is solid-state, except for low-cost offerings. These low-cost offerings have slow 250 or 320 GB HDDs installed now, and they may migrate to SSHDs soon.

Well, a Slackware full install is below 10 GB. You don't need 1 TB of flash to achieve good results. Think of a laptop with a 256 GB mSATA drive. Then double the capacity every 18 months.

Ok - If we're talking laptops and desktops then I concede defeat. :-)

For big servers using 10's of Terabytes of data we're still going to see hard disks for a while yet.

Let's talk about big data: Facebook stores a huge amount of images uploaded by users on spun down hard disks to keep the power consumption low. 99.9 % of that stuff is never accessed by anyone, but you don't know in advance, which is the important .1 %.

They're looking into a replacement technology called "cold flash", because the access times are naturally bad (up to 20 seconds for a spin-up). With cold flash (cheap, dense, slow with maybe less than 100 cycles), they could replace hard disks, save even more power, while having instant access to the data.

So even on the lowest tier of bulk storage hard disk drives still suck and will be replaced by solid-state storage sooner or later.

At the end of the day, it boils down to the system cost which translates to profit margin for the shareholders.

SSD is easily 10 times more costly than HDD. Until the day we see comparable prices for both technologies, it is unlike for a drastic shift towards SSD.

Pretty much like the war between clean energy (solar, wind) and conventional energy (fossils fuel).

Below very high capacities Flash is actually cheaper than HDD. Have you ever asked yourself, why you don't have a hard disk inside your phone or tablet?

No but I know why my artists colleagues have "old" HDD... Because they need a lot of space to store media they're working on...

And they also have a lot of RAM to work on those media...

Well one day you might be right or not... But it's always funny to see people gambling on the future being "sure against all odds"...

For me future doesn't exist yet... The only things that interest me is what I'll have to finish in the near future... Gambling on what technology will be, is a little like believing cars will fly in year 2000...

Everybody KNOWS that cars couldn't fly because of the 2k bug...

It was gambling back in 2001, where some early birds invested into the solid-state storage business in belief, that it will become a big deal ten years later. The future was set in 2010/2011 as Hitachi and Samsung did exit the hard disk business.

At the end of the day, it's not relevant, which technology you believe in, but what the markets do. HDD production will cease when it becomes unprofitable, regardless of demand. As of today SSDs already have way higher profit margins than HDDs. The rest is pure economics.

Nothing about oil shortage, no war scheduled or any "crisis" huh ? All pure economics...

Yeah yeah... But you were talking about what the market will be in 2020...

Again you might be right...

I don't "believe" in any technology. As I said I don't know future, I'm just working on it. Believe what you want, I'm not required to believe your beliefs, neither I say you're wrong for sure. I just don't like to "believe" that's my point.

And the thread was about does e4rat works now (maybe by extension "is it useful" by my fault in my question) not what the spin doctors have decided for 2020.

I also have mechanical HDDs, for data storage. That doesn't prevent me to also have SSDs for the OS. So maybe you both are right, while SSDs become more common HDDs will not go away soon.