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

Firefox - Use memory cache instead of disk cache
Description: If you use Firefox, there's a way to write cached files to RAM instead of the hard disk. This is not only faster, but will significantly reduce writes to the SSD while using the browser.
Instructions: Open Firefox -> Type about:config into the address bar -> Enter -> double-click browser.cache.disk.enable to set the value to False -> Right-Click anywhere -> New -> Integer -> Preference Name "disk.cache.memory.capacity" -> value memory size in KB. Enter 32768 for 32MB, 65536 for 64MB, 131072 for 128MB, etc. -> restart Firefox

Have you tried manually TRIMing the drive with Intels SSD toolbox?

 

Would doing this be a bad idea if I've had not problems myself. If it means keeping some writes off of my SSD, maybe its not a bad thing. Any cons or gotchas I should know about.

 

I don't know but I just tried it using FF 3.6 RC1 and it seems faster already. I'm using an Alienware i7 M15-x with a slow 500GB HDD atm though. Already a speed improvement.

 

Yeah, this is only with the new firmware v1.5 of course for the Intel G2's.

sgilmore62,

100/70=1.428571, so yes, it is over 40% faster - because everything else is equal between them.

Cheers!

 

tiller, you know as well as I do that those sequential benchmarks are practically meaningless for all intents and purposes. I wouldn't think that there would be a discernable difference between the 80gb and 160gb in real world testing. Are you seeing 40% difference in any real world benchmark comparisons aside from the sequential write tests?

 

sgilmore62,

Yes, real world copying 15GB folders back and forth (client's machine) the speed increase was very marked (he even noticed it; says a lot).

What was even 'icing on the cake' was that general usage was also noticeably snappier - again, even noticed by my pretty non-discerning client.

This 'general usage' was not an 40% increase, granted - but it felt like it was twice as fast.

So, even though I didn't compare the 80GB to the 160GB G2's against each other, effectively, I did - because pre firmware v1.5, the 160GB G2 was performance equivalent to an 80GB G2 Intel SSD.

Cheers!

 

just to make sure: the 80gb has >=80MB/s, tiller is wrong here

aaand

don't turn of the cache. you can't write too much onto an ssd, that's a myth. and removing that cache will make your web experience more slow, as it has to redownload webpages every first time after opening firefox.

useless premature optimisation

tiller, you're completely wrong. it's 100MB/s against 80MB/s so it's not 40% faster, but around 25%. and it's only on the part of an ssd which matters the least for the experience of speed. what it felt doesn't matter, as some of your situations where very slow, not based on the ssd, anyways. case in point: the 80gb gen2 here feels faster than the 160gb gen1. but it could be other system components.

so, lets see if i got my 40gb ssd mail today..

 

cool, sooo... you want a sell any 80GB G2's? lol

 

no..

well i would, but first there are family and friends close to me already lurking for getting one of those beats whatever gets unused by me gets consumed with love by them

 

ok, can't hurt to try. good luck withe the 40GB ones, I'm considering buying 2 or maybe 3 and RAIDing. the 80GB ones are 300$+shipping+tax, (350 or so). bu the 40 can be had for 120 locally.

 

yeah. especially once raid starts to support trim, the 40gb in raid0 will be the killer choice for cheap fast systems.

 

Oh! Hope you receive it soon and let us know how it performs!

Now, I'm not completely wrong; I'm quoting Intel for the ~40% speed difference.

And I'll agree that the 80GB is closer to 80MB/s, but then the 160GB is also closer to 110MB/s too, so... still closer to 40% (37.5) than 25%.

Anyway, my point is that there is a difference on a 160GB G2 between pre firmware v1.5 and after: both in 'feel' and in copying large folders back and forth (real world use). If this indicates and I believe it does, the relative speed difference between an 80GB G2 (firmware v1.5) and the 160GB G2 (firmware v1.5), then I stand by my statement that the 160GB G2 is the faster drive.

Sequential (i.e. 'non-important') speed increase or not, this is my experience.

No doubt about that in my mind - in the current firmware G2's: the 80GB will feel slower (and actually is, too) than the 160GB.

To theorize why the 160GB model is faster; I would guess that latency goes down for writes on the larger SSD. This would be felt across all aspects of the drives performance (as I did notice in actual use) and contribute to a 'faster' experience. Again, I'm not claiming 40% faster in 'general' use - I have no way to quantify that - but the difference was significant enough for it to 'feel' twice as fast.

Cheers!

 

it's not closer to 80MB/s, it's at least 80MB/s (here)

well, copying large stuff around is the only way to ever get a possible 25-37.5% gain. everything else performs identical (which is what ordinary daily usage, and snappiness is about).

and i don't trust your theory at all, latency is unaffected in any form on these devices.

anyways, to welcome my newest friend, the 40gb, my newest track is named by my intel babies.. yes, i'm a geek

 

davepermen,

let's not argue numbers here!

Have you compared (in the same system, same setup/install) the 80GB G2 to the 160GB G2 with the same firmware?

Don't trust my 'theory' - I don't even (that's what theories are for - to be proven) - don't trust me that I (and my client) noticed the jump in performance by just 'zapping' the firmware - but at least don't spread misinformation yourself.

Latency is affected by simply which drivers you load - search back in this thread for proof of this by various members.

Can we hear this new track 'intel babies'?

Yeah, we're all geeks here.

 

no i don't have a 160g2 yet, and most likely won't ever have one. but as all numbers are identical except for the write speed, which CAN (on sequential only) be UP TO 1/3rd faster, there's no way the drive as a whole can be 2x as fast it just can't.

sure, latency is driver affected. not DRIVE affected. they have the same latency specs, 80 and 160 (and even the 40 one).

I'm not allowed to post links.. but there's my sig. with some knowledge about the intels you should find out which one is named after them

 

I am still wondering if intel G1 has its own GC like OCZ and Samsung SSD to avoid the performance degradation coz I am still not clear about this point.

 

well, they came out before OCZ started to abuse GC as a marketing term.

but yeah, they clean themselves up to not degrade massively. they don't need all the fancy words. they just "do it"

 

davepermen,

Sorry, now it's my turn to not believe. Spec's are nothing; real world usage is the only thing that counts. (And, once again... benchmarks are not real world... but...)...

So, does anybody here have access to both G2's with updated firmware? In the same system/install? Willing to verify/quantify any differences between them?

I'd be very surprised if what I saw the 160GB G2 do before/after a firmware upgrade doesn't carry over to the 80GB vs. 160GB models.

Even if I'm wrong, we will still have learned something new (except for davepermen ).

Cheers!

 

well, the upgrade might help the 80gb to gain "felt speed", too? you don't know.

technically, the ONLY difference is the write speed between those two. anything ELSE is placebo. i am all for real world usage, but reality is only what your imagination tells you to believe. so it's just as untrustable as anything else.

the 80gb g2 i have feels a bit faster than my 160gb g1. it might be the actual firmware? or the pc, in my case..

 

davepermen,

I have considered that the upgrade might gain speed too for the 80GB model.

What I am partially weighing my opinion on though is Anand stating that the only thing holding back the title to 'best ever SSD' is the write speeds of the 160GB G2.

Guess I'll have to buy both (when I have the chance and the time) to test this for myself.

Man! You sure put me up to a lot of work.

Good thing.. otherwise I'd be a lazy good for nothing.

Cheers!

 

one simple "proof" of my idea of being firmware dependant, not write speed dependant: the snappiness of your system depends on what? exactly: only read speeds/access times.

so write does only matter if you copy data ONTO the ssd. mainly.

 

Yes the G1 Intel SSD's have self cleaning mechanisms that work in real time unlike Samsung and Indilinx that require the system to be idle. Due to the nature of many notebook configurations, idle time rarely happens with many notebook systems requiring tricky sequences to initiate the self cleaning mechanisms. The G2 drives have improved self cleaning mechanisms with their firmwares including TRIM and manual TRIM tools to schedule self cleaning mechanisms. Intel has statedthat it has no plans for future firmware upgrades for G1 drives.
http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&ProdId=3044&DwnldID=18363&lang=eng

http://www.reghardware.co.uk/2009/04/20/review_update_storage_ssd_intel_x25_m/

 

No, that is not 'proof' at all. Unless you're only considering snappiness of only displaying objects in an O/S. But even that doesn't hold up. Keep reading.

Proof of my idea is that Windows reads/writes constantly - even when just displaying data. It does this by updating the MFT, system files and other parts of the HD that need to write temporary files while the displayed data is being shown.

So, with writes a constant part of the O/S experience, I can see (hope you can too) how a higher write speed will have beneficial changes to the 'feel' of the system.

Cheers!

 

no, it's not reading and writing constantly. and the os-writes are mostly asynchronous anyways.

you don't have to believe that. you can verify that if you learn to use the stuff from sysinternals, and some os level programming to understand how it works.

obviously, a higher write speed CAN help the system. but it's just the MAXIMUM speed here. and that does NEVER matter for os file updates. they are tiny some-byte updates anywas.

 

lol...

Okay, I'll agree that the Maximum doesn't have any bearing on 'feel' but copying folders back and forth is as 'real world' use as a HD (whether a mechanical or an SSD HD) is going to get. Some of this is translated to an O/S and opening/closing programs, no?

But, if the Maximum is increased, then the Average is too - and, I'm guessing, so too is the Minimum. But, we're starting to repeat ourselves creatively here.

Besides - I have to go make some money now, to be able to continue this Virtual head butting in relative comfort later.

Cheers!

 

copying folders back'n'forth is what i do about once every second month.. so no, it doesn't count for me most of the time it's over networks anyways, and most of them are slower than my disks

no, so far, the random numbers for the G2 are about the same for 80 and 160, so for small things, numbers didn't change much.

^^

virtual head butting... for me, it's quite real. except for the 160g2, all is here

 

If so, if I use G1s on Raid system, it will not have degradation?

 

it will have the same as a single one. which is not really much no. it will work well. mine works well >half a year after purchase. and i guess it still will in the future.

 

"></embed></object>

 

Well ok great. I did try the tweek, didnt notice a difference, and switched back to default stage to hard drive. Google didnt net me anything helpful either... well except for this thread

 

haha not exactly, but i had to watch it. somehow scary, and still had to watch it

mine is a different type of music as you might have found out by now

and by now, it's official: today, in about 8 hours, i get it

 

what they are charging for ssds is way more violelent than what this derelict thought biting off a chickens head was 30 years ago.

 

True to form, I am planning my next attack on SSD's! Do you think there will be a clear winner to Intel in 3 months? In 6 months? Will there be a new version of the Intel by then?

 

In 6 months we should see G2.5's.

A clear winner to Intel? lol... not if you allow Intel to respond to any perceived threat.

Why am I so confident Intel will remain on top? Because all the information points to Intel holding back as much a they need to, because they can, right now.

Points:
If power consumption were not a concern, the G2's and even the G1's would be better than what is promised from other companies in a few months from now.

If a new competitor introduces a real inexpensive SSD that is also performing at or above the Intel's, Intel can drop the price in a heartbeat (they have been commanding higher prices for almost a year and a half already).

Mostly though, Intel has such a head start on everyone else with a quality, reliable and dependable product that whatever gets introduced now might compete with Intel in a couple of years at the earliest - when the discerning consumers can finally begin to trust the 'new' kid with their precious business.

Cape Consultant, I truly was sorry to read that you had to sell your G2, but I am fairly confident that what you replace it with will be equal or hopefully better - but it will be Intel.

Cheers!

 

Hey gang...question for the experts but posted abroad....if anyone has any thoughts pls jump in!

http://forums.storagereview.com/index.php?showtopic=28018

I threw it there because someone posed the question to me there originally. I just thought that the answers might lie here...

 

Les,

Somewhere around this linked post (in this same thread, btw) you might want to follow davepermen, darQ96, myself and others as we try to convince darQ96 why TRIM is important and also, what TRIM is.

See:
http://forum.notebookreview.com/showpost.php?p=5663823&postcount=2345


Quickly summarizing it for you:

TRIM is not a marketing gimmick. Mechanical HD's do not care if they overwrite previously used areas - they stay the same speed regardless.

SSD's on the other hand, care a great deal if all the cells have been used. They then must first erase those cells before they can be written to. Worse, because they can only erase in blocks of cells, they must move (needed) data to new cells before they can clean/erase the partially required nand block.

TRIM gives SSD's a chance to do this cleaning/erasing when they feel it's convenient for them (and also when we will notice it the least). This is how the drive stays fast - the O/S tells the SSD that this file is deleted or moved and the SSD clears the nand cells, when appropriate, ready to do the next writes at it's fastest possible speed. In other words, instead of doing a 'read/modify/erase/write' cycle, the SSD simply needs to do a 'write' cycle when we next write to it.

TRIM needs four things to perform:

First, the SSD, M/B and O/S needs to be running in AHCI mode.

Second, the SSD needs to be TRIM enabled.

Third, the O/S needs to be TRIM enabled.

Fourth, the SATA AHCI driver needs to be able to pass the TRIM command from the O/S to the HD.

From most of the information I have read; the only consistent driver that does this is the default MS SATA AHCI driver in Win 7, so far.

Hope this helps.

Cheers!

 

Thanks tiller! I basically agree. Intel, just like with chips, has a heck of a wiggle room when it comes to prices.

They have come darn close to putting AMD out of business a few times simply by dropping prices.

And with their talent and quality, I also suspect that the G 2.5 will be a winner. I will focus on income until then!

 

Then the question for myself would be why canI not observe or document performance drops from ssd that has been cleaned completely and one that has been filled and then overwritten???

Are there documented speed tests somewhere?

 

Les, top of the evening to you! I have been in Canada in the Winter and I know that it is very super cold!

Dave

 

Les,

See this:

http://www.anandtech.com/storage/showdoc.aspx?i=3631


Btw, are you using SLC SSD's? If so, then those Gen1 models had a different way to cleanup after themselves. In the link above, Anand is focussing on MLC (mostly) models.

 

I just got my second intel ssd 80gb G2 and update the firmware. Right now im running intel toolbox and its been 20 mins and im still at 57%. I remember when i ran my first ssd drive it didnt take this long. On this new drive, I didnt install both intel matrix storage driver, and rapid storage driver. On my first ssd I had both matrix storage and rapid storage drivers. Both are running on windows 7. Just wondering why its taking so long to run toolbox?

 

I am wondering if I have the right drivers, etc, because currently, I have 53GB available on the Intel G2 80GB drive, and when I run the Intel SSD toolbox, it takes only 2 seconds for the green bar to move from 0% to 100%, and it says "ready to use". Is this normal? Should it take longer (or will it take longer in the future as I use the drive more and more)? I am not complaining that the process is fast; I just want to make sure that it actually works

Thanks!

 

there's my 40gb baby.. say hi!! (and hi, i5, hi, mainboard, and hi, 4gb ram )

edit: and hi fan! (big shuriken, hopefully usable without fan )

 

Run CCleaner "wipe free space" overwrite with 0's and you will be documenting a slowdown. The operating system will show free space but the controller will think the drive is filled making an erase prior to each write causing a slowdown.

 

Intel X25-V 40GB, G2

 

nice benchmarks, what do you think is the best setup for notebooks, a single ssd + hdd data drive, 2 ssd's RAID 0 or 1 large capacity SSD? I have a Toshiba notebook with 2 HDD bays but no BIOS options for RAID so have to settle for 1 SSD + HDD data drive. If my notebooks BIOS allowed for RAID I would definitely be hitting the buy button @ Newegg for 2 of those 40gb Intel SSD's.

 

what you want. i personally would not buy hdds anymore ever, except to expand my home server. but your mileage sure varies => do what you like

but yeah, two 40gb in raid0 in a desktop would be quite nice.. sort of like the super-80gb intel (same write speed, everything else better). but, so far, without trim. but that'll come, too.

 

I could live with out trim, (I'm using a G1 80GB intel now in my laptop). I'd buy another G1 except they are the same prices as the G2's... so no way.

 

@ davepermen: where did You buy it?

 

Review: 'The Sandforce Cometh'

See:
http://www.pcper.com/article.php?aid=856