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

They are priced to what the market will bear. I predict $349 for the Intel 80GB by mid 2009... The sum cost of the parts is far less than that. The early adopters will pay for the R&D costs

 

Stupid question:

What are the *CURRENT* SSDs being offered by Dell in their XPS M1330 line? The "Ultra Performance 64GB" is a +$175 option and the "Ultra Performance 128GB" is a +$350 option.

(-25% with current coupon)

From a performance/bang-for-the-buck perspective, I'm wondering if it's better to get one through Dell, or simply get the base HD option and drop in 3rd party drive...

Suggestions?

 

http://www.dailytech.com/Super+Talent+Announces+Enterprise+SSD+Line/article13635.htm

new ones from super talent..

they don't use jmicron..

soon the new ocz will be there, too.. another fight-round will be fun..

 

New OCZ range being announced later today...

EDIT:

OCZ Vertex Series

http://www.ocztechnology.com/aboutocz/press/2008/320

http://www.ocztechnology.com/products/flash_drives/ocz_vertex_series_v2_sata_ii_2_5-ssd

32/64MB of on-board cache

 

Thus begins round 3 of OCZ's SSDs.

Though if they have a new controler as the press release states, hopefully these drives will be real contenders in the field. While not the fastest things out the read/write speeds are nice.

Plus the price isnt bad for a 250gb Drive. They are almost getting into useful capacities for me (come on 500GB capacity with RAID support)

 

WOW, what a day for new SSD info! Yes!

 

I don't want to subject too much, but I did want to add that 16 GB Corsair and 32 GB Kingston flash drives are on sale at newegg with promo codes ($17 and $45 respectively), and I plan to buy one or the other and try it out with Half-Life 2 engine games just for kicks and see how it compares to my current HDD.

 

At $3.5 per GB, the prices are beginning to become reasonable. Now all we have to do is wait for the other manufacturers to release 250 GB drives and watch them fall even further. Even $1 per GB by the end of 2009 is beginning to look plausible.

 

i hope they fixed the stuttering.. 200 / 160, 3.5$ / gb.. quite nice..

 

Dell uses only Samsung drives. The 64GB is SLC and 128GB is MLC. Lenovo has started using 64GB MLC drives, I hope Dell hasn't downgraded too.

 

Excellent news.
And at $250 for 60GB / $470 for 120GB, I am getting ready to pull the trigger.
Now the question becomes. Do the new controller and the on board cache deliver what they promise?

 

i have X300 with 64GB SSD how to know if its an SLC or MLC ?

 

You could pull it and write down the model number off the drive. A quick search will tell you what drive it is.

 

There is no need to pull the drive to find the model number.

Look in device manager.

 

I'm 100% sure that it is SLC, don't worry. MLC SSD has not been used by Lenovo for more than the last month back or so

 

Yeah but thats alot less fun

 

as if i have device manager, i run linux ubuntu

but well ill bot into windows heh

thanks guys

 

Hmm actually I'm getting 120MB/sec with my T60
http://www.mmu.co.id/intelssd.jpg

I bought diskeeper with hyperfast. I hope this would help.
UPDATE:
Not sure what happened but when I pressed Optimize: the file searching was stuck at 7%, it won't go any further. The HDD Light is on all the time though. and looking at the resource monitor DKService.exe going at 40-60Mb/sec, with Read and write 1GB/min. But still nothing updating on the diskeeper indicator still stuck at 7%.

Latest Update:
First Time Hyperfast defragmentation:
It stucked at 7% for 20min, and then 14% for another 20min, and 18% for another 20min, and finally at 21% it stopped updating for 60min, until suddenly, it showed that the drive already optimized. The whole optimization process completed in a little bit over 2hrs with intel drive X25-M 80GB.

Second Time Hyperfast defragmentation:
It was a lot better, you could actually see the percentage indicator progressing gradually, and the whole optimization process took only 10min.

So I guess the first time optimization did take a while and could be worrying.

 

I'm telling you, from your transfer rates the laptop controller is limiting the drive. You might say 120MB/sec is good enough blah blah, but read transfers aren't the only thing that might be missing.

You have reported the entire system freezing when you get a problem. The drive supports NCQ supporting 32 outstanding operations. Even if the problematic program might be frozen, all other aspects of the system should be responsive.

If you have a desktop with full SATA2-300 support and NCQ, try the same thing that causes the problem. If it doesn't, then its your laptops fault.

There's another user: http://forum.notebookreview.com/showthread.php?t=310097&page=5

which reports stuttering with X25-M using Windows Live Messenger.

You said you don't have problems with IM. I don't either, and the system itself never freezes even though the program itself might. What causes the same drive to have such different behaviors?

One of the things Intel put for the benchmarks is the Outlook 2007: http://www.intel.com/performance/mobile/sata/sata.htm

I think SSD users might be extra paranoid when using it as hard drive technology is stable and most hard drives do not differ drastically while SSDs are new and they differ a lot between various drives. So any kind of pauses that might not be related to the drive itself is reported as "stuttering".

To verify its the X25-M, the following has to be done:
1. Somebody else needs to verify the "stuttering" on another system
2. Someone else with similar Outlook workload should try on a desktop system with full-fledged SATA2 support and have synthetic benchmarks verify it
3. They need to explain the exact behavior. Hard drives crash and freeze too. We normally associate that to the fault of the program. For SSD users, the first thing that might pop up on the users minds are stuttering.

Unless its done, all I can say is people are spreading FUD.

EDIT: It was from a foreign site, but apparently according to them Intel is planning a firmware update for the drive which will raise the write speed from current 70MB/sec to 120MB/sec. Now if its true and the update can be downloaded that would be awesome!

 

it may be that the intels drive has issues when not on full speed (s-ata2). the throttling down may result in some sort of "resonance" that may result in stuttering in some worst cases. i don't have one, i can't test. but i guess on systems with intel in by default, there won't be stuttering. else, f.e. hp wouldn't put it into it's newest highend line (elitebooks).

 

Even better

cat /sys/bus/scsi/devices/0\:0\:0\:0/model (similar files exist for firmware version, etc.)

Or look through the output of "dmesg | less"

 

SSDs aren't leveraging the firmware and controllers used by hard drives, so trying to support older standards isn't free and affects time to market.

My Samsung 0th-Gen SSDs get about 55MB/sec reads on ATA-66 and about 2.5MB/sec on ATA-33! Clearly, they were not concerned with ATA-33 and only implemented enough to claim compatibility. It's sad but not surprising when new hardware only performs optimally (or at all) in conjunction with other new hardware.

 

Here are some tests I have done October 23, 2008:



2nd one November 4, 2008:


Test I did December 10, 2008:


See the first two screenies showed the drastic drop at ~20% point. Now its not there anymore. Dynamically optimizing controller at work...

 

Mine is capped at 80MB/sec. It must be an old motherboard revision that only supports Core Duo, which is what it has. Either way, you can only get a little bit above the 100MB/sec mark with the t6* series. Also the laptop may not be optimized at the hardware level also. I have known some cases where you can't install an SSD into a laptop at all because the BIOS doesn't support it

 

@Inteluser:
Actually my post was towards Zephir since he was saying my laptop is 3 generation behind and therefore only capable of 80MB/sec data transfer. I was going to pointed out that I was getting 120MB/sec on my T60 which isn't that bad. But T60 should be capable of higher data transfer since it's a SATA2 technology.

 

Has it improved the general speed of the SSD? Someone over at OCZ forums has done some initial tests with a Core V2 and it improved performance slightly but it didn't eradicate the stuttering like SteadyState did.

 

Guys, I ordered Solidata X1-128 SSD (128 GB SLC, 240 MB/s read, 220 MB/s write)

ETA is in 4 weeks, and I will post results as soon as I get the drive

Solidata claims some pretty incredible IOPS, so if the drive really delivers those speeds with tens of thousands of IOPS then they have the fastest mainstream SLC drive (mainstream = costs less than a car)

 

I love Spam!!! but not that kind

 

The eagle has landed

http://rocketdisk.com/product_info.php?cPath=14&products_id=109

 

Looks like you raised the price by ten bucks, Jonny?

 

Hey, that's not bad at all for an SLC -- only $6.25 / GB, though it's kind of small for my tastes. I'll buy something like it in a year or two assuming we get a drive 5 times as large for the same price or less.

 

Aha yes! But the lucky people who pre-ordered will still get the old price

And the costs associated right now with producing an SLC 2.5" 64GB SSD that performs like this Mtron SSD are still very high................but we are trying!!

 

Hmmm, according to the screen I should have gained 10% of speed. but I personally didn't notice anything on real life activity. I guess I should wait for 6 months using my SSD before implementing a HyperFast?

And yes the stuttering is still there.

 

Ok I have ordered
MTRON MOBI 3500 SERIES 2.5" 64GB SATA SLC SSD

I hope this would solve the stuttering completely compared to my intel X25-M on my Thinkpad T60

Hey JonnyRocketDisk,
Could you please double check whether I didn't get charge twice on my credit card? I accidentally pressed the final submit button twice. Just in case it went through twice. I only want one harddrive.

 

HD Tune for two 16GB 3500 Mtron's in RAID 0 (HP 8530w, Server 2008 x64, 128KB stripe size):



(Comparison borrowed from Legit Reviews; URL in image)

90% of the performance of Intel's X25 series is pretty nice, esp. considering the price difference between the two. In a year or so when prices come down I'll shoot for 400MB/s; for now, ~200MB/s is pretty nice in a laptop.

Note for the record: the Mtrons have shown 0 sign of the stuttering reported from some other SSD's.

 

hm.. in my case, i have even more than 200mb/s read speed measured.. maybe teh 64gb version. have you made the firmware update yet? (it'll be fun.. booting dos from an usb stick was fun on the quadcore.. strange but .. fun ).

well, i'd like to shoot for 1GB/s next, but > 250..300MB/s isn't really possible right now in a laptop. lets see when s-ata3 will be around, and if it's possible to max it out at 600MB/s..

i think of a hw raid controller, and some more mtrons in or so.. (in the pc)

at least, it would be fun

yes.. that article about battleship mtron was not nice to my brain.. it's now fixed in there somehow..

johny, do i get better pricing when ordering more from you?

 

Hartman - that test was done in a laptop? So that's a software RAID then?

 

maybe some hw onboard raid controller? like the one on my pc mainboard? but i don't know..

 

Why is the Mtron's power consumption so much higher than other SSDs? They consume about as much power as mechanical hard drives.

 

maybe because of their fpga chip (how's that spelled?!!?). i thought they have some reprogramable arm-processor in, or something. but, while i now messed up tons of things in that single sentence, there might be enough truth in it to eat those watts

and besides that, i don't trust any manufacturers numbers. so they may not be 100% right, differences may be smaller to other ssds.. no clue.

 

I'm just curious, but any notebook users out there with the Intel Turbo Memory feature test CrystalDiskMark on it??

I'd like to see what chip they use on the ITM module.

EDIT:

Well, Intel seems to be spot on from the battery life tests over the net. It gets more battery life than the Samsung SSDs which have lower power consumption possibly because the X25-M features better power management features.

As far as power consumption on the Mtron's are concerned, it could be multiple reasons. The Intel chips are on 50nm, and Samsung I believe also uses 50nm. The Mtron chip might use older 80nm or even 90nm process. The controller chip might also use more power.

 

Now, now. That's hardly a fair comparison. Two 16GB SSDs in RAID 0 = 32GB total disk size. A far cry from 80GB in the X25. A fairer comparison (price and capacity) would be a pair of the MTRON MOBI 3500 SERIES 2.5" 32GB SATA SLC SSD in RAID 0. 64GB is closer to the capacity of the X25, but so is the price (2 x $220 = $440 versus $540 for the X25).

The gap is definitely narrower. The extra $100 for the X25 buys you an additional 16GB and better read performance. Also, how many laptops have room for 2 drives in RAID 0?

 

i think he talked about the intel slc..? as the mtrons are slc, too

 

Ahh, my bad. Discussion in this thread has been almost universally about the X25-M, so I assumed it was that drive.

 

no problems i thought so, first, too. then i thought why 16gb ssds? hm..

anyone having the intel slc worst case (4kb) random write iops at hand?
and for the mlc, too.. so we'd be able to compare to a raid0 of two mtron in price, performance, etc..

 

according to intel (first hit on google, hehe), it's 3300iops at 4k random writes.. that's the slc. quite a bit more than my 125iops in raid0

or is that the mlc.. anyways, infos are sort of mixed up. hope anandtech soon returns with a new roundup. intel slc, intel mlc, actual mtrons, new ocz vertex, actual samsungs.. would be nice.

 

MTRON's controller is, unfortunately, implemented in FPGA (Xilinx Spartan if I am not mistaken). I guess this is because volume is too low for MTRON to produce a piece of dedicated sillicon.

FPGAs are very flexible, you can implement the logic and it would behave "almost" like a custom-made sillcion, however the downside is considerably higher power draw.

 

the picture is clearly labeled x25-m 80gb

 

The mtron will hopefully drop in price if/when the 3rd gen MLC drives with CACHE and non-JMicron controllers come out end of this year/behinning of next.
If they can solve the stuttering ...

 

........................