Solid State Drive (SSD) - my Tests!!!

[Snoopy69]

I wanted to take a report on the Hama SSDs deliver, I did get on 03.01.2008.

The polished aluminum housing is really very nicely processed. These drives are MTRON OEM-SSDs only for Hama (it´s a reseller), depending on the size of up to 300 € more than the Hama's cost.

http://www.mbu.de/news/SolidDisk.html





System is a ASUS P5N32-E SLI + E6600...

And here are Benches with the S-ATA-Raid-Controller Areca ARC-1210 (256MB ECC-Ram-Cache).

RAID 0-Array...

The write performance with FC test with onboard controller.

Left: Samsung HD501LJ - Right: Hama SSD...The weakness in writing SSD can be seen clearly here ...

The same again, but now with the controller and still ATTO-Bench...

One SSD without Cache...

The majority of controller performance can be seen immediately ...

One SSD with cache. The benefits are really enormous ...

RAID 0 + Cache...


Nearly a halving extractiontime... *ggg*

Without Cache...

With Cache...

RAID 0 with 3 SSDs...

RAID 0 with 4 SSDs... *ggg*

RAID 5 with 3 SSDs...

RAID 5 with 4 SDDs...

Even a few benches with IOmeter...

Mit Workstation-Pattern... *<-klick*

2x SSD @ RAID 0 with Write-Cache

2x SSD @ RAID 0 without Write-Cache

150GB Raptor

Samsung HD501LJ

Mit Fileserver-Pattern... *<-klick*


2x SSDs @ RAID 0 mit Write-Cache

2x SSD @ RAID 0 ohne Write-Cache

150GB Raptor

Samsung HD501LJ

Mit Database-Pattern... *<-klick*

2x SSDs @ RAID 0 with Write-Cache

2x SSDs @ RAID 0 without Write-Cache

150GB Raptor

Samsung HD501LJ

From here all the tests with only one SSD...

The Windows XP Start time:
The difference is really noticeable - of course, at work. One the other hand, is a Raptor Nothing!
This I simply beautiful packed XP on the Samsung Hama cloned, as a fresh OS little anfordernd.

Samsung = 69sek.
SSD = 46sek.

Difference of 50%


Copy / move of large files:
I have a beautiful large selbstgenerierte file of exactly 5GB taken and copying at the click abgestoppt exactly.

Copy within a SDD 36MB/sek approx.
Copying the other SSD SSD 81MB/sek approx.

2.25 times faster ...


HL2 - Lost Coast:
This time I have the time from the double click until the appearance of the menus stopped. More specifically, until the ink on the background image appears. It was always only ONCE HL2 started and abgestoppt since the 2nd Load times already a lot of data in RAM or Videoram. There were each stopped 5 times (also 5-boot), a precise value funds.

Samsung HD501LJ = 27.6 sec.
Hama SSD = 19.5 sec.

Difference nearly 42%

The repeated time after loading end of HL2 showed little difference (basic stands above). Since it is just 0.1 sec. (If any). It is important but the first after loading Windows Start.


Crysis:
This I Crysis on c: \ XP installed (such as HL2). I assumed that the loading time also depends on the settings - so that high resolution + AA / AF / details more data in RAM or Videoram land than with moderate attitude. Therefore, I have the highest setting (1920x1200 + + 16xAA high). I then a batch file for the GPU bench, as the intros, and the menu is being circumvented.
Stopped Auführen was from the batch file to the disappearance of the loading screen. Re with 5 passes (each after a restart).
The drives were (like HL2 also) first with Perfect Disk defragt the game installed and the same again defragt. So all very clean.

Samsung HD501LJ = 44.2 sec.
Hama SSD = 30.2 sec.

Difference approximately 46%


WinRar 3.70:
Here I simply the folder "dllcache" c: \ windows \ system32 \. It is about 450MB in size and includes 2460 files of a few hundred bytes of up to 13MB. Only small files would be naturally better for testing, but here is the Hama very strongly on the Samsung. The result would be even more dramatic with only small files.
Thus, the CPU is not the whole limited WinRar I have the option to "save".

Samsung = 28.5 sec.
Hama = 18.3 sec.

Approximately 56% difference! *OMG*


Windows boot time - vs. Hama. Seagate ST3160021A:
I have the OS on my work PC (Venice A64 3200 +, MSI K8N Neo 2 Platinum) on the SSD cloned. In the PC handycrafts been 3 1 / 2 years with a 24/7-Betrieb Seagate ST3160021A (160GB - 2MB cache). Since the OS is already very crowded and some in the Quick Launch toolbar lands needs the chest about 3:25 minutes until HDD activity no longer be seen. What should I say? The result almost made me away ...

ST3160021A (160GB - 2MB cache) = 3:28 min
Hama = 1:28 minutes! *Aaargh*

Bright 2.36 times faster! *OMG*


Raid array:
So it looks like I can not Raid 0 on the ASUS create when you press F7 should immediately so a pop. It does not (freeze)
That would be very unfortunate if that is not suffering. Had me on Raid 0 pleased ...

The problem with raid on the ASUS is likely to be found. Either Drives are too small (some controllers to accept no small Drives) or the IDs of the drives are not allowed. So even with the 9650 series of 3ware. These controllers support SSD's NOT! Only the 9590 Series supports this, but merely coincidental or unintentionally (as you told me).
The controller of Areca but run well with SSDs.


http://www.nextlevelhardware.com/storage/battleship/

http://www.forum-3dcenter.org/vbulletin/showthread.php?t=400991

http://mtron.net/Upload_Data/Spec/ASiC/MSP-SATA7035.pdf

http://de.wikipedia.org/wiki/Solid_State_Drive

http://www.ssdisk.eu/benchmark_pc_extreme_raid.php

http://www.forumdeluxx.de/forum/showpost.php?p=7989640&postcount=336

http://www.hmk.de/datasheets/PM_HMK_SSD_0307.pdf

http://www.cis.nctu.edu.tw/~lpchang/papers/sac2007_ppt.pdf

 

[Ockie]

Benchmarks looks good, but I'd like to see benchmarks using storage of equivalent value. People like to compare to the stock raptor X, but the mtrons are so expensive, why not compare to SAS with a good controller and RAID? Or what about other meduims?

 

[Snoopy69]

Sorry, i haven´t SAS-Drives, but i know - the SSDs beat SAS
The magic word is "accesstime"

 

[Ockie]

Sorry, i haven´t SAS-Drives, but i know - the SSDs beat SAS
The magic word is "accesstime"

If that is what you want to believe, more power to you.

IMO there is no point in comparing a drive setup that cost a staggering number higher than the other drive you are competing to. You need to compare apples to apples.


Sure a raptor is a fast drive, but comparing a $1,500+ setup to a $150 raptor hardly makes sense or to any other much cheaper drive for that matter.

If you want to start comparing apples to apples, your price range would lie in the FusionIO territory, http://www.fusionio.com/ Or the IRAM, HyperDrive, etc.

 

[Snoopy69]

Your array can be faster - or not (your accesstime)
I don´t like so big arrays of laudly HDDs...(do you have tests, benches?)
When you need compares with apples an apples, than we can foget every tests with SSD.

I know - my setup is very expensive (1060&#8364, but this is only my problem (or not )

What do you know about IOdrive?
I know IOdrives are only suported by linux
And - this drives is NOT bootable

 

[Ockie]

Your array can be faster - or not (your accesstime)
I don´t like so big arrays of laudly HDDs...(do you have tests, benches?)
When you need compares with apples an apples, than we can foget every tests with SSD.

I know - my setup is very expensive (1060&#8364, but this is only my problem (or not )

What do you know about IOdrive?
I know IOdrives are only suported by linux
And - this drives is NOT bootable

Like I said, you are comparing an entire setup worth $1,634 to a single raptor worth 150 bucks.... hardly call that apples to apples.

You tout access time as if it's the only thing that matters. And considering that this is your first post, I am going to guess that your some advertiser.


Oh and lastly, in before the lock:

(17) Posting is to be in ENGLISH only. Non-English posts will be locked or removed.

 

[MC FLMJIG]

Very interesting. Price is a bit out of touch but then again it is SSD.

Nice and can't wait til SSD is MUCH more cheaper.

 

[[LYL]Homer]

Sorry - it´s german...

I can translate - it says, " My e-penis is is bigger than yours."

 

[Snoopy69]

@ Ockie

Forget the money for a moment, please...
This test shows nice benches - NOISELESS!!!

I know - $ 1.634 is lof of money, but not for every user in this great forum (i think).
May be some users are interessted in SSDs (or not).

Because the rules...
Sorry, i don´t seen this rules
What is the problem with non-english posts?

I can to limit the german text of a few englisch words.

I can translate - it says, " My e-penis is is bigger than yours."

This is YOUR translation!!! (not xenophilous )
I came at peace

 

[RangerSVT]

@ Ockie

Forget the money for a moment, please...
This test shows nice benches - NOISELESS!!!

I know - $ 1.634 is lof of money, but not for every user in this great forum (i think).
May be some users are interessted in SSDs (or not).

Because the rules...
Sorry, i don´t seen this rules
What is the problem with non-english posts?

I can the german text to limit of a few englisch words.

Ignore Ockie, he's obviously just being bitter, I really appreciate the benchmarks and that's a great setup you have there, though I would try to get better comparisons along the lines of price ranges ect...

As far as I'm concerned, this post is english, and you're typing in english, there is no difference in someone else posting a link to a chinese website with benchmarks, than having a forum user contribute with his personal benchmarks on this website.

 

[Snoopy69]

@ RangerSVT

thx

 

[iSkylla]

Interesting but the prices of these SSD drives are just ridiculous, just look at the MacBook Air. When these come down in price, I can't wait to get a few.

 

[Old Hippie]

Hey Snoopy69, I liked it!

You did a nice job.

Toooooo rich for my blood, but nice!

 

[Ockie]

Ignore Ockie, he's obviously just being bitter, I really appreciate the benchmarks and that's a great setup you have there, though I would try to get better comparisons along the lines of price ranges ect...

As far as I'm concerned, this post is english, and you're typing in english, there is no difference in someone else posting a link to a chinese website with benchmarks, than having a forum user contribute with his personal benchmarks on this website.

How am I being bitter? All I wanted was some proper benchmarks to something thats more in line with the costs.


Let me quote what you said:

though I would try to get better comparisons along the lines of price ranges ect

That is EXACTLY what I've been trying to say here. Apples to Apples.


I'm not bitter in the sense that I have the money laying around and I am looking for the fastest solution currently, however, I do get a little pissed off when someone is quick to tout their drive (array) as being the fastest on the market and compares it to a single desktop drive at 10% of the value and fails to submit scores of more similar drives.... hardly a fair comparison.

With the price offset, his SSD setup should be compared to a $16,000 array.... as it would be the same offset to compare his $1,600 unit to a $150 raptor (about 10%).


All I want is a apples to apples comparison, not someone trying to tell everyone that it's the fastest drive out there with inadequate benchmarks. You have SAS, HyperDrive, IRAM, FusionIO, etc that all fits the bill and will obliterate a raptor too.


This does look like a promising drive, which is why I really want to see something scaled to the same price metrics.

 

[Megalomaniac]

Nice snoopy!

yes, too rich for my blood, but i see this as being the next "raptor", a raptor-killer?

there might be solutions that are faster but:
1) they are not for regular home use ($1,600 is not there yet, but intel is entering the SSD market, price wars will start)
2) this is totally silent, unless you are using RAM drive with battery, this is best you can get. anyone who has ever been in a server room knows how loud those are.
3) back to 1, a lot of players are entering THIS market. so it will get cheaper. the others will get obsolete. (but hey, i still have an IDE raid 5 array in an old machine, its old, but still works)

Snoopy, heres a possible dilemma - how do you think the RAID chip will scale? will it be able to handle these speeds? with more drives? I can see a problem with the speed of the controller chip not being able to scale with these speeds. In other words, the controller will become the bottleneck.

try a RAID 10?

 

[Term-X]

Man, I don't understand why some people have a bone with this guy. He came here and posted some of his findings, which he didn't have to do. And we have people giving it to him for his lack of english; there is a lot more people with first language being english that don't come off as good as he is. Give him a break and enjoy the results. Yeesh.

Thanks for posting these, Snoopy.

 

[[LYL]Homer]

Hey Snoopy, I didn't mean to sound like a goober - was said in jest and meant to be funny. Maybe it wasn't taken that way.

Or maybe my e-penis was feeling just a little small after seeing your benchmarks.

 

[Ockie]

Thanks for converting it to english. I noticed in your other posts you say this is a MTron MOBI 3000 16GB, but on your pictures it shows 120/80 performance... the MOBI 3000 does not have 120/80 performance, it has a 100/80 performance, which exact model do you have?

Is the SATA interface identical to the SATA interface of regular drives? Especially on the 2.5".... reason I'm asking is if I would be able to use hot swap units with these.

 

[gjvrieze]

Is the SATA interface identical to the SATA interface of regular drives? Especially on the 2.5".... reason I'm asking is if I would be able to use hot swap units with these.

Ockie, I believe the manufactures of the SSDs made them to the same size as 2.5" SATA Notebook drives, so that they would be easily interchangeable.. The connector is in a standard location, just like your 2.5" SAS back plane....

 

[Ockie]

Ockie, I believe the manufactures of the SSDs made them to the same size as 2.5" SATA Notebook drives, so that they would be easily interchangeable.. The connector is in a standard location, just like your 2.5" SAS back plane....

Heh, you said SAS backplane... I think you already know what I'm thinking

 

[J-Mag]

What happens when you enable power management? Does the SSD "spin up" as fast as a standard random access?

 

[MooN]

Thanks for the post. Although the drive is unattainable for me, I enjoyed seeing how it destroys my Raptor. It was fun reading the post, for shits and giggles.

 

[Snoopy69]

Snoopy, heres a possible dilemma - how do you think the RAID chip will scale? will it be able to handle these speeds? with more drives? I can see a problem with the speed of the controller chip not being able to scale with these speeds. In other words, the controller will become the bottleneck.

try a RAID 10?

Yes, you are right. The Areca sclae good with 1, 2 or 3 SSDs, but unidimensional at the 4. SSD...

Hey Snoopy, I didn't mean to sound like a goober - was said in jest and meant to be funny. Maybe it wasn't taken that way.

Or maybe my e-penis was feeling just a little small after seeing your benchmarks.

Ok

Thanks for converting it to english. I noticed in your other posts you say this is a MTron MOBI 3000 16GB, but on your pictures it shows 120/80 performance... the MOBI 3000 does not have 120/80 performance, it has a 100/80 performance, which exact model do you have?

Is the SATA interface identical to the SATA interface of regular drives? Especially on the 2.5".... reason I'm asking is if I would be able to use hot swap units with these.

"Hama" is only a reseller.The declaration 120/90 by "Hama" is wrong - sadly. In this housing an original MTRON Mobi 3000 is working. It´s a OEM-SSD only for "Hama". Hardware-Info shows a "MTRON MSD-SATA3035" (HDtune), but in fact it´s only a MSD-SATA3025 2,5"with 100/80. "Hama" is very bad http://www.hama.de/portal/searchSel...00/action*2563/searchMode*1/bySearch*00090847

What happens when you enable power management? Does the SSD "spin up" as fast as a standard random access?

Hm...
I see no alternation.

 

[mjz_5]

nice job!!! i want one!!! but for 1600 dollars, i'll pass. If they just made an iRAM drive with 8 slots, i can get a 16gb (8x2GB) worth of ram for 400 bucks and it will be alot faster.. . i just dont understand why this is not happening.

 

[Snoopy69]

I payed 1600 dollars (or 1060&#8364 for the first fastet SSDs in germany (2x 400&#8364; for SSDs + 260&#8364; for ARC-1210). This set ist fulsome. I have profits only by copying, extracting an by complex apps.
For normaly (fast) work only one MTRON Mobi 3000 16GB-SSD (280&#8364; or 432Dollar) for WIN XP is enough
Look at the tests after the pics only one SSD). A new installed XP boots in 10-15sec.

 

[DN]

Wow, know thats what i call a real review/test, but i guess ill wait untill prices go down ...

 

[drizzt81]

What happens when you enable power management? Does the SSD "spin up" as fast as a standard random access?

Why would you want to do that in the first place? The SSDs have significantly lower power consumption than standard HDDs anyway.

@Snoopy:
Where did you buy them from? [edit: I live in Germany]

 

[Snoopy69]

@Snoopy:
Where did you buy them from? [edit: I live in Germany]

Bin aus der Südpfalz (sorry for this german words, guys )

 

[drizzt81]

Bin aus der Südpfalz (sorry for this german words, guys )

nice, but what place did you buy them from? A direct link would be nice. Or were these sourced locally?

 

[Snoopy69]

Here... http://www.alternate.de/html/product/details.html?artno=IMEMH5
Or you buy these one. It´s only plastik, but original from MTRON with 3 years warranty. "Hama" has the same MTRON Mobi 3000 inside (as OEM), but they give only 2 years warranty.

http://www.1deins.de/content.php?group=1624
http://www.snogard.de/content.php?group=1622

 

[Eickst]

I have one in a Lenovo X300 ultra-light notebook, 64GB Samsung SSD.

Benchies are similar to the micron, and I believe it was a $700 option (I didn't order it so don't quote me on it).


Comparing even a $700 drive to a $150 raptor (with twice the capacity I might add) is like comparing a 9800GX2 against a 9600GT.

They are nice, but what a ridiculous amount of money they are fetching right now.

 

[Snoopy69]

is like comparing a 9800GX2 against a 9600GT.

I love I-Ram (12.000 IO´s )

http://www.forumdeluxx.de/forum/showpost.php?p=8625990&postcount=2721

 

[Snoopy69]

http://geizhals.at/deutschland/a325793.html

= US-$ 3.816

 

[J-Mag]

Why would you want to do that in the first place? The SSDs have significantly lower power consumption than standard HDDs anyway.

Apologies for the thread res, but I am still interested in this issue.

I agree there is no reason to "spin down" an SSD, but with XP you can't select individual drives for power management. If you have a mixed environment of an SSDs and regular HDs a no lag spin up on the SSD would be nice.