SSD Raid0 Question
Page 1 of 1
bxrdj




Posts: 1469
Location: Far from Home
PostPosted: Wed, 7th Sep 2011 01:14    Post subject: SSD Raid0 Question
Does anyone know if Raid0 is possible with a dual SSD setup over Sata3 connector, and what would the read/write benefit be compared to single drive?

I have a EVGA board - evga x58 sli3 with dual Vertex3 SSD's ... so if my single ssd read is 550M what would it be under raid0 (if possible) and how would that work ... also, I have additional 4 spinners that I use for storage and curious how would I connect these in the case I run the ssd's in raid 0?

Thank you very much in advance!


fuck ...
Back to top
Werelds
Special Little Man



Posts: 15098
Location: 0100111001001100
PostPosted: Wed, 7th Sep 2011 11:23    Post subject:
How you connect the storage drives wouldn't change.

Running SSDs in RAID 0 is not a good idea though, because you'll lose TRIM. Intel said they were working on TRIM+RAID support about a year and a half ago, but they never released drivers. That board uses Intel's controller right?
Back to top
bxrdj




Posts: 1469
Location: Far from Home
PostPosted: Wed, 7th Sep 2011 18:48    Post subject:
yea it does ...I pretty much got to the same conclusion ... oh well, I can just drop it as an additional storage over the sata3 connector and just install games and resource demanding apps on it Wink


fuck ...
Back to top
Werelds
Special Little Man



Posts: 15098
Location: 0100111001001100
PostPosted: Wed, 7th Sep 2011 19:44    Post subject:
Keep one drive for Windows and the apps you gotta reinstall on format/image restore anyway Smile

Stuff like Steam fits perfectly on that second SSD, as all you gotta do is reinstall the executables and it'll pick up your cache files again.
Back to top
Shoshomiga




Posts: 2378
Location: Bulgaria
PostPosted: Wed, 7th Sep 2011 21:01    Post subject: I have left.
I have left.
Back to top
timechange01
VIP Member



Posts: 6650

PostPosted: Thu, 8th Sep 2011 08:00    Post subject: Re: SSD Raid0 Question
bxrdj wrote:
Does anyone know if Raid0 is possible with a dual SSD setup over Sata3 connector, and what would the read/write benefit be compared to single drive?

I have a EVGA board - evga x58 sli3 with dual Vertex3 SSD's ... so if my single ssd read is 550M what would it be under raid0 (if possible) and how would that work ... also, I have additional 4 spinners that I use for storage and curious how would I connect these in the case I run the ssd's in raid 0?

Thank you very much in advance!


Are you actually getting those read speeds though? Run a drive benchmarking tool (AS SSD or ATTO) to double check your speeds please



ASUS Maximus XII Formula | Core i9 10900k @ 5.2Ghz | 32GB G.Skill DDR4 4200Mhz | EVGA RTX 3080 Ti FTW3 Ultra Hydro Copper | ASUS ROG PG35VQ
Back to top
Werelds
Special Little Man



Posts: 15098
Location: 0100111001001100
PostPosted: Thu, 8th Sep 2011 09:16    Post subject:
Please don't, benchmarks are bad for SSDs ;p
Back to top
timechange01
VIP Member



Posts: 6650

PostPosted: Thu, 8th Sep 2011 12:18    Post subject:
No its not. ive had my SSD for several months and haven't lost a drop of performance.

The reason I asked him to benchmark his Vertex 3 is because he might be bottlenecked at 400MB/s read speed with his onboard Marvel 91xx SATA III controller, which would defeat the whole purpose of getting a second Vertex 3 for raid 0

So yes, please benchmark your drive before you go ahead and purchase a second drive for raid 0



ASUS Maximus XII Formula | Core i9 10900k @ 5.2Ghz | 32GB G.Skill DDR4 4200Mhz | EVGA RTX 3080 Ti FTW3 Ultra Hydro Copper | ASUS ROG PG35VQ
Back to top
Werelds
Special Little Man



Posts: 15098
Location: 0100111001001100
PostPosted: Thu, 8th Sep 2011 14:48    Post subject:
timechange01 wrote:
No its not. ive had my SSD for several months and haven't lost a drop of performance.

The reason I asked him to benchmark his Vertex 3 is because he might be bottlenecked at 400MB/s read speed with his onboard Marvel 91xx SATA III controller, which would defeat the whole purpose of getting a second Vertex 3 for raid 0

So yes, please benchmark your drive before you go ahead and purchase a second drive for raid 0

Most benchmarks are bad, because they use up write cycles. Even if you don't notice it straight away, it's still a whole lot of wasted write cycles. If you benchmark it weekly, you are going to see it degrading months (if not more) earlier than it would with normal usage. "Several" months is hardly long enough to judge that. How the fuck do you suppose sites used to test longevity of these things? Wink
Back to top
bxrdj




Posts: 1469
Location: Far from Home
PostPosted: Thu, 8th Sep 2011 21:00    Post subject:
So basically this is what happened ... I rder a lot of stuff from newegg to my office, so when I received my first ssd I noticed that it was dropped off by DHL which was odd since packages always get here via UPS. Anyway, on the newegg website - there was a note - "dropped off at the main desk". So even tho I received the SSD i connected to newegg support chat and told them I never received it ... so they just sent me another one via ups Wink (im going to hell hahaha)

Anyway, I am not going to do a Raid0 setup with these after reading everything, im just going to use it as a secondary game storage drive ...


fuck ...
Back to top
timechange01
VIP Member



Posts: 6650

PostPosted: Fri, 9th Sep 2011 02:17    Post subject:
Werelds wrote:
timechange01 wrote:
No its not. ive had my SSD for several months and haven't lost a drop of performance.

The reason I asked him to benchmark his Vertex 3 is because he might be bottlenecked at 400MB/s read speed with his onboard Marvel 91xx SATA III controller, which would defeat the whole purpose of getting a second Vertex 3 for raid 0

So yes, please benchmark your drive before you go ahead and purchase a second drive for raid 0

Most benchmarks are bad, because they use up write cycles. Even if you don't notice it straight away, it's still a whole lot of wasted write cycles. If you benchmark it weekly, you are going to see it degrading months (if not more) earlier than it would with normal usage. "Several" months is hardly long enough to judge that. How the fuck do you suppose sites used to test longevity of these things? Wink


So then he is just supposed to assume that hes getting the same read speeds as manufacturer specs??

If you look up the issue of SandForce 2200 + onboard Marvell 91xx SATA III controller (which his motherboard has), I think you will change your mind about that.

And Ive had my SSD since the very beginning of February and Ive been running benchmarks very often on this drive because I was curious to see if performance does degrade and my benchmarks are still identical to day 1. If it takes over a year or more for a 64GB SSD to degrade by a few MB/s, who cares? Hes not using it for storage and Im sure hel upgrade by then.

So yea dont just believe manufacturer specs for the drive, you gotta run your own benchmarks because you dont know if your SATA III controller is bottlenecking the drive or not

bxrdj wrote:
So basically this is what happened ... I rder a lot of stuff from newegg to my office, so when I received my first ssd I noticed that it was dropped off by DHL which was odd since packages always get here via UPS. Anyway, on the newegg website - there was a note - "dropped off at the main desk". So even tho I received the SSD i connected to newegg support chat and told them I never received it ... so they just sent me another one via ups Wink (im going to hell hahaha)




Yea they dont send these SSDs via UPS because the box is so small. Im wondering if I should do the same as you Laughing But they already know that Im aware they send SSDs via DHL since I got my current SSD from newegg and also I have everything delivered to my home



ASUS Maximus XII Formula | Core i9 10900k @ 5.2Ghz | 32GB G.Skill DDR4 4200Mhz | EVGA RTX 3080 Ti FTW3 Ultra Hydro Copper | ASUS ROG PG35VQ


Last edited by timechange01 on Fri, 9th Sep 2011 02:50; edited 1 time in total
Back to top
timechange01
VIP Member



Posts: 6650

PostPosted: Fri, 9th Sep 2011 02:39    Post subject:
According to Vertex 3 manufacturer specs, the SSD should give read speeds above 500MB/s. Heres a random Vertex 3 running on Marvell 91xx SATA III controller:

...not exactly in the 500MB/s range as the manufacturer specs state

Source: http://www.ocztechnologyforum.com/forum/showthread.php?85384-Vertex-3-on-X58-w-Marvell-9128-or-9130

I have a 120GB Corsair Force 3 coming in the mail either tomorrow or Saturday. Manufacturer specs for the Corsair drive state 550MB/S read speed but Il have to see for myself if Marvell SATA III controller bottlenecks the drive or not



ASUS Maximus XII Formula | Core i9 10900k @ 5.2Ghz | 32GB G.Skill DDR4 4200Mhz | EVGA RTX 3080 Ti FTW3 Ultra Hydro Copper | ASUS ROG PG35VQ
Back to top
Frant
King's Bounty



Posts: 24679
Location: Your Mom
PostPosted: Fri, 9th Sep 2011 05:14    Post subject:
If you really want to run latest generation SSD in RAID0 you should get a separate RAID-card that use the PCIe-bus.


Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn!

"The sky was the color of a TV tuned to a dead station" - Neuromancer
Back to top
timechange01
VIP Member



Posts: 6650

PostPosted: Sat, 10th Sep 2011 07:51    Post subject:
So I received my 120GB Corsair Force 3 SSD in the mail today. A very sexy looking drive but just as I suspected it didn't perform ANYWHERE NEAR manufacturer specs. The manufacturer specs for the Corsair Force 3 state max sequential read should be 550MB/s and write speed should be 500MB/s.

Im using an EVGA X58 FTW3 board which has a Marvell 91xx SATA III controller....same controller as bxrdj has.

I ran AS SSD benchmark on the empty drive and guess what, max write speeds around 150MB/s and max read speed was at a shameful 200MB/s....This is WORSE than my 64GB Corsair Performance 3 SATA III (non sandforce based) SSD. I first benchmarked the drive completely empty, and then I removed all harddrivesd and installed only windows 7 on the SSD and still 200MB/s read speed with both AHCI and IDE modes.

So there you have it, SandForce 2200 + Marvell 91xx = shit performance. Only get SandForce based drive if youve got sandybridge or AMD setup. Im calling Newegg tomorrow morning and returning the drive for a larger capacity Performance 3 drive.



ASUS Maximus XII Formula | Core i9 10900k @ 5.2Ghz | 32GB G.Skill DDR4 4200Mhz | EVGA RTX 3080 Ti FTW3 Ultra Hydro Copper | ASUS ROG PG35VQ
Back to top
Przepraszam
VIP Member



Posts: 14548
Location: Poland. New York.
PostPosted: Sat, 10th Sep 2011 09:01    Post subject:
wow, newegg ships ssd using DHL?...

wow what a fail


Back to top
Frant
King's Bounty



Posts: 24679
Location: Your Mom
PostPosted: Sat, 10th Sep 2011 18:44    Post subject:
timechange01 wrote:
So I received my 120GB Corsair Force 3 SSD in the mail today. A very sexy looking drive but just as I suspected it didn't perform ANYWHERE NEAR manufacturer specs. The manufacturer specs for the Corsair Force 3 state max sequential read should be 550MB/s and write speed should be 500MB/s.

Im using an EVGA X58 FTW3 board which has a Marvell 91xx SATA III controller....same controller as bxrdj has.

I ran AS SSD benchmark on the empty drive and guess what, max write speeds around 150MB/s and max read speed was at a shameful 200MB/s....This is WORSE than my 64GB Corsair Performance 3 SATA III (non sandforce based) SSD. I first benchmarked the drive completely empty, and then I removed all harddrivesd and installed only windows 7 on the SSD and still 200MB/s read speed with both AHCI and IDE modes.

So there you have it, SandForce 2200 + Marvell 91xx = shit performance. Only get SandForce based drive if youve got sandybridge or AMD setup. Im calling Newegg tomorrow morning and returning the drive for a larger capacity Performance 3 drive.


But.. most motherboards have multiple controllers, first the built-in controller (Intel or AMD, mostly 2-4 ports), then they have a Marvell controller for 4-6 more ports.

EDIT: Oh, you have a board based on X58 with the übershitty Marvell-controller. No native SATA3.

You COULD get a separate SATA-controller, but I don't know what's out there.


Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn!

"The sky was the color of a TV tuned to a dead station" - Neuromancer
Back to top
bxrdj




Posts: 1469
Location: Far from Home
PostPosted: Thu, 15th Sep 2011 01:32    Post subject:
hey timechange which benchamarking tool do you use? I want to test out my vertex Smile


fuck ...
Back to top
timechange01
VIP Member



Posts: 6650

PostPosted: Thu, 15th Sep 2011 01:37    Post subject:
bxrdj wrote:
hey timechange which benchamarking tool do you use? I want to test out my vertex Smile


AS SSD, CrystalDiskMark, and ATTO Disk

I mainly use AS SSD to get the sequential read and write speeds because it takes less than 20 seconds to benchmark the drive



ASUS Maximus XII Formula | Core i9 10900k @ 5.2Ghz | 32GB G.Skill DDR4 4200Mhz | EVGA RTX 3080 Ti FTW3 Ultra Hydro Copper | ASUS ROG PG35VQ
Back to top
Page 1 of 1 All times are GMT + 1 Hour
NFOHump.com Forum Index - Hardware Zone
Signature/Avatar nuking: none (can be changed in your profile)  


Display posts from previous:   

Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB 2.0.8 © 2001, 2002 phpBB Group