Forum Vs Naplam - Fasted real world storage solution

Quote:

---

Originally Posted by SteveRo

Yes, I have a problem, I know - should seek counseling.

---

I've always wondered why nobody here has got one yet. People have storage setups worth around $2k, but nobody gets it... No need for any RAID overheads, the CPU usage isn't as high and it does 100k IOPs with the lowest access times. I am kind of stupid myself as well as I will have 4 X25-es any day now lol.

Quote:

---

Originally Posted by One_Hertz

I've always wondered why nobody here has got one yet. People have storage setups worth around $2k, but nobody gets it... No need for any RAID overheads, the CPU usage isn't as high and it does 100k IOPs with the lowest access times. I am kind of stupid myself as well as I will have 4 X25-es any day now lol.

---

It doesn't seem like so much when you spread the cost over several months, an SSD here, an SSD there. For me its much harder to pull the trigger on a single $2K item vs lower cost piece by piece over time. It's a nonlogical-psycho problem.

Quote:

---

Originally Posted by SteveRo

It doesn't seem like so much when you spread the cost over several months, an SSD here, an SSD there. For me its much harder to pull the trigger on a single $2K item vs lower cost piece by piece over time. It's a nonlogical-psycho problem.

---

http://smiliesftw.com/x/werd.gif

After a while human brain will justify anything... That's the beauty of it! :D:D

Quote:

---

Originally Posted by NeedMoMegaHurtZ

@F.E.A.R - why the high 43% CPU usage ? What CPU configuration? Is it the controller, ACARD or the something else?

SteveRO you don't seem to have the same CPU% usage but a similar setup as F.E.A.R ?

---

I don´t know why.
I think the system is my "problem"!?!?!? :shrug:
When i take a look in the taskmanager, i can´t find a thread with a high cpu-load.
So it must be the system. (ASUS Striker II Extreme nf790i)

SteveRo use a Core i7-sys with the better bandwidth - i use a QX9650


btw: I´m using the highest stripesize - 128k atm.

Quote:

---

Originally Posted by One_Hertz

For about the same price the original SLC IOdrive would give 80GB of space with roughly the same performance, if not better. Not bootable though.

---

Quote:

---

Originally Posted by One_Hertz

I've always wondered why nobody here has got one yet. People have storage setups worth around $2k, but nobody gets it... No need for any RAID overheads, the CPU usage isn't as high and it does 100k IOPs with the lowest access times. I am kind of stupid myself as well as I will have 4 X25-es any day now lol.

---

ive spent $4k on controllers alone over the last 2-3 years.. and instead of dealing with diff controllers/x # of drives.. id rather have 1x iodrive but its not bootable.. ahhh!

just what is it gonna take for them to make it bootable..

Quote:

---

Originally Posted by Mic

Hey Napalm, any chance you could run the same test with the original 512mb memory on your controller if you still got it. Curious how much of a difference, if any cache, memory gives for read access like boot times and app. load.

---

all vidz were done @ 512mb cache

Quote:

---

Originally Posted by F.E.A.R.

@ Napalm

This is my next project (in my dreams *lol*)

---

theres no sata3 drives/sata3 controllers :shrug:

where are the sas2 ssds ??

as for firmware 1.47 vs 1.46.. yeh on xp 1.47 is slower but i got slower ssd firmware aswell.. im gonna flash back the ssd firmware and then ill see better how 1.47 vs 1.46 compares

im sick and tired of these jmicrons.. but im still waiting on x25e g2 :shrug: how much longer..

Quote:

---

Originally Posted by NapalmV5

just what is it gonna take for them to make it bootable..

---

my guess is never.

Hell I wouldn't even mind throwing Windows on an X25-M and the rest of an ioxtreme. But I don't think they are even out yet.

why never.. theyd sell more if bootable.. no ?

or they want to keep it for the elites only.. is that the point ?

Quote:

---

Originally Posted by NapalmV5

why never.. theyd sell more if bootable.. no ?

or they want to keep it for the elites only.. is that the point ?

---

I think it has to do the w/ the way the writes are cached to the ram. Its similar to why MFT can't be booted from. This is what I've gathered anyways. They have said since its release (damn near two years ago?) that it would be bootable soon. Which is why I think it will never come.

Quote:

---

Originally Posted by NapalmV5

ive spent $4k on controllers alone over the last 2-3 years.. and instead of dealing with diff controllers/x # of drives.. id rather have 1x iodrive but its not bootable.. ahhh!

just what is it gonna take for them to make it bootable..

---

;) http://www.dvnation.com/Fusion-IO-IO...Review-p5.html

Quote:

---

The other disadvantage is that the ioDrive is not yet bootable. This feature is coming though in the form of a simple firmware update in Q4 2008.

---

Quote:

---

Originally Posted by NapalmV5

where are the sas2 ssds ??

---

Voilà ;)

SAS 2 or Fibre Channel up to 350MB/s. and 800GB - what you want.
But i hope you moneybag is big enough http://www.abload.de/img/uglylaughk53.gif

30X 15K HDDs vs. 1x ZeusIOPS http://www.stec-inc.com/ssd/videos/ssdvideo2.php so many IOps :eek:

So, if you were building a high-end system today, what would it look like in terms of number and type of drives, controller with cache, I/O bus, etc? Would you stay with the 1231ML and PCIe 1.1, or move to something like the LSI 9260-8i on PCIe 2.0?

Also, given a fixed budget for the project, would you use any X25-E drives (if so, where/for what?), or just stripe more X25-Ms together?

Quote:

---

Originally Posted by F.E.A.R.

;) http://www.dvnation.com/Fusion-IO-IO...Review-p5.html

---

Exactly. It was suppose to be bootable LAST YEAR

Quote:

---

Originally Posted by F.E.A.R.

Voilà ;)

SAS 2 or Fibre Channel up to 350MB/s. and 800GB - what you want.
But i hope you moneybag is big enough http://www.abload.de/img/uglylaughk53.gif

30X 15K HDDs vs. 1x ZeusIOPS http://www.stec-inc.com/ssd/videos/ssdvideo2.php

---

Not out till Q4 and doesn't even list a price.

I don´t know the price of ZEUS IOPS, but i know STEC is very expensive :(

Can anyone ask for the price of Zeus IOps?

(949) 260-8345 - California

Quote:

---

Originally Posted by F.E.A.R.

I don´t know the price of ZEUS IOPS, but i know STEC is very expensive :(

Can anyone ask for the price of Zeus IOps?

(949) 260-8345 - California

---

The people who buy STEC products does not ask for price, they just buy it :yepp:

Quote:

---

Originally Posted by F.E.A.R.

;) http://www.dvnation.com/Fusion-IO-IO...Review-p5.html





Voilà ;)

SAS 2 or Fibre Channel up to 350MB/s. and 800GB - what you want.
But i hope you moneybag is big enough http://www.abload.de/img/uglylaughk53.gif

30X 15K HDDs vs. 1x ZeusIOPS http://www.stec-inc.com/ssd/videos/ssdvideo2.php so many IOps :eek:

---

lol sas2 ssd must cost double that of their sas ssd

2x/4x of that @ areca 1800 of baby!! the cost of the controller is pocket change..

Quote:

---

Originally Posted by AceNZ

So, if you were building a high-end system today, what would it look like in terms of number and type of drives, controller with cache, I/O bus, etc? Would you stay with the 1231ML and PCIe 1.1, or move to something like the LSI 9260-8i on PCIe 2.0?

Also, given a fixed budget for the project, would you use any X25-E drives (if so, where/for what?), or just stripe more X25-Ms together?

---

1231/4x x25e

still no controller like it

Quote:

---

Originally Posted by NapalmV5

1231/4x x25e

still no controller like it

---

Can you say more? What differentiates the 1231? Why would you prefer something like two X25-E drives instead of eight X25-M (based on their relative price)? How much controller memory?

I'm not disagreeing, and I realize that benchmarks don't tell the whole story, but Tom's Hardware claims they can get 1.7 GB/s throughput from the LSI 9260-8i, using 8 drives in a PCIe 2.0 x8 slot in RAID 0 (212 MB/s per drive) -- that's well above Areca's published results of 825 MB/s for 5 drives (165 MB/s per drive).

Is it the small block random I/O throughput that's the differentiating issue?

Would your choice be different if you were using RAID 5 or 6?

I think, i have the price for one STEC Zeus IOps - US-$7500 for 30GB ! ! ! :(
You can find it in the video...

wrong thread FEAR?
Post it again in Pliant thread :D

Quote:

---

Originally Posted by AceNZ

Can you say more? What differentiates the 1231? Why would you prefer something like two X25-E drives instead of eight X25-M (based on their relative price)? How much controller memory?

I'm not disagreeing, and I realize that benchmarks don't tell the whole story, but Tom's Hardware claims they can get 1.7 GB/s throughput from the LSI 9260-8i, using 8 drives in a PCIe 2.0 x8 slot in RAID 0 (212 MB/s per drive) -- that's well above Areca's published results of 825 MB/s for 5 drives (165 MB/s per drive).

Is it the small block random I/O throughput that's the differentiating issue?

Would your choice be different if you were using RAID 5 or 6?

---

lol say more??

the best ssd support for the last ~2 years.. low access/latency/cpu usage.. native sataII.. high performance small block

why go 8x mlc when 4x slc decimates the world ??

get 8x x25m + 4x x25e and see for yourself

the fewer the ssds the better the controller handles/controls.. the lower the cpu load

since 4x is enough to saturate the controller 4x is the max i go on the 1231

512mb cache is enough for most tasks.. but youd have to try 256/512/1024/2048 to see which fits better your workload/apps

some lunatics say the highest bandwidth/throughput the highest performance

some lunatics say the highest iops the highest performance

me?.. iops/bandwidth/access/latency/cpu usage/ssd support/everything about the controller is equally important

thats why 1231 fits my needs/wants perfectly

thats why ive said time and time again.. 1231 is king of controllers

:toast:

@ XS Janus

No ;)
See post # 312 (answer to Napalm)


@ Napalm

You mean 1231/1261/1280ML :D

indeed :D

:cat:

meow?

http://t1.gstatic.com/images?q=tbn:Q...ty_800x600.jpg *curr*

:rofl:

haha
:D:D

I tried your little 40app test with my Xeon rig, was something like 7-8s and with a 4gb RAMdrive with readyboost was only like 3-4s to load the 40apps

@ Napalm

2TB Supertalent Raid-Card with 1,4/1,2GB/s. r/w and Intel IOP348. Fast enough? ;) http://translate.google.com/translat...ie%2F447831%2F

Quote:

---

Originally Posted by NapalmV5

why go 8x mlc when 4x slc decimates the world ??

---

Given the name of this forum, do you really not know the answer to that question?

Frankly, I'm much more interested in building the fastest possible system. If 4x SLC decimates the world, and 8x MLC is even faster, then that's what I want.

Quote:

---

Originally Posted by NapalmV5

get 8x x25m + 4x x25e and see for yourself

---

If I had the time and money to buy and test one of every possible system, I would. Unfortunately, I have to rely on pre-purchase analysis and feedback from current owners and users instead.

Quote:

---

Originally Posted by NapalmV5

since 4x is enough to saturate the controller 4x is the max i go on the 1231

---

It's exactly that point which makes me think that a PCIe 2.0 controller has the potential of being faster overall, since they have twice the bus bandwidth available.

Quote:

---

Originally Posted by NapalmV5

the fewer the ssds the better the controller handles/controls.. the lower the cpu load

---

Really? Do you have any tests to back that up, or is it more of a gut feeling? I suppose zero SSDs would be the lowest CPU load, but it definitely wouldn't be the fastest.

Quote:

---

Originally Posted by F.E.A.R.

@ Napalm

2TB Supertalent Raid-Card with 1,4/1,2GB/s. r/w and Intel IOP348. Fast enough? ;) http://translate.google.com/translat...ie%2F447831%2F

---

id be interested in the 192gb slc but iop348 ?

thats sour kraut on ice cream :shakes:

goddamn sas controllers!

so how they do 1.4/1.2gb/s r/w ? custom iop348 ?

whats the interface between the ssd and the iop ?

Quote:

---

Originally Posted by AceNZ

Given the name of this forum, do you really not know the answer to that question?

Frankly, I'm much more interested in building the fastest possible system. If 4x SLC decimates the world, and 8x MLC is even faster, then that's what I want.



If I had the time and money to buy and test one of every possible system, I would. Unfortunately, I have to rely on pre-purchase analysis and feedback from current owners and users instead.



It's exactly that point which makes me think that a PCIe 2.0 controller has the potential of being faster overall, since they have twice the bus bandwidth available.



Really? Do you have any tests to back that up, or is it more of a gut feeling? I suppose zero SSDs would be the lowest CPU load, but it definitely wouldn't be the fastest.

---

look you think you know better than me?.. no problem.. but dont be surprised when and why my 1231/4x x25e decimates your 8x mlc/pcie2 sas

Quote:

---

Originally Posted by NapalmV5

look you think you know better than me?.. no problem.. but dont be surprised when and why my 1231/4x x25e decimates your 8x mlc/pcie2 sas

---

I wasn't trying to assert that the 9260 with 8x MLC would be faster than the 1231 with 4x SLC. I've seen the benchmarks in the other thread. In my environment, small block random I/O rates determine performance much more than sequential I/O. The lower latency on the 1231 makes complete sense.

The original question I was trying to ask is something completely different: how would you build the fastest possible system? First choose the fastest drives and controller. If the 1231 is the best, and if more drives or a PCIe 2.0-based controller won't help, then what's next? Stripe a bunch of them together? Are there configuration tweaks, such as NTFS cluster size, strip/stripe size, disabling cache, or what?

Or maybe switching to a native PCIe-only controller would be better? That should eliminate the SATA latency, which would help a lot with small block random I/O.

Quote:

---

Originally Posted by NapalmV5

id be interested in the 192gb slc but iop348 ?

thats sour kraut on ice cream :shakes:

goddamn sas controllers!

so how they do 1.4/1.2gb/s r/w ? custom iop348 ?

whats the interface between the ssd and the iop ?

---

I think 1,4GB/s. with cache.

if thats the case.. 1231 does 1.6+gb/s :rofl:

sas controller just cant beat 1231/1261 no way no how :ROTF:

Quote:

---

Originally Posted by AceNZ

I wasn't trying to assert that the 9260 with 8x MLC would be faster than the 1231 with 4x SLC. I've seen the benchmarks in the other thread. In my environment, small block random I/O rates determine performance much more than sequential I/O. The lower latency on the 1231 makes complete sense.

The original question I was trying to ask is something completely different: how would you build the fastest possible system? First choose the fastest drives and controller. If the 1231 is the best, and if more drives or a PCIe 2.0-based controller won't help, then what's next? Stripe a bunch of them together? Are there configuration tweaks, such as NTFS cluster size, strip/stripe size, disabling cache, or what?

Or maybe switching to a native PCIe-only controller would be better? That should eliminate the SATA latency, which would help a lot with small block random I/O.

---

whats next is sata3 pcie2 controllers/sata3 drives.. none out and about yet sata6g+sas6g controller = sata3g performance <- you wouldnt want that..

the other option is sas6g drives + sas6g controller: areca 1800 though 4x/8x sas6g drives cost more than a highend mercedes benz

so at the moment if you want the fastest possible system without the hassle of raid cards/x # of drives theres pcie iodrive/ramsan

ramsan: http://www.xtremesystems.org/forums/...d.php?t=230107

I think we can all agree that the card to wait for is a new gen. one but without SAS overhead...
The question is who will be the first and how much cheaper will it be in regard to 1231...
:D

Quote:

---

Originally Posted by NapalmV5

if thats the case.. 1231 does 1.6+gb/s :rofl:

sas controller just cant beat 1231/1261 no way no how :ROTF:

---

I give you competition with my Areca 1680ix ;)

1680ix rules with SAS drives. I scales all the way up to 1250MB/s with 10x 15k.5 sas drives. And I got over 1gb/s read speed with 5x Vertex.

I tested the latency. You remember. My latency was not so bad :)

*sigh* i thought we went over this nizzen.. no ?

you dont got 1gb/s read.. 5x 150 = 750mb/s at best

if you wanna be like that.. i do over 1gb/s too.. @ h2benchw and if you prefer everest.. 1.6gb/s

5x vertex? i thought its my 4x slc vs your 6x mlc?

let me make it easier..

nizzen vs napalm
1680 vs 1231
2gb cache vs 2gb cache
5x/6x mlc (vertex) vs 4x slc (jmicron)
942mb/s vs 1070mb/s @ h2benchw
1164mb/s vs 1608mb/s @ everest

5x vs 4x
http://i30.tinypic.com/6z1n4i.jpg vs http://i31.tinypic.com/64gzf4.png

6x vs 4x
http://i29.tinypic.com/103sm09.png vs http://i27.tinypic.com/977fr4.jpg


and thats my 100mhz pcie vs your xxxmhz pcie ??

I beat you @ average read still in everest :p

u should u got ~6x drives but not 1gb/s.. my avg is higher too due to cache

what pcie ? 125mhz ? higher ?

your h2benchw was @ 5x not 6x and everest @ 6x

nizzen: 5x * 150 ~ 750mb/s hence the 711mb/s @ h2benchw
napalm: 4x * 188 ~ 750mb/s hence the 716mb/s @ h2benchw

so whether you like it or not your 6x * 150 ~ 900mb/s at best

LSI + 7xintel

http://i413.photobucket.com/albums/p...lsi7xintel.jpg

:p:

note: block size: 512K

lol show me some real apps on that ^ and then tell me you still beat me

what is up with 194mb/s minimum ?? :rofl:

Quote:

---

Originally Posted by NapalmV5

lol show me some real apps on that ^ and then tell me you still beat me

what is up with 194mb/s minimum ?? :rofl:

---

But do you beat that :p

Did you tested Pcmark Vantage? With youre controller you should beat the WR easy ;)

yeh all i care about is highest bandwidth.. i can get 9260 too if id want to show off high bandwidth numbers.. you aint gonna see 9260 numbers from me

i dont do vista/win7/vantage.. if i wanted to break WRs i wouldve submitted results..

im more into real apps.. since thats what i do 99% of the time :)

Quote:

---

Originally Posted by NapalmV5

yeh all i care about is highest bandwidth.. i can get 9260 too if id want to show off high bandwidth numbers.. you aint gonna see 9260 numbers from me

i dont do vista/win7/vantage.. if i wanted to break WRs i wouldve submitted results..

im more into real apps.. since thats what i do 99% of the time :)

---

Quote:

---

if i wanted to break WRs i wouldve submitted results..

---

:lol2:

How does the Areca 1231ML do with parity RAID (5 or 6)? Still a top performer, or are the 1680 cards better?

i dont do raid 5/6.. only do raid 0.. but why would raid 5/6 performance be diff than raid 0 performance ??

if id do raid 5/6 id still do it on the 1231 as long as the drives are sata



btw: i just noticed rhys been banned.. what happened ??

Quote:

---

Originally Posted by NapalmV5

i dont do raid 5/6.. only do raid 0.. but why would raid 5/6 performance be diff than raid 0 performance ??

---

Small block writes are slower, since there are two reads and two writes for each block (data and parity both have to be read and written).

Also, the parity calculations themselves take time and can introduce latency.

exactly and what better/faster controller @ that for sata2 drives ??

for sata2 drives theres still nothing better/faster than 1231/1261/1280.. be it raid 0/5/6

or wait till sata3 controllers are out if you go sata3 drives @ raid 5/6 or 0

the lsi guys claim/think/believe they are kings

the ich guys claim/think/believe they are kings

the acard guys claim/think/believe they are kings

the ramdisk guys claim/think/believe the are kings

ignorance ^ truly is bliss lol


i never claimed such.. never claimed awesomeness.. ive only claimed the 1231 to be the king of controllers and to this day still is

what/who i am and what i do is simple.. i napalm king arses with a 1231 a 512mb cache and 4x jmicron freakin jmicron!! not really my intention and you guys brought it upon yourselves

you havent seen anything yet from napalm ;) im just getting warmed up! :toast:

These guys really take this stuff personally, don't they Napalm? I don't see why everyone is attacking you, you're just throwin the numbers up and having fun. Seems like the other guys are getting pissed.. Oh well-

LOL!
The war is on!
hahahaha :D:D

Holy S:banana::banana:T Balls man, just came across this thread, those are some insane numbers, may I ask about how much you have total invested in the cards and such?

Quote:

---

Originally Posted by Griff805

These guys really take this stuff personally, don't they Napalm? I don't see why everyone is attacking you, you're just throwin the numbers up and having fun. Seems like the other guys are getting pissed.. Oh well-

---

i only throw up numbers the system throws up at me.. and yes im trying to have fun while at it (mft results).. but it gets personal with everyone

yeh theyre getting pissed! cause even their sandforce ssd cant keep up :rofl: :D :D

btw why/how come youre being nice to me griff ?? :)

Quote:

---

Originally Posted by XS Janus

LOL!
The war is on!
hahahaha :D:D

---

:cheer2:

you embedded/posted vid @ pcekspert ? :D lol

Quote:

---

Originally Posted by GameGuru

Holy S:banana::banana:T Balls man, just came across this thread, those are some insane numbers, may I ask about how much you have total invested in the cards and such?

---

well 1231:800$ 4x slc jmicron:1000$ exactly a year ago

Quote:

---

Originally Posted by NapalmV5

i only throw up numbers the system throws up at me.. and yes im trying to have fun while at it (mft results).. but it gets personal with everyone

yeh theyre getting pissed! cause even their sandforce ssd cant keep up :rofl: :D :D

btw why/how come youre being nice to me griff ?? :)

---

I don't really remember being a jerk to you, other than when I thought maybe your pride was kicking in a bit too heavy ;)

I think you may have interpreted one of my posts a ways as an attack rather than just a sincere inquiry.

On PCexpert? :)
I haven't embedded it there. Do you have a link?
I have almost forgotten about that page. :D

Quote:

---

Originally Posted by Griff805

I don't really remember being a jerk to you, other than when I thought maybe your pride was kicking in a bit too heavy ;)

I think you may have interpreted one of my posts a ways as an attack rather than just a sincere inquiry.

---

pride is the root of all evil.. prides been my enemy and will be till the end.. just because im joking/kidding around dont mean my pride is kicking in at all.. though i get a kick out of napalming those full of pride and having a bit of fun at it.. i do know that no matter how fast i get someone else always is/will be faster.. and i look at every aspect of my life that way

truly sincere inquiry? my response is just as sincere.. alright griff all cool :toast:

Quote:

---

Originally Posted by XS Janus

On PCexpert? :)
I haven't embedded it there. Do you have a link?
I have almost forgotten about that page. :D

---

lol i noticed that @ youtube insight (discovery) theres no link to the embedded vid just the forum http://forum.pcekspert.com/

Well, two months have passed. Is there anything new out there yet that improves on the 1231ML + X25-E combo?

Seems like some of the recent ICH10R benches look better than the 1231ML. What's the deal with that?

there is many different platforms that can outperform that combo, or be very very close, at much better pricepoints as well :)

Quote:

---

Originally Posted by Computurd

there is many different platforms that can outperform that combo, or be very very close, at much better pricepoints as well :)

---

Such as?

OK so my take on this.
Assuming we are talking all around general workstation and Raid 0 performance (as measured by pcmark vantage for example) -
1. If Areca 1231ML-4GB and 4 Intel x25e is 100% performance then -
2. LSI 9260-8i and 4 vertex 30GB (or vertex turbo) is 95%+
3. LSI 9211-8i and 4 vertex is 90-95%+
Substituting Intel X25m's for the vertex might be a little faster - but also more expensive (30 going to 80GB).
The Kingston 40GB might also be an option to consider.
The LSI cards are also more future proof (pcie 2.0).

Now lets compare cost - how much do you want to spend for that last 5% of performance? -
1. (from above) 1K+1.5K=2.5K$
2. 490+520=1010$
3. 240+520=760$
You will pay a heavy premium for less than 5-10% increase - my opinion.

SteveRo,

I would wait for the new Vertex 2 (SandForce) or Micron C300 SSD's, they're just around the corner.

Option 3 sounds sweet to me, especially since the 9211-4 is so cheap (4x C300 is enough for me speed-wise :D )
Anybody know the bootup init time for those LSIs? If below 15s I might just ditch the ICH10R ;)

Napalm has JMicron SLC drives, not X25-Es.

Quote:

---

Originally Posted by SteveRo

OK so my take on this.
Assuming we are talking all around general workstation and Raid 0 performance (as measured by pcmark vantage for example) -
1. If Areca 1231ML-4GB and 4 Intel x25e is 100% performance then -
2. LSI 9260-8i and 4 vertex 30GB (or vertex turbo) is 95%+
3. LSI 9211-8i and 4 vertex is 90-95%+
Substituting Intel X25m's for the vertex might be a little faster - but also more expensive (30 going to 80GB).
The Kingston 40GB might also be an option to consider.
The LSI cards are also more future proof (pcie 2.0).

Now lets compare cost - how much do you want to spend for that last 5% of performance? -
1. (from above) 1K+1.5K=2.5K$
2. 490+520=1010$
3. 240+520=760$
You will pay a heavy premium for less than 5-10% increase - my opinion.

---

Interesting. Are you talking about large block / sequential access or random I/O, or both?

How does the X25-V compare to the Vertex?

I like to look at configs from a constant-dollar perspective. Let's say I have $5K to spend on an SSD disk subsystem. What's the fastest setup (maybe one config that's bootable and one that isn't)?

I know 4 drives saturate the 1231ML. Is the same thing true for the LSI cards? Or can they handle more?

Quote:

---

Originally Posted by One_Hertz

Napalm has JMicron SLC drives, not X25-Es.

---

True. But he also said he thought the X25-E would be faster than what he has.

there isnt a aaray out there that has saturated the 9260 that i have seen. the only way would be via a 6gb/s expander. there isnt even one on the market yet (coming very soon)
you are speaking of a controller that has 2.8 gb/s sequential speed potential and 1.8 write :)

Quote:

---

Originally Posted by Computurd

there isnt a aaray out there that has saturated the 9260 that i have seen. the only way would be via a 6gb/s expander. there isnt even one on the market yet (coming very soon)

---

Are you thinking about something like the HP SAS expander?

http://h10057.www1.hp.com/ecomcat/hp...468406-B21.htm

The documentation on it is poor, but it is a 6gbps expander that's based on the PMC-Sierra chip.

Quote:

---

Originally Posted by Computurd

you are speaking of a controller that has 2.8 gb/s sequential speed potential and 1.8 write :)

---

What about small block IOPS?

AceNZ,

The question is, how many iops do you need or are you just looking for bandwidth or both?
I find the 9260 to be a good allround performer, for iops it's not in the same league as the Areca 1231ML , Areca 1680, LSI 9211 series though.
In fact ICH is a better iops performer than all of these at qd >8 using small blocks but as you know limited on bandwidth.

The X25-V is an excellent iops performer but writes are limited to ~45MB/s.

This chart shows 4K random iops @ QD 64 so it doesn't tell the full story. (most drives in this chart are HDDerased, so it would be its peak performance)
http://www.diskusjon.no/index.php?ac...74:iometer.PNG

Visit the Measuring QD in win7 thread Link where I have compiled a few charts showing 4K random iops.

you have to be very careful of asessing the performance of any drive or storage subsystem if you aree simply using 4k as the yardstick. 4k speeds do not have NEAR the impact on performance that it is made out to be. you can have a device with low 4k random (9260) literally smoke a device with TRIPLE the amount of 4k (9211) performance, yet the slower 4k performer wins with regular tasks. and yes, cache off.

Computurd,

Thats why I said I find the 9260 to be a good allround performer.
I'm very much satisfied by the 9260-8i :) and yet it doesn't do well in most iometer tests at lower block sizes.

4K random is just a popular test size and one should of course look at other benchmarks and test sizes before making a decision.

heres the last vid on 4x jmicron slc ssd.. am finally getting faster ssds :)

as promised: light loadup: 300 x firefox @ default home page "light" :D loadup

unlike all the other vids q9650 @ 4ghz only ;)

http://www.youtube.com/watch?v=sHJ8i2D-NmQ

after youtube vid processing the vid dont look that great.. sorry about the quality

Good morning Mr Napalm, nice to hear from you and nice video!
What are you thinking of replacing your SSDs with?

i redid/posted new better quality vid^

good morning mr steve :) thanks! :toast:

just faster slc.. nothing extraordinary.. not 100% on anything atm but ill be getting smtg this week

Hi Napalm, looking forward to seeing the 4x jmicron replacement. I'm a little surprised you are buying now when the new gen SATA 6 drives will be out before the end of the year.

I take it the 1231 is staying for a bit longer?

Regarding your batch test. I tried it on my single ssd. Task manager tells me that 300 processes start up but only 29 instance of IE actually open. :shrug:

I tried opening 300 copies of Notepad and that worked OK. It took 22 seconds. It used around 3GB of RAM and the CPU seemed to only max out at 65% (I guess because the SSD could not cope).

After opening around 100 copies the pc seemed to lock up for a few seconds before the rest opened. The QD seemed to max out around 60, so it would seem that the lock up occurred when the QD got too much to handle.

For anyone interested this is the code for the batch file to open one instance of IE. (Just repeat on separate lines to add more instances).

Start "" "C:\Program Files\Internet Explorer\iexplore.exe"

To open Notepad:
Start "" "C:\Windows\NOTEPAD.EXE

hi audienceofone.. sata6 drives/controllers dont budge me at all.. 1231 staying a megabit longer :)

try 100x/150x/200x ie probably certain apps are limited to a certain # of instances

notepad.. light loads i dont know how to do light loads.. wasnt gonna post light load vid/s.. but alright

ill post a vid of 100x/200x/300x firefox.. ill see how many notepad/iexplorer/calculator

100x ff = 5 sec
200x ff = 10 sec
300x ff = 15 sec

if you look closely at the vid youll see the 300x ff loading in 15 sec the rest of the loadup is loading the home page


lol @ onehertz - what are you going to be getting then ?? :)

i just uploaded one more vid.. still processing..

i tried ie6 @ 51x max instances and since im not using it didnt bother to update it

instead i did 100x firefox/notepad/calculator.. keep in mind there are two loadings going on since group similar is enabled

was gonna do 200x/300x/400x ff.. never loaded more than 400 so went as far as ff goes 498x firefox

here you have it 498x ff in 27 sec loadup

http://www.youtube.com/watch?v=9vyKUc7CVO4

no more 4x jmicron! onto faster ssd :D

28-32sec for 498 ff for me. Seems to be a bit different every time I run it.

I will probably change to a dual port acard for the more important stuff and an x25-m for less important. All this raid business does nothing for me.

Quote:

---

Originally Posted by audienceofone

Hi Napalm, looking forward to seeing the 4x jmicron replacement. I'm a little surprised you are buying now when the new gen SATA 6 drives will be out before the end of the year.

I take it the 1231 is staying for a bit longer?

Regarding your batch test. I tried it on my single ssd. Task manager tells me that 300 processes start up but only 29 instance of IE actually open. :shrug:

I tried opening 300 copies of Notepad and that worked OK. It took 22 seconds. It used around 3GB of RAM and the CPU seemed to only max out at 65% (I guess because the SSD could not cope).

After opening around 100 copies the pc seemed to lock up for a few seconds before the rest opened. The QD seemed to max out around 60, so it would seem that the lock up occurred when the QD got too much to handle.

For anyone interested this is the code for the batch file to open one instance of IE. (Just repeat on separate lines to add more instances).

Start "" "C:\Program Files\Internet Explorer\iexplore.exe"

To open Notepad:
Start "" "C:\Windows\NOTEPAD.EXE

---

I just tried this on my laptop - bad idea :rofl:
Even though it has a 120GB Ultradrive, its CPU and Ram weren't really up to the task :rolleyes: :p:

Hey napalm, you figure a few of those new Sandforce SSDs could beat ur setup? They should, considering the insane 4k random numbers...
I'm getting two of those tomorrow, will do your FF loadup test then :)

I must state i have issues with a script simply opening the same program 100x times or more, it could simply be cached. If you want more of a challenge, try opening 50-100 different programs and timing it.
I did that on my 1,5 year old rigg with 2R0 Mtron Pro 7025 32GB off AMD SB650, Phenom x4 @ 2,8Ghz, and 8GB DDR2 800mhz. 60 apps (some heavy) loading in ca 40 sec. Prefetch and superfetch off, and recorded it in 1920x1200 (wich added 10 seconds, it takes ca 30sec when not recording).
http://www.youtube.com/watch?v=tCCPS0Hoe28

Anvil had to 1-up me, so he did a simelar thing: http://www.youtube.com/watch?v=sax5wk300u4
And then again: http://www.youtube.com/watch?v=NFY_JEEDlK4

If you'd like to demonstrate the superiority of your storage solution, I have an IOmeter config i'd like you to run. It's basically mapping small block IO, both sequential and random, read and write, and mixes of them.

Quote:

---

Originally Posted by One_Hertz

28-32sec for 498 ff for me. Seems to be a bit different every time I run it.

I will probably change to a dual port acard for the more important stuff and an x25-m for less important. All this raid business does nothing for me.

---

lol try it @ 4ghz/120 dpi/extra large fonts

i can loadup 498x ff in 20 sec @ normal settings/4.5-6ghz

but then what would be the point in getting faster ssd.. let me upgrade! lol :D

Quote:

---

Originally Posted by jcool

I just tried this on my laptop - bad idea :rofl:
Even though it has a 120GB Ultradrive, its CPU and Ram weren't really up to the task :rolleyes: :p:

Hey napalm, you figure a few of those new Sandforce SSDs could beat ur setup? They should, considering the insane 4k random numbers...
I'm getting two of those tomorrow, will do your FF loadup test then :)

---

sandforce faster huh ?? lol yeh show me tomorrow :D

Quote:

---

Originally Posted by GullLars

I must state i have issues with a script simply opening the same program 100x times or more, it could simply be cached. If you want more of a challenge, try opening 50-100 different programs and timing it.
I did that on my 1,5 year old rigg with 2R0 Mtron Pro 7025 32GB off AMD SB650, Phenom x4 @ 2,8Ghz, and 8GB DDR2 800mhz. 60 apps (some heavy) loading in ca 40 sec. Prefetch and superfetch off, and recorded it in 1920x1200 (wich added 10 seconds, it takes ca 30sec when not recording).
http://www.youtube.com/watch?v=tCCPS0Hoe28

Anvil had to 1-up me, so he did a simelar thing: http://www.youtube.com/watch?v=sax5wk300u4
And then again: http://www.youtube.com/watch?v=NFY_JEEDlK4

If you'd like to demonstrate the superiority of your storage solution, I have an IOmeter config i'd like you to run. It's basically mapping small block IO, both sequential and random, read and write, and mixes of them.

---

here we go again :shakes:

FYI: its not chached.. look @ ram peak @ task manager :shakes:

FYI: http://www.youtube.com/watch?v=MvBKB2NndbI

FYI: i had this setup since nov. 2008

FYI: iometer ?? :rofl:

Quote:

---

Originally Posted by GullLars

If you'd like to demonstrate the superiority of your storage solution

---

That's the thing, we are not. It is pretty much a full system test. Who cares how fast your storage is if your rig is slow.

True dat ;)

From what i've seen, the Areca controllers coupled with a few SSDs are lightning fast for real world usage.

Quote:

---

Originally Posted by One_Hertz

That's the thing, we are not. It is pretty much a full system test. Who cares how fast your storage is if your rig is slow.

---

thanks onehertz :toast:

Quote:

---

Originally Posted by GullLars

True dat ;)

From what i've seen, the Areca controllers coupled with a few SSDs are lightning fast for real world usage.

---

gulllars lol now youve seen :) i never claimed superiority.. though ill go up against such "awesomeness" mostly for fun :up:

300 x FF = 29 seconds
900 x FF = lock up after 80 seconds, then lost track whilst trying to avoid reboot. :ROTF:

If you insist on spamming one application, try making a "multi-thread" batch setup, make one batch launching 10 other, and then have those 10 launch N instances of the program each, like 30 for a total of 300. Should be interresting to see if the QD skyrockets.
I stilll suspect the ML1231 reads every FF instance afte the first 2-10 from cache, and thereby get cache speeds instead of disk speed, so you'd basically be benching the RAID controllers caching performance.

I made this one for fun just to show a few n00b friends how much difference even a simple SSD setup makes. Includes startup from BIOS, 10 app Batch, and then shutdown.

http://www.youtube.com/watch?v=7bndK-cbokU

I can do the 40 app batch that's posted earlier in the thread but kind of pointless since I'm just on a single Vertex. ;)

Quote:

---

Originally Posted by audienceofone

300 x FF = 29 seconds
900 x FF = lock up after 80 seconds, then lost track whilst trying to avoid reboot. :ROTF:

---

nice :up: lol @ 900 ff wow

since your on win7 i take it.. idk how many max ff you can launch without issues.. start @ 498 and add a bunch more after each launch and youll reach a limit.. let me know what that limit is on win7

Quote:

---

Originally Posted by GullLars

If you insist on spamming one application, try making a "multi-thread" batch setup, make one batch launching 10 other, and then have those 10 launch N instances of the program each, like 30 for a total of 300. Should be interresting to see if the QD skyrockets.
I stilll suspect the ML1231 reads every FF instance afte the first 2-10 from cache, and thereby get cache speeds instead of disk speed, so you'd basically be benching the RAID controllers caching performance.

---

:spill: :ROTF: :rofl: so all im doing is benching cache :wth: :clap:

that statement takes the cake!! :up:

Quote:

---

Originally Posted by FlawleZ

I made this one for fun just to show a few n00b friends how much difference even a simple SSD setup makes. Includes startup from BIOS, 10 app Batch, and then shutdown.

http://www.youtube.com/watch?v=7bndK-cbokU

I can do the 40 app batch that's posted earlier in the thread but kind of pointless since I'm just on a single Vertex. ;)

---

nice :toast: indeed.. ssd can make any system fly.. and you dont need large number of ssds in order to enjoy an xtremely fast system :up:

Quote:

---

Originally Posted by NapalmV5

heres the last vid on 4x jmicron slc ssd.. am finally getting faster ssds :)

as promised: light loadup: 300 x firefox @ default home page "light" :D loadup

unlike all the other vids q9650 @ 4ghz only ;)

http://www.youtube.com/watch?v=sHJ8i2D-NmQ

after youtube vid processing the vid dont look that great.. sorry about the quality

---

:shakes: :shakes: :shakes: I just tried this on a 5,400RPM drive :shakes:

Here's something bone-shattering for you, watch the beginning of this video. I could feel the non-existant platters breaking :( http://www.youtube.com/watch?v=MBvQuin7Y9w

Quote:

---

so all im doing is benching cache

that statement takes the cake!!

---

umm yea...you are.
doesnt take away from the fact that your system IS fast. other things do come into play here for uber loads, etc. nice job :)
however.
from a storage standpoint, yes that is all you are doing. do it with ich10r that fast and prove him wrong if that is your intention.

it is a nice trick though. very fast!

Or he could disable the RAID controllers cache, prefetch, and superfetch.

The results are impressive, and i don't question the speed, just the point of opening 100's of instances of the same program.
BTW, did anyone try to run a multi-threaded version? either a batch launching 5-10 other batches wich opens the programs, or just making 5-10 copies of a batch lauching the programs. And could someone compare a single batch to multiple batches with regards to queue depth? I would like to see if it matters :)

If you run a program more than once, the program is loading it from the cache.

This way you can get 800MB/s + win xp load in pcmark 05. Too bad the 220mb/s? limit is there for valid score or not.

If you are running a benchmark like crystal diskmark or hdtune more than once and the testfile is smaller than cache, the benchmark is loading from the cache. :yepp:

If it is not using the cache, why have cache? :rolleyes: