Will there be an xtremesytems discount?:clap:
Printable View
Will there be an xtremesytems discount?:clap:
Doesn't mean the device will work with raid adapters (or rather one over the others). If it does end up supporting major raid controllers this device can actually possibly find its way into enterprise use rather than being limited to "desktop enthusiasts" like ourselves.Quote:
Originally Posted by IanB
Come to think about it, using 4GB sticks I would be fine with a single dual sata port model. ~32GB of storage and 400MB/s
I wonder if there will be another revision of the device with a better dram controller than can give closer to the full SATAII bandwidth per port in the works.
Huh? :confused: Do you ask if a new hard drive supports a specific RAID controller? The SATA standard should be plug and play, as long as the controller and drive speak the same dialect there shouldn't ever be any incompatability. Controllers don't need tables of data to speak to every different hard drive make and model, else they'd need to be software-updated every time a new one came out. Same for drives talking to controllers. Do you update drive firmware every time you plug an old drive into a brand new mainboard? :confused:Quote:
Originally Posted by Levish
As long as they use a common subset of the available commands and don't do anything unexpected, there should be no problem linking any SATA device with any controller, short of a bug in the implementation somewhere. No explicit "support" is necessary on the drive side. The only issue I can see is that these things return data so fast to the controller that the firmware may not be set up to handle that. That requires support for SSDs on the controller side instead.
In theory you are right, in practice not all hard disks work well with all controllers and the other way around.Quote:
Originally Posted by IanB
Thats why some storage systems are qualified to run with some hard drives and so on.
No, I got the new Areca 1680ix-16 card, fairly high end, and it didn't support SSDs for boot until the third firmware revision. It handled them fine for a data drive, but wouldn't boot. "SATA standard" doesn't guarantee compatibility.Quote:
Originally Posted by IanB
I am very much looking forward to these. As I have 4 spare SATA connectors on my ARC1231ML these units would be Great for that spot.
4x 9010 with full RAM load would be a great test against my 8x MTRON's
Quote:
Originally Posted by IanB
Er, isn't that what I said? :shrug: Levish's suggestion was that these drives should support specific controllers, not the other way around. If a specific controller can't use these drives, then they are not at fault, the controller card is.Quote:
Originally Posted by Speederlander
Actually, no, that's not what you said. Both the controller AND the drive can be at fault when things don't work. There's more going on than just meeting the SATA standard. The WD 640GB harddrives that came out several months ago also had issues with several controllers. That turned out specifically to be a harddrive firmware issue from WD. But whatever, what people want to know is if the drives will work with the major controllers. Doesn't matter who didn't support who (controller or drive), doesn't work = doesn't work. That's why we need the tests and being a first time buyer is to be a beta tester in this case.Quote:
Originally Posted by IanB
What we are disagreeing on is the word "support". Having a bug in the firmware doesn't mean drives don't "support" some controllers. It means the drives had a bug, that will affect multiple controller hosts.
"Support" implies special firmware code or handling for specific hardware. In your Areca example that's clearly what they fixed to allow booting from SSDs. So you proved my point, that controllers need to support specific drives, not the other way around. :rolleyes:
Since the 9010 series has a SATA1 compatability mode, I think we can assume that ACARD are interested in making the thing as widely compatible as possible, otherwise they wouldn't be getting that many sales of what is already a niche product...
The WD 640s had issues. This ACARD may have issues. The new Arecas had issues with some drives, just like they may have with these. My point is, just because these things are sata means squat. While I certainly appreciate a good argument degenerating into semantics, in this case it's pointless. All that matters is that major controllers will play well with it. Problems could exist on either the controller or the ACARD side. That was the only point. Without results we don't know.Quote:
Originally Posted by IanB
You earlier said:
As though as long as ACARD is SATA it will be problem free. In reality that's simply not the case. "Shouldn't ever be any incompatibility" is a pretty absolute statement considering I have seen many incompatibilities on many different drives between many different controllers.Quote:
Originally Posted by IanB
so when are these going to hit retail? I found two on ebay. Gosh I hope they are cheaper than that. :(
No and those are the less expensive versions. I think the others are around $400:eek:Quote:
Originally Posted by Eternalightwith
http://www.acard.com/upload/20081030...%20150x150.gif
http://www.acard.com/english/fb01-batcar.jsp#Quote:
ANS-9010
5.25 inch SATA x 2-to-DDRII RAM Disk
(RAM module not including)
more information
Price:USD399.00 Pre-Order
The 9010B are $299
Hello guys. Here's my first post:
I have a 9010(8 slot) box with 16GB of ram in it, probably 32GB tomorrow. I'd be more than glad to run some benchmarks. Any requests on what programs to use or benchmarks to perform? I'm doing some tests with HD Tach and PCMark05 and I'll post results tomorrrow.
If you're going to nitpick aggressively about my phrasing, I prefer the obvious stress intended when I wrote it: "Shouldn't ever be any incompatibility". I believe that was the sense of my argument. If there IS incompatability with a particular drive or controller, it's clearly a bug, it's not a case of "support" or no "support".Quote:
Originally Posted by Speederlander
Heh, I felt you were nitpicking mine. ;) I don't care if it's due to bad support or no support or a bug or a calculated choice, either it works or it doesn't. There is a better than vanishing chance that it may have troubles, which ever side those troubles might come from. My original point was the sata standard it follows won't matter. Theoretically, sure, if both sides perfectly implement it would be peachy, but practical experience has proven time and time again that both controller manufacturers and hard drive manufacturers will find a way for something not to work, either by mistake or omission. Hence, I want to see the results of reviewers trying this thing on some controllers before I spend hundreds of dollars to beta test it.Quote:
Originally Posted by IanB
Those are good. Can you also post same results of the box running in RAID vs single connector?Quote:
Originally Posted by josh1980
cant remember if this has been asked: can u use diff capacity DIMMs?
Will the ANS-9010 work with 512x64 HD memory on 4gb sticks? My I-RAM works fine with the cheaper, high density modules. That would make it a lot nicer trying to populate this thing with enough memory to make it usefull.
So here's the long story on the troubles I've been going through.
The manual that comes with it says the following:
The ANS-9010/9010B supports both ECC and none ECC DR SDRAM DIMM. However, it is strongly recommended to use ECC DDR SDRAM DIMM to ensure data integrity. ANS9010 supports both ECC and Non-ECC registered DDR-2 memory. Users can obtain full capacity when ECC registered memories are used.The manual for download at the Acard website says:
The ANS-9010 supports both ECC and Non-ECC DDR2 SDRAM DIMM. It does not support Buffered nor Registered DDR2 SDRAM DIMM.The bad English is copied from the manual. It reads like it was written in Japanese and then quickly translated to English. I can't figure out if it means to say you need registered memory from my manual or not. First it says you can't use it, then it says you can. ECC does not mean it's Registered. I called Acard about this, and the guy barely speaks English. He couldn't really tell me much of anything I didn't figure out on my own after almost 7 hours of troubleshooting this thing. Needless, don't try to call them for help. Your dog/cat can probably help out more than they can.
I was unable to get my 8 4GB sticks of ram to work. The ANS-9010 does NOT work with cheaper high density modules. Took only 7 hours of troubleshooting and 3 phone calls to finally prove that I'm screwed and I'll have to return the memory I just got in the mail. Putting the ones I bought in cause completely unexplainable results. Errors both on reads and writes, and crazy crazy memory sizes like 15.462GB when there was 32GB of RAM installed. So no, don't try the high density modules or you'll be doing a big fat return for lots of $$$ like I am doing today.
I tested both DDR2-533 and DDR2-800 memory sticks. Both performed to within 2%. I think the 2% is within the tolerance of the test. I will be buying whatever is cheaper this weekend :P.
You can populate any size stick in any slots. Choose any combination of sizes and fill whatever slots you feel like filling, it all works the same.
Before I give the results, I would like to mention that using PCMARK and h2benchw would max out one of my cores(I have a q6600) for the tests, so these may be CPU limited.
Here's the Benchmark results:
Single Port tests
----------------
HD Tach -
Random Access: 0.1ms
CPU Utilization:3%
Average Read: 167.8 MB/sec
Average Write: 140.3 MB/sec
Burst Speed: 171.9MB/sec
The graph was perfectly horizontal, as expected.
h2benchw -
Interface transfer rate w/ block size 128 sectors at 0.0% of capacity:
Sequential read rate medium (w/out delay): 157501 KByte/s
Sequential transfer rate w/ read-ahead (delay: 0.45 ms): 149527 KByte/s
Repetitive sequential read ("core test"): 157399 KByte/s
Sequential write rate medium (w/out delay): 131372 KByte/s
Sequential transfer rate write cache (delay: 0.54 ms): 125681 KByte/s
Repetitive sequential write: 129962 KByte/s
Sustained transfer rate (block size: 128 sectors):
Reading: average 159140.5, min 82580.7, max 166318.5 [KByte/s]
Writing: average 124803.8, min 61775.2, max 138881.6 [KByte/s]
Random access read: average 0.05, min 0.03, max 0.07 [ms]
Random access write: average 0.05, min 0.04, max 0.08 [ms]
Random access read (<504 MByte): average 0.05, min 0.03, max 0.08 [ms]
Random access write (<504 MByte): average 0.05, min 0.04, max 0.07 [ms]
Zone measurements for read rates were all about 166MB/sec
Zone measurements for write rates were all about 138MB/sec
PCMark05 -
XP Startup - 124.747MB/sec
Application Loading: 109.865MB/sec
General Usage: 81.776MB/sec
Virus Scan - 102.181MB/sec
File Write - 119.448MB/sec
Next I ran benchmarks using the 2 port operation. I couldn't use RAID mode because the C drive is not RAID, so the drive would not be bootable. Software RAID0 was used to perform the tests. HD Tach cannot test software RAIDs.
First I ran benchmarks on the 2 "drives" simultaneously to see what would happen. Most people would choose to test them sequentially. When run simultaneously, each drive shows the full performance of single port mode. This would hint that when I set them up as a RAID, the performance will approximately double. That is actually not the case.
Results for RAID0 software:
PCMark05 -As you can see, some benchmarks changed, others did not.
XP Startup - 184.984MB/sec
Application Loading: 101.014MB/sec
General Usage: 106.912MB/sec
Virus Scan - 129.055MB/sec
File Write - 143.953MB/sec
h2benchw -
Sequential read rate medium (w/out delay): 157501 KByte/s
Sequential transfer rate w/ read-ahead (delay: 0.45 ms): 149527 KByte/s
Repetitive sequential read ("core test"): 157399 KByte/s
Sequential write rate medium (w/out delay): 131372 KByte/s
Sequential transfer rate write cache (delay: 0.54 ms): 125681 KByte/s
Repetitive sequential write: 129962 KByte/s
Sustained transfer rate (block size: 128 sectors):
Reading: average 159140.5, min 82580.7, max 166318.5 [KByte/s]
Writing: average 124803.8, min 61775.2, max 138881.6 [KByte/s]
Random access read: average 0.05, min 0.03, max 0.07 [ms]
Random access write: average 0.05, min 0.04, max 0.08 [ms]
Random access read (<504 MByte): average 0.05, min 0.03, max 0.08 [ms]
Random access write (<504 MByte): average 0.05, min 0.04, max 0.07 [ms]
Zone measurements for read rates were all about 166MB/sec
Zone measurements for write rates were all about 138MB/sec
Next I decided to plug the ANS-9010 into my RAID controller. I used a Highpoint Technology 2320 PCI-Express card to perform the tests.
RAID0 using hardware:
HD Tach -
Random Access: 0.1ms
CPU Utilization: 3%
Average Read: 222.9MB/sec
Average Write: 183.6 MB/sec
Burst Speed: 325.5MB/sec
The graph looks like waves from an ocean. See the attachment HDTACH.jpg to see for yourself.
PCMark05 -
XP Startup - 217.013MB/sec
Application Loading: 143.856MB/sec
General Usage: 150.068MB/sec
Virus Scan - 196.521MB/sec
File Write - 170.235MB/sec
h2benchw -
Sequential read rate medium (w/out delay): 166471 KByte/s
Sequential transfer rate w/ read-ahead (delay: 0.42 ms): 166272 KByte/s
Repetitive sequential read ("core test"): 166620 KByte/s
Sequential write rate medium (w/out delay): 138353 KByte/s
Sequential transfer rate write cache (delay: 0.51 ms): 138472 KByte/s
Repetitive sequential write: 138405 KByte/s
Sustained transfer rate (block size: 128 sectors):
Reading: average 167238.2, min 166625.9, max 167782.4 [KByte/s]
Writing: average 138959.6, min 138269.7, max 139480.7 [KByte/s]
Random access read: average 0.05, min 0.04, max 0.10 [ms]
Random access write: average 0.05, min 0.04, max 0.09 [ms]
Random access read (<504 MByte): average 0.05, min 0.04, max 0.12 [ms]
Random access write (<504 MByte): average 0.05, min 0.04, max 0.11 [ms]
Zone measurements for read rates were all about 166MB/sec
Zone measurements for write rates were all about 138MB/sec
As you can see, some benchmarks changed alot, some didn't change at all. PCMARK05 seems to have increased in all tests by using the RAID card. This makes me wonder if there is some kind of bottleneck when using my onboard controller. I'm thinking the SATA controller has a 1x PCIe lane to the southbridge, which is probably nearing saturation when I run RAID0. Strangely this doesn't show itself when I benchmarked the 2 drives independently
Overall, this thing is crazy fast. I don't care what others say, this thing is wicked fast. Most people here will admit that if you tried to copy 7000 little files totaling less than 500MB, it would take a while because of the number of files, not the actual size of data. I extracted about 7000 files totaling about 500MB from a .7z file using maximum compression to the drive in less than 10 seconds. I then duplicated the data on the drive in less than 10 seconds. I don't know anyone that has copied 7000 files in 10 seconds, let alone even 1 minute. I'm impressed!
I don't know if this box can use registered memory or not, but I am going to find out. ECC memory is cheaper online if it's also Registered, so I'm looking for the cheapest solution to fill this drive with 4GB sticks.
Also, the drive name in the BIOS will clue you in on what this drive is doing. If you look at BIOSPICT.jpg, you'll see how it works.
Each "field" is separated by an underscore.
The ANS9010 is the model.
The first 0 is the port. If you use the dual port mode, the first drive is 0, the second drive is 1.
The last one is the slots used. In my pict slots 0 and 1 were populated. The *'s indicate empty slots.
Also, for anyone playing around, here's the slot numbers for the slots from left to right when the front of the box faces you:
7,5,6,4 <Center> 0,2,1,3
I'll let everyone know how the Registered memory works and hopefully we can all find a cheap supplier for 4GB sticks.
Let me know if you have any other questions, and I'll try to answer them as best as I can.
Josh, I would certainly like to see the results with an onboard chipset in RAID mode if you get around to it. I have an ANS-9010 due in the mail Monday; glad I waited on the RAM purchase. I currently have an old Platypus QikDrive8 and two Gigabyte I-RAM Box style drives on my system, and the ACARD looks like the Holy Grail....
Wade
Ok, I can confirm that Registered memory does NOT work. I also investigated buying registered memory and removing the register chip, but that won't work either. Looks like we're stick buying either 2GB sticks on the cheap, or 4GB sticks at about $110 or so each :(. Anyone else know of any place that sells 4GB memory sticks that are unregistered and low density for less than $100 a stick?
I'll see about getting that done for you today wmaciv.
Just for giggles, I tried using one of those SATA to USB converters and benchmarked the box. I get a cool, crisp 30MB/sec. WOOHOO! ROFL. I did this to see what the actual maximum speed for USB2 is because USB2 wasn't designed for transferring data like is commonly used for thumbdrives and memory sticks.
Also, on my box is a sticker that tells you what all the jumper settings are. There is only 3 on the chart(there's 5 jumpers). But 1 of the 3 jumper descriptions is covered with a sticker that says "RESERVED". If you put a jumper in that location, it disables the ECC feature that takes non-ECC memory and makes it ECC. You lose 1/9 of your RAM size for this ECC feature. I think that they chose to not show this feature because of some kind of problem they had requiring the ECC feature to work.
Wow great results.
I read the device has enough battery to back up its contents on a CF card. Have you also tested this? Do you have more info on this feature?
I haven't bought a CF card for it, but the battery lasted about 2.5 hours when I tested it after leaving it to charge overnight.. that should be plenty of time to back up to CF.
Sorry, can you clarify that, as it seems you are contradicting yourself... :confused:Quote:
Originally Posted by josh1980
Are you saying the quasi-ECC feature with non-ECC memory IS working and the jumper to DISable it doesn't work, or that the quasi-ECC feature ISN'T working and the jumper to enable it doesn't work?
Sorry, it does sound confusing. The normal operation for non-ECC memory is to enable the quasi-ECC feature. This feature uses 1/9th of the total size of the memory, so whatever size you have installed you will lose 1/9th. If you put a jumper over the "RESERVED" jumper, which is the second jumper from the right when looking at it, this disables the quasi-ECC feature giving you the total capacity of the RAM sticks. I have a copy of their manual from 2 months ago, and it did have a feature to disable the quasi-ECC feature, but they strongly recommended you use it. I disabled it and ran read/write tests all night with no problems.
I would still recommend you not use a jumper there since they seem to have intentionally wanted to remove the 'feature' by not explaining the jumper in the manual.
I just had to know what was the purpose of the jumper since they deliberately covered up the sticker on the case for the jumper configuration.
Anything 8 gig and up acting as a permanent RAM drive you want to run with ECC. Doing it on the cheap with non-ECC is the wrong approach. Large ram drives simply have to use ECC RAM. So not using ECC and then disabling their "quasi-ECC" (however well that works) is a bad plan IMO.Quote:
Originally Posted by josh1980
From what you said earlier, I'm thinking of splurging for 2 4gb sticks at an unsavory cost, then finishing out the slots with 2gb sticks for a rough total of 20 gig, since 16 is cutting it a little close for what I want to install (OS and a few key apps). I really wonder how this thing will behave with the Intel ICH9R chipset on my MB (Supermicro C2SBX). It had some early issues with the I-RAM's, before a BIOS update.
Speederlander, why 8GB? What happens at 8GB that makes ECC important? Or are you just trying to say that if you use it as a permanent RAM drive you should use it? I wouldn't recommend ever disabling the feature. 1/9th of your capacity isn't much to lose to prove your data is correct via ECC.
Me personally, I intend to buy 4GB sticks for all 8 slots and make it my boot drive, with 2x1TB hard drives for storage space. I have a RAID5 array consisting of 8x1TB drives, so I don't really need to keep alot of storage space on my workstation.
I'm sorry I haven't had a change to test the RAID0 using the onboard RAID chipset. Work has kept me busy this weekend, but I intend to get results as soon as I can. I'm just as curious as others to see what the RAID chipset can do. My motherboard is an Asus P5E if anyone wants to look up the RAID chipset. I think it's the ICH9R.
I say 8GB only because people tend to add in 4GB increments. At 4GB, RAM suffers from the occasional transient bit error but probably not enough to worry, especially when used as simple volitile memory (though I would, since I really value my data).Quote:
Originally Posted by josh1980
The original Corsair rule I recall is: 1 bit error occurs in 256MB of ram every month.
4GB = 15 bit errors/month
8GB = 31 bit errors/month
16GB = 62 bit errors/month
32GB = 125 bit errors/month
However, other more recent sources maintain 1 bit error per gigabyte per month, so that cuts those numbers by a factor of 4, i.e.:
4GB = 4 bit errors/month
8GB = 8 bit errors/month
16GB = 15 bit errors/month
32GB = 31 bit errors/month
Which is more correct? My guess is that it's somewhere in between, governed by the quality of the RAM, the luck of the draw, and a host of other issues.
There is a reason that server boards only use ECC RAM.
At 8GB, is the incidence still acceptable? You have to decide. But if that RAM is now operating as an always on, permanent drive, to not use ECC when you will probably be using 16GB or more would be nuts.
+1 QFT. Which is why the software quasi-ECC feature is such a brilliant idea, at the cost of a small amount of capacity, if the cost of ECC memory is too high.Quote:
Originally Posted by Speederlander
My ANS-9010 just arrived this evening. I have to go out of town tomorrow morning for 10 days... argggghhhhh! Anyway, I don't have any memory for it yet, think I may go with 16 gigs of 4 x 4gb dimms initially... And will spring for the 32 meg CF card also... wonder if it's worth spending the premium on high transfer rate CF memory? Bet not...
Hopefully I can post some Platypus to ACARD and IRAM to ACARD transfer stats in the near future.. I can't WAIT to see this as my boot drive...
Wade
Who has it in stock? Website still saying pre-order!! Here's an interesting thread re the Acard on the Anandtech forum.
http://forums.anandtech.com/messagev...VIEWTMP=Linear
:( I thought single port would have maxed out 1x SATAII port = ~300MB/s ... why is it only 168? :(Quote:
Originally Posted by josh1980
So only ECC memory can be used with it. That sounds pricey.
No one should be surprised. ECC is the only serious option for larger RAM drives.Quote:
Originally Posted by B.E.E.F.
I hope I can eventually get my hands on a unit to give it a thorough investigation, but I currently theorize that the reason is the controller on the ram disk. Technical specs show that it provides ~200MB/s throughput maximum (which is one reason it's sure not going to hit 300MB/s), but the question is whether that includes SATA overhead or not.Quote:
Originally Posted by tiro_uspsss
If it does, then something else is to blame and that bears further investigation.
If it does not, then 168 is pretty close to right if the controller can provide a little over 200MB/s performance.
mm.. well.. personally I had hoped for more considering the price :(Quote:
Originally Posted by Serra
*sigh*
The unit does not require ECC memory, but can use it. If ECC memory is installed, the unit will use the ECC feature to fix errors. If ECC memory is not installed, then 1/9th of the total capacity of the RAM is used for ECC functions, and is not used. The big difference between ECC memory and non-ECC memory is a 9th chip on a RAM stick. This 9th chip is identical to the other 8, but the data stored in the 9th chip is similar to a checksum of the RAM bank. If you chose not to buy ECC memory for this unit, then it will create it's own ECC function utilizing 1/9th the space.
I agree with tiro_uspsss that 168MB/sec is somewhat limiting, but it could be a limit from something else besides the box. I do own a Gigabyte I-RAM card, and I benchmark it at 120MB/sec(the limit for SATA 150).
I also did order 4x4GB of RAM today for my box. I'll buy 4 more 4GB sticks next week when I get paid again. This box is going to make my core i7 computer boot at light speed. :D
Also, the motherboard RAID benchmark's I've been promising to do will be performed tonight. I'll post the results when I get them done. :)
Just catching up on this thread and have a question. I see reading back a bit that Access time is 0.1ms ?
I thought that these units are faster than that ?
I'm surprised. They can increase sales by making it non-ECC compatible.Quote:
Originally Posted by Speederlander
:fact:Features:-Quote:
Originally Posted by B.E.E.F.
Supports up to 64GB amount of memory
Supports ECC/Non-ECC DDR2 400/533/667/800**
Automatic data backup/restore between DDR2 memory and CF card
Built-in Lithium Battery
LED indicators for battery capacity, power status, SATA ports activity, backup status
Driver less: need no driver on host side
**Note: Please refer to Compatibility List for DDR2 RAM Module
source:
http://www.acard.com.tw/english/fb01...ve&idno_no=270
Running a big RAM drive without ECC means the person has no concern whatsoever for data integrity.Quote:
Originally Posted by B.E.E.F.
Josh, what kind of ram and where are you getting 4gb density sticks? Are you paying a premium, or have you found an inexpensive source? I would sure like to find some for my 9010 without spending what I think I am going to have to.....
Wade
You are correct. But most benchmark programs are written to give the results for the hard drive access time in ms, and they can't calculate below 0.1ms. They all round, and in this case, they round up to the smallest possible division of 0.1ms.Quote:
Originally Posted by Buckeye
The h2benchw tests showed the following:
Random access read: average 0.05, min 0.03, max 0.07 [ms]
Random access write: average 0.05, min 0.04, max 0.08 [ms]
Even at 0.1ms, you wouldn't hear a peep out of me. I don't know of any other hardware out there with these kinds of seek times. These are 100ths of a millisecond. All I can say is "Awesome".
I was playing around some today, and I figured out something else that I figured I would share. My BIOS has SATA options of Disabled/Compatible/Enhanced. If I chose "Enhanced" I lose about 10% performance vice "Compatible". Also worth mentioning is disabling ECC feature does not result in any change in performance.
I spent some time attempting to use the onboard RAID controller for my motherboard, and I was unable to get it to work. Unfortunately I do not have a spare hard drive to install another copy of Windows on, so we'll have to wait until either I get my new core i7 computer next week, or someone else can perform some benchmarks.
Absolutely not. These are my RAM drive.Quote:
Originally Posted by josh1980
That's right.Quote:
Originally Posted by Speederlander
My bad. Thanks.Quote:
Originally Posted by RADCOM
Yes, but I assume that is a drive carved out of your system RAM, not a peripheral. To speak with your physical RAM there is effectively no wait time... but the question is, does the program perhaps perceive that there is a "non-zero" wait-time to a hard-drive (or SSD) and therefore just round up to 0.1? I personally think it is likely.Quote:
Originally Posted by Speederlander
I was responding to:Quote:
Originally Posted by Serra
Also, I-ram shows as 0.0ms, and it's a peripheral, not directly created from system RAM. And IIRC, doesn't it also use the SATA interface, just using PCI for power? Not that .1ms is a lot.Quote:
they can't calculate below 0.1ms
Great ! Thanks for that info on access time.
I am looking at 2 of the 9010 installed onto my ARC-1231ML and set in Raid 0. From what I gather the 9010 has 2 SATA ports on each unit, so I would need 4 SATA ports total. This matches up just fine with the 4 SATA ports I have left on my controller card.
I wounder now how well these will scale in a setup like this. That would be 128gig total and a OS installed on that. I am thinking that would be in the range of 800mb/sec bandwidth with 0.05ms access time. Pretty fast.
However with my current SSD Raid 0 I am in the range of 950mb/sec bandwidth and 0.1 access time.
Side by side these 2 setups should perform pretty close but with the ACRAD mopping the floor with writes and small file access.
Should be an interesting test thats for sure.
I-RAMS are looking rather interesting to me also, I don't like the fact that they need a slot for power. I can just see 12 I-RAMs connected to my ACR-1231ML as a Raid 0, but as far as being usefull for anything other than a bench is not good.
i-rams dont work with areca cards :down:Quote:
Originally Posted by Buckeye
Bummer :down:Quote:
Originally Posted by tiro_uspsss
:yepp:Quote:
Originally Posted by Buckeye
Josh, I only had time to pop the cover on mine before I left on my trip... The slots for the RAM looked pretty tightly spaced. Do you think many heat spreader equipped DIMMS are too big to place adjacently? It almost looks that way, just curious about your thoughts on the matter. BTW, any playing around with the CF slot yet?
Wade
Ok, I got 4x4GB in my box along with 4x2GB sticks. I ordered the Gskil 4GB sticks from newegg. Strangely(or maybe not) the heatsinks aren't glued to the RAM at all. In fact, I turned the box upside down while I was talking on the phone, and a heatsink fell off. I haven't ever seen this before. They have always been firmly glued onto the chips. All 4 sticks are like this, so I had a great opportunity to see about heat problems inside the box both with and without heatsinks.
I tested the temps using the slots that have RAM installed on both sides and after heavy use of the drive to ensure the sticks got as I could resonably expect.
Overall I noticed no significant difference in temperature between the sticks with the heatsink and the sticks without a heatsink. However the center chip on the board with the heatsink gets extremely hot.
I have decided that I am going to look into installing a fan somehow to keep some air flowing through the box to keep everything cool. Since these RAM sticks have heatsinks that are easily removed, I have removed all of the heatsinks from the 4GB sticks. There is some room, maybe 1.5mm tops, between the memory sticks when the heatsinks are installed. I just think that there is more opportunity for cooling without the heatsinks. Besides, if the heatsinks aren't securely fastened to the ram chips, how much heat can you reasonably expect to be transferred?
I do not have a CF card to test in my box. Right now I'm not planning to get a CF card. I did buy the external power supply for the ANS-9010, and I do have a UPS that the computer will be hooked up to, so I don't expect to have any problems with losing data in RAM. I also plan to do periodic backup images of the drive for safe keeping just in case. When CF get's a bit cheaper or there is a fire sale I'll get a good CF card. I'd just rather continue to put money into more RAM for the box before I get the CF card.
KVR667D2E5/4G, KVR667D2E5/2G and KVR667D2E5/1G are Kingston ECC Unbuffered/Unregistered sticks according to kingston sales dept
the 1GB and 2GB variety are very reasonably priced and will let you avoid any processing overhead of running the virtual ecc mode
The difference between the non-ECC and quasi-ECC feature was not significant(test results between the two were less than 1%). With this knowledge, I would avoid buying ECC unless the price is less than 1/9th the difference of non-ECC memory, or you do have a need for that 1/9th more RAM.
Now, there may be something unique to the design of the box, where ECC memory will perform faster than non-ECC(or using the quasi-ECC feature), but as I do not have ECC memory to test this theory, it is unproven.
I honestly wouldn't expect ECC memory to perform any faster than non-ECC(see above), but stranger things have happened.
the kingston 1gb sticks are ~14.00$, the 2GB sticks are ~27.00$ and quite unfortunately the 4Gb sticks are around 330.00$
Well, the 4GB sticks on newegg are $109 each.... If you price shop, you can find some good deals where 2Gb sticks are <$27 each.
Obviously though if you are getting 4GB sticks, you'd be crazy to buy ECC at that price. Today my core i7 computer will be assembled. The motherboard is the GA-EX58-UD5. I'll be installing Windows XP on the drive, and you bet I'll be performing benchmarks shortly after. I'll post them as soon as I get them.
Here's the new benchmarks for those interested. These were performed using the hardware RAID-0 of the ICH10R chip(in this case on a Gigabyte EX58-UD5 motherboard). I did the full benchmarks for h2benchw and HDTach using all available stripe choices. It appears that the ideal stripe size for these boxes are 16k. Since there was so many results, I decided to provide zip files of them instead of pasting it all here. The 16k stripe benchmarks for h2benchw are show below.
I did install Windows XP Pro on the box. This thing is so freakin' fast I can't believe it. I updated a program, and the update screen flashed so fast that I didn't even realize the program updated. I was able to install Windows XP in less than 10 minutes, then upgrade from SP2 to SP3 in 117 seconds. That's just AMAZING!
Ok, here's the results for 16k stripe size.....
nterface transfer rate w/ block size 128 sectors at 0.0% of capacity:
Sequential read rate medium (w/out delay): 304925 KByte/s
Sequential transfer rate w/ read-ahead (delay: 0.23 ms): 311081 KByte/s
Repetitive sequential read ("core test"): 304936 KByte/s
Sequential write rate medium (w/out delay): 262211 KByte/s
Sequential transfer rate write cache (delay: 0.27 ms): 266516 KByte/s
Repetitive sequential write: 261967 KByte/s
Sustained transfer rate (block size: 128 sectors):
Reading: average 306144.9, min 285985.0, max 309198.2 [KByte/s]
Writing: average 262580.0, min 243938.3, max 265505.7 [KByte/s]
Random access read: average 0.05, min 0.04, max 0.06 [ms]
Random access write: average 0.05, min 0.04, max 0.06 [ms]
Random access read (<504 MByte): average 0.05, min 0.04, max 0.06 [ms]
Random access write (<504 MByte): average 0.05, min 0.04, max 2.56 [ms]
Application profile `swapping': 116846.4 KByte/s
Application profile `installing': 351531.5 KByte/s
Application profile `Word': 253346.4 KByte/s
Application profile `Photoshop': 250514.4 KByte/s
Application profile `copying': 386762.0 KByte/s
Application profile `F-Prot': 168442.1 KByte/s
Result: application index = 229.8
If you look at the HDTach benchmarks, you'll notice that the benchmarks become really odd at larger stripe sizes.
Sorry if the attachments aren't fancy. I'm tired and stayed up to make sure to get these benchmarks done tonight. Enjoy everyone!
Thanks Josh !
I really like these babies :)
Thanks Josh1980!
From your previous posts, I see you have used a number of different memory configurations. What you are using now and how much? Are you using ECC? Also...is it possible to throw for example 4 x 4GB sticks with 4 x 2GB sticks in the box?
And one more question for everyone... has anyone tried using two 9010a or b's off an onboard or discrete raid controller?
I have 4x4GB, 2x2GB, and 2x512MB sticks. Some are 533 and some are 800. There is no requirement for pairs, and you can put different sizes in the box in any order and any speeds. None of the chips are ECC. The cost versus benefit wasn't worth it(cost is quite high for 4GB sticks).
Spoiler - There is a set of benchmarks I did with a pci-e RAID card I had. Not sure if that helps you at all since you wanted onboard and discrete. The benchmarks I posted yesterday are from an onboard card, while the first set I did was using a software raid as well as pci-e card.
Does anyone else notice that when you post messages it says it got posted, then it shows you the thread with the new message. Yet when you go and check later the thread is missing the post? I've had this happen at least a half dozen times on several different computers using both IE and Firefox.
What card?Quote:
Originally Posted by josh1980
This thing seems to be so much faster than any other hard disks out there. The only drawback is we need to spend on lots of RAM :(
My card is a Highpoint RocketRAID 2320.Quote:
Originally Posted by Speederlander
As comparison my 2x core v2 ssds in raid 0 on ICH9R get about half of those rates at app index 125.9.Quote:
Originally Posted by josh1980
Ok, so here's my latest and greatest thoughts on this hardware....
I will have spent over $1000 on this little box with 32GB of RAM by the time all is said and done. I have a UPS connected to my computer, with the ANS-9010 connected to a second power supply. I bought one of those IDE/SATA to USB convert kits and used the power supply for the ANS-9010. The cord comes in through the back of the computer to the ANS-9010. This gives me UPS power to the drive and it has power even when the computer is off. There is a power supply kit you can buy for it, but it costs $30 and my idea is identical to theirs but less expensive. The internal battery lasts about 2.5 hours if the box loses external power.
This drive is my boot drive. I installed the software that usually has alot of boot time loading requirements like Yahoo IM, MSN IM, Defrag(defragmenting this drive is essentially pointless), Antivirus, etc. Any other software that doesn't have high boot loading requirements I install onto my 1TB drive. My boot times have suffered very little from this method. Even with all of my software installed my boot times are less than 1 minute. It's still so much fun to load Windows and watch all of the programs in the toolbar load in seconds. I also have a Gigabyte I-RAM card with 4GB of RAM on it. I chose to use the I-RAM as a swapfile location.
I am quite impressed with the performance of this box. Being that I had money to throw at it, I find it a worthwhile investment(I could have spent it on alot more things that do less for me). I can't help but wonder if this was really worth the $1000 I'll spend on it. I don't really do video editing, or any of the stuff that would see the biggest performance boost from this drive. It's for my workstation that I do everything on from games to programming to Office related work. I will say that I do not fear the 5 minute reboots like I used to.
This box is amazing, but I would venture to say that 2GB sticks would have probably sufficed for my goal.
This box is definitely not for those with small pocketbooks. These things are not cheap to set up, and there is a higher probability of losing one's data since the box losing power = wiped drive.
Now that the Intel X25-E is being released, I would have some mixed feelings too about spending $1000. The intel drive is somewhat comparable in speed, but cost almost $300 cheaper.Quote:
Originally Posted by josh1980
it is not comparible in speed in any way shape or form.Quote:
Originally Posted by Spoiler
For what Josh will be doing with it, I would believe that the speed of the two devices would be very much comparable. I would definately agree that the ACARD device surely wins in the benchmarks, but I can't imagine myself noticing much difference between the two devices when the main usage would be for gaming and general office work.
Well, once you get fast enough, doubling the speed isn't easily noticed anymore by the user. It only becomes an issue of faster benchmarks. It might feel a little snappier, and a few non-critical things like boot time may be faster, but somewhere down the line you are bottlenecked by the requirement to push things through the CPU, etc.Quote:
Originally Posted by One_Hertz
I would like to see this drive benched for real-world feel and compared to SSDs on anand or somewhere else using their full suite of apps.
the fact that this ram drive doesnt get raped by random writes or just a simple combination of reads and writes means a lot. The agree that the STR means very little.
Where do you see the X25-E getting raped in random writes? For one, I am only basing my opinions off the latest review from the techreport.com And second, I could be wrong, but I thought the SLC devices didn't suffer from random writes nearly as bad as the MLC ones.Quote:
Originally Posted by One_Hertz
http://www.alternativerecursion.info/?p=106
When they run benchmarks, the drive showed it as top of the line in seek times, read and write performance. BUT, when it comes to random writes, it was horribly slow(4IOPS ROFL!). This was a MLC drive, not SLC. I'm very interested to see a similar breakdown in an SLC drive. If anyone finds a good benchmark for random writes of an SLC drive an link would be appreciated.
I've always been a little skeptical of SSDs ever since reading that article. Tom's Hardware did a big review of a bunch of SSD drives in August, and the 1 benchmark they didn't provide was random writes. When I can see SSD boot as fast as my computer does, i'll be a believer again. A ramdrive would make a significantly better pagefile drive than an SSD just because you don't 'wear out' ram drives.
Of course, SSD is new, and it will only improve. I see SSD drives being made with 2 principle purposes:
1. MLC - Drive that are designed for backup and long term storage(think your video collection, pictures, application zips, etc).
2. SLC - Drives that are your boot drives(your installed OS, Programs, User Profile).
Of course, some company could invent a 3rd hybrid that uses a combination of both chips.
I'm expecting that in 5 years I'll be buying MLC drives for that "big drive" need, and SLC for boot drives.
In the test that you linked to the problem is in JMicron controller, not MLCs.Quote:
Originally Posted by josh1980
X25-E review.
MLC drives can be very fast, just as good as SLC except for writes, especially small and random ones.
They are very good whenever you're streaming large data portions and are much cheaper than SLC.
I expect that the top consumer drives will be MLC + large RAM cache, that seems to be a better option than SLC...unless you have a server and need extra durability.
Ironically HyperOs have released a 64GB £269/$399 drive....looks very similar to the Acard too.......lets play spot the difference :) This is just after this months custompc magazine review put a "crazy but cool" and ""recommended tag on the Acard 9010 and 9010B. The older HyperoS drives cost in excess of £1200 and used registered PC2700 DDR1 memory.
http://www.hyperossystems.co.uk/
We are looking at the same place. Look at this page:Quote:
Originally Posted by Spoiler
http://techreport.com/articles.x/15931/7
It dips down to 50-60mb/s. Do the test on josh's setup and you will see 150+mb/s. I do not see how you can call that comparable performance, but maybe that is just me. Windows boot time would be A LOT faster on the ram drive and program load time would be somewhat faster.
I might be looking at this the wrong way, but I figure the ACARD would be a tad faster than the old i-RAM. With that reasoning, I woud imagine the Intel SLC SSD would be right around the same speed in game/system load times. Take a look here at the i-Ram boot and load times :Quote:
Originally Posted by One_Hertz
http://techreport.com/articles.x/9312/3
I know it's not a fair comparison, but it give us a good idea.
The ACARD 9010 does seem like a real screamer though. I thought about purchasing one last week, but nobody had them in stock. Lucky for my wallet I guess. :)
ACARD should be about 2x faster than the old IRAM (which was limited by the PCI bus)... Although there is nothing to back that up, just what I think.Quote:
Originally Posted by Spoiler
Nope, not limited by the PCI bus, as it doesn't use that to transfer data, just supply power. It connects via SATA1, same speed as the ACARD now appears to (even in SATAII mode). So they are almost on par in performance. That's the disappointment, really, that 4 or 5 years of massive tech advances hasn't produced a faster RAM drive, simply becase it's being bottlenecked by poor interfacing. :(
Well, well. That IS a comedown from their crazy stupid prices on the old models. But for my money, with otherwise identical specs (and it IS odd, look at the design, and the spec for transfer rates!) the ACARD wins with the quasi ECC feature. It's just more reliable over time as long-term storage.Quote:
Originally Posted by RADCOM
The IRAM was not limited by the pci bus, but limits by sata I. The acard uses sata II, so should not be as limited there, but look here:Quote:
Originally Posted by One_Hertz
http://www.wideopenwest.com/~dcason6634/Acard.html
comparable
Nox
I could be wrong but i believe they are just reselling the Acard produced device and possibly have pictures of slightly different revisions of the same device.Quote:
Originally Posted by IanB
I see two IRAM devices being put against one 9010? The benchmarks are right in this thread as well. 270mb/s versus the 135 of the iram. Access times should be similar.Quote:
Originally Posted by nox_uk
The ACARD has a similar streaming speed to the i-RAM PER SATA PORT. The bigger model has two SATA ports which you can access separately, each offering half the installed RAM as a disk device. So two ports in RAID0 is roughly equivalent (a little faster) to two i-RAMs in RAID0. If the single-port option is used, then you will get a disk the full size of the RAM, so bigger but not much faster than an i-RAM.Quote:
Originally Posted by One_Hertz
Who the heck is talking about the one port version? It is pure garbage more or less. The fact remains, the proper version is 2x faster than the old IRAM.Quote:
Originally Posted by IanB
NO, it's not. You are missing the point. It's not about the unit, it's the fact that you have to use TWO SATA ports to interface to the bigger unit in RAID0. So equating this to a SINGLE i-RAM that interfaces to a SINGLE SATA port is just ludicrous. It's also twice the cost and takes 2 to 4 times the memory. :rolleyes:Quote:
Originally Posted by One_Hertz
Using the bigger unit in two-port mode is effectively having two half-size units similar to an i-RAM each, needing the same interface requirement, so it's only fair to compare this to two i-RAMs similarly RAIDed. :shakes:
The point is, it *should* be twice as fast per SATA port!!!! The SATA II on the acard appears to be no faster than the SATA I on the IRAM, well OK a little faster, but certainly not up to double the speed. That ACARD is like having two IRAMS in one box, near enough, and it should be faster, nearer double the speed. If you use your sata ports on your mobo, you could plug in 2 acards, or twice as many irams, net result - roughly the same, give or take.Quote:
Originally Posted by One_Hertz
The REALLY annoying thing is I was holding off for this tech, and may still get one, but the top end SSD's are starting to get into the same ball park for comparisons, if you take into consideration capacity and retaining data i'm now leaning towards those... If I do get an acard, then I can see it being used as a page file drive to try to help the lifespan of any ssd's i get. Had the ACARD been released 2 years ago, or even 1 it would of been a no brainer...
Nox
The price of the single port ACARD device is pretty nice, it's just too bad the maximum transfer rate through that single SATA connection isn't anywhere close to the theoretical maximum of SATA II.
As someone posted earlier, it sounds like the ACARD controller isn't fast enough. I saw something about another ACARD device coming out scheduled for early next year. Could this be a revision?
if it is, it's just in time for SATA III :D (or whatever they plan to call it)
Nox
As it's been already mentioned (i.e. 1 post above:rolleyes:), it's a fault of ACARD, not SATA2.Quote:
Originally Posted by nox_uk
Intel could get well over 200 MB/s from the same interface.
yep, which is pretty much why i'm leaning towards SSD at the moment...
Market on this area is changing so fast though I do plan to wait 6+ months, see where things like the iodrive and this are standing compared to SSD's when the dust begins to settle. Was very tempted with a couple of OCZ Core V2's, or could get 3 or 4 now for the same price as one of these & eight 2gb sticks... then again the v3's are around the corner...
Nox
Do you not have any empty sata ports? What is your point? What does it matter if it uses another one? Equating two hardware devices to one and saying that it is the same is ludicrious. Why dont you take two 4850s put them into CF and say that 4850 is comparable to 280gtx(which would take up the extra slots with its huge cooler)?Quote:
Originally Posted by IanB
If it costs pretty much the same and gives pretty much the same results, then it's not ludicrous. :shrug: In both cases the single box is merely more convenient.
As others have just pointed out, this isn't revolutionary tech, it's barely evolutionary tech equivalent to putting two i-RAMs in a box using DDR2. For the market it's aimed at (boot device), no-one would use a single i-RAM, it's too small, you have to RAID anyway to get a usable drive size. So the comparison is obvious. The killer feature here isn't the streaming speed as we might have hoped, it's merely the memory capacity and ECC feature, and a neat integrated backup gizmo.
nox_uk is right, it's still a great product, but there's nothing here that couldn't have been produced years ago, and it may be too late to make a niche with faster and cheaper SSDs widely available. And since I have two i-RAMs, there's simply not enough difference in the performance to justify a switch to this. :(
I would say my 4870x2 is more comparable to two 4870's in crossfire, than one on its own...Quote:
Originally Posted by One_Hertz
cost is near enough the same, performance, near enough the same, just one takes up half the space and is more convenient...
Nox
Josh (and whoever else)... Have 24 gig and a 32 gig flash card installed and running on the 9010 via ICH9R. Was a little trouble at first, but smoking now. Holy smokes, the flash card backup takes FOREVER! I will try to post some relative times, but definitely going the way josh did for fulltime power backup. What an awesome kit, but really want to go for the full 32gb.
Wade
I was just browsing the memory compatability sheet listed on ACARD's website. It looks like more sticks have been added! :up:
I was backing up a 8GB ramdisk on shutdown (to the hard disk) and it'd take roughly 15 minutes. Most CF doesn't write nearly that speed (even though it wasn't a particularly fast HDD) and you might be dealing with more than 8GB of storage.Quote:
Originally Posted by wmaciv
C2SBX w/Q6600 @ 3.0 gHz 1333 mHz FSB 2GB RAM
ANS-9010 has 24GB of ddr-2 800, and a 32mb Transcend CF card (133x)
After playing with the ACARD for about a week now, I pulled my two I-RAM box drives out, and left in my Platypus Qikdrive8. I may end up pulling it too when my last 8GB for the ACARD shows up this week. This drive is everything I had hoped for. OS is XP sp3 w/ all recent patches, Office 2007 Enterprise, Adobe Reader 9, WinRar 3.71, assorted Malware, Symantec AntiVirus corporate, video and sound drivers, etc. All major applications are on a WD Velociraptor 300GB.
Wade
Interesting, I'm not hyper impressed by the HDTach numbers. That's a dual sata RAID 0 so it's actually 140 per sata port. Not drastically better than the raptor (or a pair of raptors with a small fast OS partition). I understand the random I/O capability is the key here, but it just doesn't seem compelling given the cost-to-capacity issue compared to the raptors or even several SSDs out now. I DO like the ECC function for data protection. To me that's a primary selling point. Something a typical RAID 0 or single drive would not enjoy the benefit of. An array of these would certainly be impressive, but so very expensive.Quote:
Originally Posted by wmaciv