RAID0 WD800HLFS HD Tune Pro benchmarks: Intel ICH10R vs Areca 1200 vs HighPoint 3520

E

EnderW

[H]F Junkie
Joined
Sep 25, 2003
Messages
11,249
CPU: Intel Q9450
Motherboard: Asus P5Q Deluxe
Boot Disk: Intel X25-M SSD (80GB)
Memory: Corsair Dominator TWIN2X4096-8500C5DF 4GB PC2-8500
Video Card: Sapphire AMD HD4870
PSU: Corsair HX620W
OS: Windows Vista Ultimate 64-bit
Intel ICH10R Driver: Intel Matrix Storage Manager 8.6.0.1007


Summary Table thanks to coo-coo-clocker
wd%20bench.jpg



Single WD800HLFS Read
HDTune_Benchmark_WDC_WD800HLFS-75G6U0_drive01_read.png



Single WD800HLFS Write
HDTune_Benchmark_WDC_WD800HLFS-75G6U0_drive01_write.png



2 x WD800HLFS RAID0 Intel ICH10R Read
HDTune_Benchmark_Intel___Raid_0_Volume%20-%20read.png



2 x WD800HLFS RAID0 Intel ICH10R Write
HDTune_Benchmark_Intel___Raid_0_Volume%20-%20write.png



2 x WD800HLFS RAID0 Areca 1200 Read
HDTune_Benchmark_Areca___ARC-1200-VOL%2300%202xWD800HLFS%20read.png



2 x WD800HLFS RAID0 Areca 1200 Write
HDTune_Benchmark_Areca___ARC-1200-VOL%2300%202xWD800HLFS%20write.png



2 x WD800HLFS RAID0 HighPoint 3520 Read
HDTune_Benchmark_HPT_____DISK_0_0%202xWD800HLFS%20read.png



2 x WD800HLFS RAID0 HighPoint 3520 Write
HDTune_Benchmark_HPT_____DISK_0_0%202xWD800HLFS%20write.png



3 x WD800HLFS RAID0 HighPoint 3520 Read
HDTune_Benchmark_HPT_____DISK_0_0%203xWD800HLFS%20read.png



3 x WD800HLFS RAID0 HighPoint 3520 Write
HDTune_Benchmark_HPT_____DISK_0_0%203xWD800HLFS%20write.png



4 x WD800HLFS RAID0 HighPoint 3520 Read
HDTune_Benchmark_HPT_____DISK_0_0%204xWD800HLFS%20read.png



4 x WD800HLFS RAID0 HighPoint 3520 Write
HDTune_Benchmark_HPT_____DISK_0_0%204xWD800HLFS%20write.png



4 x WD800HLFS RAID10 HighPoint 3520 Read
HDTune_Benchmark_HPT_____DISK_0_0%204xWD800HLFS%20RAID10%20read.png



4 x WD800HLFS RAID10 HighPoint 3520 Write
HDTune_Benchmark_HPT_____DISK_0_0%204xWD800HLFS%20RAID10%20write.png



4 x WD800HLFS RAID5 HighPoint 3520 Read
HDTune_Benchmark_HPT_____DISK_0_0%204xWD800HLFS%20RAID5%20read.png



4 x WD800HLFS RAID5 HighPoint 3520 Write
HDTune_Benchmark_HPT_____DISK_0_0%204xWD800HLFS%20RAID5%20write.png



4 x WD800HLFS RAID6 HighPoint 3520 Read
HDTune_Benchmark_HPT_____DISK_0_0%204xWD800HLFS%20RAID6%20read.png



4 x WD800HLFS RAID6 HighPoint 3520 Write
HDTune_Benchmark_HPT_____DISK_0_0%204xWD800HLFS%20RAID6%20write.png




Couple of pics
Lian%20Li%20HTPC%20057%20%5B1024x768%5D.JPG



Lian%20Li%20HTPC%20058%20%5B1024x768%5D.JPG




Update: A couple of benchmarks from when I had some SSDs

4 x Mtron MOBI 3000 RAID0 HighPoint 3520 Read
52xjqs.png


4 x Mtron MOBI 3000 RAID0 HighPoint 3520 Write
2myz0qu.png
 
Those Highpoint 3520 results are amazing. *checks price*

ICH10R did well enough.

Edit: As I suspected... too expensive for my needs :)
 
Why is the areca so slow on the writes compared to everything else? I figured it would perform equally as well as the ich10r
 
You would think that with write caching enabled, it would be faster. Weird...
 
Archmage said:
Those Highpoint 3520 results are amazing. *checks price*

ICH10R did well enough.

Edit: As I suspected... too expensive for my needs :)
Click to expand...
Yeah, it's not cheap. There is a 4 port version available too, the 3510, which is about $100 less.

nitrobass24 said:
Why is the areca so slow on the writes compared to everything else? I figured it would perform equally as well as the ich10r
Click to expand...
No idea. It's worse than a single drive. I kept thinking something must be configured wrong, but ran out of things to change.

Blue Fox said:
You would think that with write caching enabled, it would be faster. Weird...
Click to expand...

With write caching off, it was getting around 40 MB/s.
 
EnderW said:
With write caching off, it was getting around 40 MB/s.
Click to expand...
Now that's just horrible. What makes me laugh is the fact that Areca advertises OCE and RAID level migration support for this card on their website. :p
 
Blue Fox said:
Now that's just horrible. What makes me laugh is the fact that Areca advertises OCE and RAID level migration support for this card on their website. :p
Click to expand...
I'd really like to see some other results. I'm surprised to see this level of performance knowing Areca's reputation for fast controllers.
Basically what I'm saying is take these with a grain of salt and let me know if there is anything you think I should try.
I played around with the stripe size, tried 64kb and 128kb, if I remember correctly, 128kb was the default and what the results you see are from, and 64kb was a bit worse.
 
These are my settings for my main array:
Code: 
Volume Set Name 	ARC-1280-VOL#00
Raid Set Name 	 	Raid Set # 00
Volume Capacity 	11000.0GB
SCSI Ch/Id/Lun 	 	0/0/0
Raid Level 	 	Raid 5
Stripe Size 	 	64KBytes
Block Size 	 	4096Bytes
Member Disks 	 	12
Cache Mode 	 	Write Back
Tagged Queuing 	 	Enabled
Volume State 	 	Normal
 
Only thing I can think of is that sometimes Areca Controllers are just picky about drives
I remember Ockie had probs with WD1Tb drives on his Areca Card.
 
Added RAID10 benchmarks at the request of someone at XS.
Gonna do RAID5 and RAID6, then I'm going to bed :)

Blue Fox said:
These are my settings for my main array:
Code: 
Volume Set Name 	ARC-1280-VOL#00
Raid Set Name 	 	Raid Set # 00
Volume Capacity 	11000.0GB
SCSI Ch/Id/Lun 	 	0/0/0
Raid Level 	 	Raid 5
Stripe Size 	 	64KBytes
Block Size 	 	4096Bytes
Member Disks 	 	12
Cache Mode 	 	Write Back
Tagged Queuing 	 	Enabled
Volume State 	 	Normal
Click to expand...
Looks pretty close. I have the HighPoint in right now so I can't compare at the moment. I will try a couple more things before I get rid of it.


nitrobass24 said:
Only thing I can think of is that sometimes Areca Controllers are just picky about drives
I remember Ockie had probs with WD1Tb drives on his Areca Card.
Click to expand...
Yeah I've heard a lot about how picky Aerca controllers are, but I was under the impression most of those issues were related to drives being dropped from arrays.
 
RAID5 and RAID6 benchmarks added
this HighPoint card is a beast!!
 
Not too shabby. Price is pretty competitive too. I have a feeling it would be a bit faster than the Areca 1220 due to the newer processor. PERC 5/i still beats it in price though.
 
Are you going to use the VRs in raid as your boot drive? How do you like them compared to your SSD?
 
Thanks a ton Ender for these benches... good stuff. Takes time I know. These are the things we all wanna see but don't wanna take the time...
 
anodize said:
Are you going to use the VRs in raid as your boot drive? How do you like them compared to your SSD?
Click to expand...
No, I had planned on having the Intel X25-M as boot and then having these VRs in RAID0 as a "workspace" for anything that needed fast write speeds, but I decided it's really overkill to spend this much just to save a few seconds on extracting TV shows and stuff :p
I'm sure it would be great as a Photoshop scratch disc or something similar, but I don't do any of that stuff.
I haven't tried booting from the RAID0 array, but I wouldn't expect it to be as fast as the SSD. It doesn't matter how high your throughput is, access times can never touch the SSD.

I'm still considering 2 x Intel X25-M RAID0, but again, it's overkill for sure. I'm probably going to get a second one, but I might put it in my MacBook Pro since regular 2.5 drives are a huge step down, and I hate how it feels "sluggish" somestimes. I'm pretty sure it's the disk at fault, since it has 4GB of RAM and a 2.4GHz Core 2 Duo.

Brahmzy said:
Thanks a ton Ender for these benches... good stuff. Takes time I know. These are the things we all wanna see but don't wanna take the time...
Click to expand...
I wanted to do some 4 x RAID0 on the Intel ICH10R, but most of the SATA ports are on the side, and I can't get to them very well with it in the case.
Yeah it definitely takes time, but lots of fun. I wish I could do stuff like this for a living.
Should have more time when I get through with school in a couple of years.
 
Thanks for the post! I had no idea software raid was that good. Hardware raid is a bit out of my price range and overkill, but I've several drives laying around and could do software raid for no additional cost on my next build and it looks like it would give a bit more speed:)
 
BikePilot said:
Thanks for the post! I had no idea software raid was that good. Hardware raid is a bit out of my price range and overkill, but I've several drives laying around and could do software raid for no additional cost on my next build and it looks like it would give a bit more speed:)
Click to expand...
Not to get off track here, but for a boot drive, a single VR would be better than most combinations of 7200 RPM drives for general usage due to the lower random access time. How old are the drives you have?
 
Great series of benchies, and I liked the HDD pr0n. +1 to those who were surprised by the weak write speeds by the Areca. +1 more for being wowed by the HighPoint R5 speeds.

The onboard ICH10R solution did pretty well, too, imho.

Again, nice work and thanks. You've clarified my future a bit with the ideas of buying an addin RAID card, if I pick up more 1500HLFS drives at some point.
 
dayam, some nice results.
what controller was the single drive attached to?
 
Any reason for not testing the 1220 out with 4 drives to give a head to head on the Highpoint? The 2x R0 results are only a single data point.
 
pissboy said:
Any reason for not testing the 1220 out with 4 drives to give a head to head on the Highpoint? The 2x R0 results are only a single data point.
Click to expand...
It's a Areca 1200, a 2 port card, not the Areca 1220.
What do you mean by single data point?
 
How about performance with a single drive? Say, ICH10R against the Areca 1200 with write caching enabled.
 
Hey Brahmzy, what kind of speeds do you get with your 3 VRs on Intel RAID?

Blue Fox said:
How about performance with a single drive? Say, ICH10R against the Areca 1200 with write caching enabled.
Click to expand...
I'll try to do that tomorrow.
 
EnderW said:
Hey Brahmzy, what kind of speeds do you get with your 3 VRs on Intel RAID?
Click to expand...

This is on my ICH9R (ASUS P5E). This is my boot drive slice - notice it's only a 60GB shortstroke, so the numbers are at their greatest, and of course ignore the ridiculous burst number as we all know that's what happens when Volume Write Cache is enabled on an ICHxR. For the life of me, I couldn't get that dip in the middle out after 3 runs.

FWIW
3vr300raid0hdtune01bt9.png
 
Brahmzy said:
This is on my ICH9R (ASUS P5E). This is my boot drive slice - notice it's only a 60GB shortstroke, so the numbers are at their greatest, and of course ignore the ridiculous burst number as we all know that's what happens when Volume Write Cache is enabled on an ICHxR. For the life of me, I couldn't get that dip in the middle out after 3 runs.

FWIW
3vr300raid0hdtune01bt9.png
Click to expand...
interesting
it's like the Intel controller has room to scale, but it can't get the max performance out of the drives it has
so even though the highpoint kicks ass, your money would better be spent on more disks
at least to a point
I wonder what kind of ceiling it has

if I were you, I'd sell off all those VRs, and get an Intel SSD for boot and then a couple of the small VRs or some of the 320GB platter 7200 RPM drives for when you need fast writes
 
EnderW said:
Not to get off track here, but for a boot drive, a single VR would be better than most combinations of 7200 RPM drives for general usage due to the lower random access time. How old are the drives you have?
Click to expand...

I have 3x WD5000YS drives and 2x WD4000YR's. A dedicated VR boot drive would be sweet, but I'm working on a budget build for a schoolwork/email/photo editing/storage machine and a VR is probably out of my budget. Planning on buying a P5Q-Pro mobo and the cheapest 45nm C2D I can find.
 
here was my single disk performance... YMMV

hdtune80.gif


this was on my machine at work... Q6700 some nVidia chipset cant remember..

80gbvrhome.JPG


this is on my machine at home... T7200 975x/ICH7
 
Old Hippie said:
I never said I was dissatisfied.

Here's your benchmark...



Nobody's gotten anywhere near close to that.

Got a secret?
Click to expand...

no i dont, if you want wednesday when i go back to work i can do the same bench on the same computer and see if it comes out the same... i did it multiple times and i get very similar benches...

unless microsoft installed some omg1337drivers for me in windows update or something, i have normal stock drivers from nvidia/microsoft

it was either a MSI P6NGM-L with a pentium D945...

or an ASUS P5WDG2-WS with a QX6700, those are my 2 workstations at work
 
FYI, the hardware Highpoints have native linux support for most models in kernels > 2.6.25.
 
no i dont, if you want wednesday when i go back to work i can do the same bench on the same computer and see if it comes out the same... i did it multiple times and i get very similar benches...
Click to expand...

It's pretty obvious those two tests are waaaay different and the second test is more of a representation of the results of others that purchased your drives are getting.

I'm no mealy mouth and would have contacted you if I wasn't satisfied with the purchase.

But, the discrepancy between the two tests is unusual.
 
Old Hippie said:
It's pretty obvious those two tests are waaaay different and the second test is more of a representation of the results of others that purchased your drives are getting.

I'm no mealy mouth and would have contacted you if I wasn't satisfied with the purchase.

But, the discrepancy between the two tests is unusual.
Click to expand...

maybe its something between the intel/nvidia chipsets? also the CPU's are way different (check out the CPU% usage on the benchmarks)

the motherboard i have at home is pretty old... its one of those aopens that uses a laptop CPU (between the cpu and the VR and an 8600gt my machine draws <120w from the wall)

like i said, i can try it again at work and see what the results are again, but why would i make them up? moreso how would i make them up? if it was a raid or something the name of the disk wouldent show up in the benchmark utility? and i mean, in theory i could have done that in photoshop, but trust me, i do NOT have the time nor desire to do silly stuff like that

i agree there IS a discrepancy obviously, i dont know why tho, my concern is you seem to think i tried to rip people off, and like i said, i didnt even post benchmarks until people started hounding me for them, and then i just threw a drive into a computer at work and ran ATTO and HD Tune really quick and thats what i got
 
Any chance you run run ATTO on a 4 drive RAID 0 (I'm interested in comparing it to my raptor array, however my 3ware card doesn't seem to like any Windows test other than ATTO)?

Thanks for the thorough list of benchies :D
 
it is definately the mobo/cpu differences

I have a Workstation board with duall lga771 sockets with 5000x chipset, and it gets better performance then my amd 780g chipset board hands down...that is just difference between desktop and WS/Server equipment.
 
my concern is you seem to think i tried to rip people off,
Click to expand...

Not at all.

I'm very satisfied with the drive and have heard no complaints from other purchasers.

It serves as my main OS drive and I've turned down more than a few purchase offers. :)

If it indeed is as you and nitrobass24 suspect, all I can say is "Wow".

I had no idea there wold be this dramatic of a difference in performance! :eek:

But, now that you've seen the difference yourself, you'll have to admit, even if we throw out the "valleys" of the graphs, it's pretty dramatic.

The next question is, now that we've seen the difference, how many of us will need to move to server MBs? :D
 
the areca 1220 is kind of old news/slow by today standards. I would love to see the high point compared to the ARC-1222 though!
 
houkouonchi said:
the areca 1220 is kind of old news/slow by today standards. I would love to see the high point compared to the ARC-1222 though!
Click to expand...
my tests are with the 1200, not 1220
I would also like to see 1220 and 1222 results
 
EnderW said:
It's a Areca 1200, a 2 port card, not the Areca 1220.
What do you mean by single data point?
Click to expand...

Oh sorry about that, I honestly didn't think anyone ever bothered to buy one of those. They're really not a RAID card IMO, not anymore than the onboard or any cheap 2 port card from best buy. :)
 
You must log in or register to reply here.
Back
Top