Overclock.net › Forums › Components › Hard Drives & Storage › RAID Controllers and Software › PERC 5/i RAID Card: Tips and Benchmarks
New Posts  All Forums:Forum Nav:

PERC 5/i RAID Card: Tips and Benchmarks - Page 508

post #5071 of 7150
Quote:
Originally Posted by LuciferX View Post
You have like 2 times my resutls! Can anyone explain me why my numbers are so low? Something is wrong?


Config: PERC6 - 3x750 WD RE2 GP RAID5 - WB_ARA_128k
Fileserver: AMD Sempron 2600+ s754 - 768MB RAM DDR - Nforce 405 - WinXP X86

Benchs:



Your running 750gb vs my 2tb so I have denser platters.
 
MacBook Pro 6.1
(16 items)
 
Server
(19 items)
 
CPUMotherboardGraphicsRAM
Core i7 Q720 Asus ATI Mobility Radeon HD 5870 1GB DDR5 12 GB DDR3 
Hard DriveOSMonitorKeyboard
Seagate Momentus XT 500GB 7200RPM & Seagate 750... Windows 7 Ultimate x64 17.3" LED LCD Factory 
PowerCaseMouseMouse Pad
Factory Factory Logitech G9X Mine 
CPUMotherboardGraphicsGraphics
Intel Core i5 MacBook Pro 6.1 17" Intel HD Graphics NVIDIA GeForce GT 330M 
RAMHard DriveOptical DriveCooling
4 GB DDR3 1067 MHz Hitachi HTS545050B9SA02 HL-DT-ST DVDRW GS23N Factory 
OSMonitorKeyboardPower
Mac OS X 10.6.8 17" LED Factory Factory 
CaseMouseMouse PadAudio
Factory Factory None Factory 
CPUMotherboardGraphicsRAM
Intel Core 2 Quad Q9550 Abit IP35 Pro PNY VCQ290NVS-PCIEX16-PB Quadro NVS 290 256MB 64-b G.Skill F2-8000CL5D-4GBPQ 
Hard DriveHard DriveHard DriveHard Drive
Dell Perc5/i Samsung HD204UI Seagate Constellation ES SAS Drive Model: ST310... Hitachi Deskstar 
Hard DriveHard DriveOptical DriveCooling
Western Digital WD1002FAEX Black Western Digital Black 2 x ASUS DRW-2014L1T Corsair H50 
OSMonitorKeyboardPower
Windows Server 2008 R2 Standard x64 ASUS VK222U Black 22" 2ms Logitech G15 (Original) - U.S. Dvorak Layout Enermax INFINITI EIN720AWT 720w 
CaseMouseMouse Pad
LIAN LI V SERIES PC-V2100A Silver Aluminum ATX ... Logitech G9x Steven's Computer Service 
  hide details  
Reply
 
MacBook Pro 6.1
(16 items)
 
Server
(19 items)
 
CPUMotherboardGraphicsRAM
Core i7 Q720 Asus ATI Mobility Radeon HD 5870 1GB DDR5 12 GB DDR3 
Hard DriveOSMonitorKeyboard
Seagate Momentus XT 500GB 7200RPM & Seagate 750... Windows 7 Ultimate x64 17.3" LED LCD Factory 
PowerCaseMouseMouse Pad
Factory Factory Logitech G9X Mine 
CPUMotherboardGraphicsGraphics
Intel Core i5 MacBook Pro 6.1 17" Intel HD Graphics NVIDIA GeForce GT 330M 
RAMHard DriveOptical DriveCooling
4 GB DDR3 1067 MHz Hitachi HTS545050B9SA02 HL-DT-ST DVDRW GS23N Factory 
OSMonitorKeyboardPower
Mac OS X 10.6.8 17" LED Factory Factory 
CaseMouseMouse PadAudio
Factory Factory None Factory 
CPUMotherboardGraphicsRAM
Intel Core 2 Quad Q9550 Abit IP35 Pro PNY VCQ290NVS-PCIEX16-PB Quadro NVS 290 256MB 64-b G.Skill F2-8000CL5D-4GBPQ 
Hard DriveHard DriveHard DriveHard Drive
Dell Perc5/i Samsung HD204UI Seagate Constellation ES SAS Drive Model: ST310... Hitachi Deskstar 
Hard DriveHard DriveOptical DriveCooling
Western Digital WD1002FAEX Black Western Digital Black 2 x ASUS DRW-2014L1T Corsair H50 
OSMonitorKeyboardPower
Windows Server 2008 R2 Standard x64 ASUS VK222U Black 22" 2ms Logitech G15 (Original) - U.S. Dvorak Layout Enermax INFINITI EIN720AWT 720w 
CaseMouseMouse Pad
LIAN LI V SERIES PC-V2100A Silver Aluminum ATX ... Logitech G9x Steven's Computer Service 
  hide details  
Reply
post #5072 of 7150
Quote:
Originally Posted by Jaco View Post
Actually, the previous tool did support ssl for gmail and that was what I've used until I've switched to 7... the problem with the previous is that is much more complicated for no reason from a normal user's perspective and also is not actually compatible with win7... Also the new tool is not xp compatible as xp and 7 have a different scheduler...

You can use whatever email provider you like (with ssl) and all you have to do is to fill the email parameters with the correct provider's values...
Hmm, I was never able to get it to work properly... Maybe because its not Win7 compatible...
    
CPUMotherboardGraphicsRAM
Core i7 970 ASUS P6T Deluxe eVGA GTX 680 6x4GB Kingston Hyper X DDR3-1600 
Hard DriveOptical DriveOSMonitor
Boot R0: 2x256GB Corsair SSD; Scratch R0: 3x750GB Samsung 22x LightScribe DVD+/-DL [SATA] Windows 7 x64 Ultimate Dell 2407FPWx2 
KeyboardPowerCaseMouse
Microsoft Sidewinder X6 Antec TruPower Q 1KW Custom Fabricated 18" cube (Mountain Mods UFO) Logitech G9 Laser 
Mouse Pad
Faux Woodgrain (My Desk) 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Core i7 970 ASUS P6T Deluxe eVGA GTX 680 6x4GB Kingston Hyper X DDR3-1600 
Hard DriveOptical DriveOSMonitor
Boot R0: 2x256GB Corsair SSD; Scratch R0: 3x750GB Samsung 22x LightScribe DVD+/-DL [SATA] Windows 7 x64 Ultimate Dell 2407FPWx2 
KeyboardPowerCaseMouse
Microsoft Sidewinder X6 Antec TruPower Q 1KW Custom Fabricated 18" cube (Mountain Mods UFO) Logitech G9 Laser 
Mouse Pad
Faux Woodgrain (My Desk) 
  hide details  
Reply
post #5073 of 7150
You 're right, the old tool isn't compatible with 7 (not without some modifications)... Anyway, so far I'm convinced that simple is better, therefore better use the new one, it's just sends the controller's status or your custom megacli command's result...

--------------

(edit: The following text is for those that are complaining that their benchmark results are not the expected)

Now, regarding the performance benchmarks I have seen around (not only here but also in other forums) most users doesn't really understand their setup, what they test and how to analyze the results... Maybe this topic is already discussed here, but it's repetition doesn't harm...

If the test case is a raid5 array of 5 disks for example, then if you want to measure the maximum performance of the array then you need to adjust your test block size... the reason is that the array topology actually defines the block size that must be used, this means that if you have 5 disks on a raid5 setup then there's no reason to test 4K reads when your chunk size is 128 or 64K because in every read only one disk actually throws data...

Therefore, you have to use a benchmark tool that can adjust the used block size during the test and such tool is hdtune for example... To sum up for optimum results you have to use the following equations

raid0: block size=(number of disks)*(chunk size)
raid5: block size=(number of disks-1)*(chunk size)

This is for a single read, but for a more realistic result you have to multiply the above result by 2 or 3 in order to perform 2 or 3 sequential reads...

So, for a 5 disks raid5 setup with 128K chunk size you'll have to use (5-1)*128K*2=1024K block size to the benchmark tool...

Considering the above then a new need arises and this is the need of a good analysis before set up a raid cluster, so you need to visualize it's everyday use... For example if your array is located on your local machine then you can obtain it's maximum performance and there's a meaning on trim the array parameters regarding the data that will be stored on the array... so if the majority of the files are large and you have 8 disks then you can use 64K for chunk size, or if you have 4 disks better choose 128K size... if your data are small files, then raid5 is of no use really so better setup a raid1 array...

On the other hand if your array is located on a remote location, i.e. a home file server then if you have a gigabit network then you're limited from the frame size, the tcp window size and the network's bandwidth of course... therefore use smaller chunk sizes...

Also have in mind that raid arrays (especially raid5) after some use they degrade their performance because the fragmentation is more 'chaotic' compared to a single disk and that has impact in the cluster's sequential reads/writes and the speed is limited to the speed of the disk which is more fragmented and has more latency in random read/writes...
Edited by Jaco - 3/4/11 at 12:36pm
post #5074 of 7150
Quote:
Originally Posted by eclsnowman View Post
I hope its not a problem with the cross mark cables. I am still waiting for mine. I already went through this issue with outletpc.com cables. I would be so mad.

Sent from my Droid using Tapatalk
I emailed the store and they said they are straight through cables even though I got crossover ones. Hopefully you get straight-through unlike me.
    
CPUMotherboardGraphicsRAM
Intel Core i7 5820K EVGA X99 Micro2 EVGA GTX 980 32GB DDR4-2400 
Hard DriveOSMonitorPower
Samsung 850 Pro Windows 10 x64 Pro Qnix 1440p EVGA 850W Gold 
Case
CaseLabs Mercury S5 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Core i7 5820K EVGA X99 Micro2 EVGA GTX 980 32GB DDR4-2400 
Hard DriveOSMonitorPower
Samsung 850 Pro Windows 10 x64 Pro Qnix 1440p EVGA 850W Gold 
Case
CaseLabs Mercury S5 
  hide details  
Reply
post #5075 of 7150
Here's hoping they work. I should get them tomorrow


Sent from my Droid using Tapatalk
post #5076 of 7150
Nope. They sent (4x7 Sata)host-to-(32pin SAS)backplane cables instead of the advertised "32-pin SFF-8484 to 4 x 7-pin internal Serial Attached SCSI(SAS) cable is idea for connecting SAS controllers to individual SAS/SATA hard drives"

Here are some pictures of what they sent:





And this is what it was supposed to look like from the CrossMark Website





I am so Pissed I could Scream.
post #5077 of 7150
Quote:
Originally Posted by eclsnowman View Post
Nope. They sent (4x7 Sata)host-to-(32pin SAS)backplane cables instead of the advertised "32-pin SFF-8484 to 4 x 7-pin internal Serial Attached SCSI(SAS) cable is idea for connecting SAS controllers to individual SAS/SATA hard drives"

Here are some pictures of what they sent:





And this is what it was supposed to look like from the CrossMark Website





I am so Pissed I could Scream.
Yep, those are the EXACT cables I got.
    
CPUMotherboardGraphicsRAM
Intel Core i7 5820K EVGA X99 Micro2 EVGA GTX 980 32GB DDR4-2400 
Hard DriveOSMonitorPower
Samsung 850 Pro Windows 10 x64 Pro Qnix 1440p EVGA 850W Gold 
Case
CaseLabs Mercury S5 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Core i7 5820K EVGA X99 Micro2 EVGA GTX 980 32GB DDR4-2400 
Hard DriveOSMonitorPower
Samsung 850 Pro Windows 10 x64 Pro Qnix 1440p EVGA 850W Gold 
Case
CaseLabs Mercury S5 
  hide details  
Reply
post #5078 of 7150
Here is a transcript of the email I sent cross-mark:

Quote:
Originally Posted by me
Hi, I recently ordered these cables:

http://www.cross-mark.com/50cm-seria...ble-p-842.html

and I was wondering whether they are \\"straight-through\\" or \\"crossover\\" type cables. I trying to connect SATA disks to a RAID controller and they aren\\'t working. I thought maybe I got the wrong ones since they are also red instead of black like in the picture.

Thanks,
Erik
Quote:
Originally Posted by them
Hi Erik,

It is straight-throught cable.

Thanks
Cliff Lee
Sales Manager - Cross Mark Co.
Website: www.Cross-Mark.com
Quote:
Originally Posted by me
Hi, I just checked the cables using a multimeter and I can confirm that they are actually crossover/backplane cables and even have a sticker labeling them so.

This diagram of a straight-through cable is what I used for reference

http://dl.dropbox.com/u/2865529/SA-84S7P-WD.pdf
Awaiting their reply
    
CPUMotherboardGraphicsRAM
Intel Core i7 5820K EVGA X99 Micro2 EVGA GTX 980 32GB DDR4-2400 
Hard DriveOSMonitorPower
Samsung 850 Pro Windows 10 x64 Pro Qnix 1440p EVGA 850W Gold 
Case
CaseLabs Mercury S5 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Core i7 5820K EVGA X99 Micro2 EVGA GTX 980 32GB DDR4-2400 
Hard DriveOSMonitorPower
Samsung 850 Pro Windows 10 x64 Pro Qnix 1440p EVGA 850W Gold 
Case
CaseLabs Mercury S5 
  hide details  
Reply
post #5079 of 7150
Anyone know where I can get good, cheap, and mainly correct cables? This is the second time i got cables because a vendor put the wrong description. First OutletPC them Cross-Mark. I am so freaking bummed.
post #5080 of 7150
It is really weird because I bought the cables from them and they sent me exactly what they show on their website.
Sorry for you guys. Fill their mailboxes with complaints.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: RAID Controllers and Software
Overclock.net › Forums › Components › Hard Drives & Storage › RAID Controllers and Software › PERC 5/i RAID Card: Tips and Benchmarks