Originally Posted by
cadtag
>> 2 x 250 GB Samsung EVO SSD in RAID 0 (OS)
Are there any benchmarks that show a significant performance improvement for Raid 0 on SSD? Seems that approach would simply double the chances of a catastrophic data loss - since a failure of either SSD means the loss of all the data.
Yes... The RAID 5 array [#2] was slower than a single SSD by itself [#1], and the RAID 0 array [#3] was faster than both, hence the configuration:
[#1]
Code:
-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]
Sequential Read : 272.269 MB/s
Sequential Write : 258.993 MB/s
Random Read 512KB : 259.273 MB/s
Random Write 512KB : 250.074 MB/s
Random Read 4KB (QD=1) : 31.421 MB/s [ 7671.1 IOPS]
Random Write 4KB (QD=1) : 46.671 MB/s [ 11394.3 IOPS]
Random Read 4KB (QD=32) : 183.238 MB/s [ 44735.8 IOPS]
Random Write 4KB (QD=32) : 157.587 MB/s [ 38473.4 IOPS]
Test : 100 MB [F: 0.1% (0.1/232.9 GB)] (x5)
Date : 2014/11/07 8:42:56
OS : Windows 8.1 Pro [6.3 Build 9600] (x64)
[#2]
Code:
-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]
Sequential Read : 459.007 MB/s
Sequential Write : 252.466 MB/s
Random Read 512KB : 280.713 MB/s
Random Write 512KB : 240.411 MB/s
Random Read 4KB (QD=1) : 26.707 MB/s [ 6520.3 IOPS]
Random Write 4KB (QD=1) : 35.198 MB/s [ 8593.3 IOPS]
Random Read 4KB (QD=32) : 227.423 MB/s [ 55523.3 IOPS]
Random Write 4KB (QD=32) : 157.275 MB/s [ 38397.2 IOPS]
Test : 100 MB [F: 0.1% (0.1/221.2 GB)] (x5)
Date : 2014/11/07 8:59:33
OS : Windows 8.1 Pro [6.3 Build 9600] (x64)
[#3]
Code:
-----------------------------------------------------------------------
CrystalDiskMark 3.0.3 x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]
Sequential Read : 513.337 MB/s
Sequential Write : 483.692 MB/s
Random Read 512KB : 524.201 MB/s
Random Write 512KB : 446.906 MB/s
Random Read 4KB (QD=1) : 28.608 MB/s [ 6984.3 IOPS]
Random Write 4KB (QD=1) : 36.863 MB/s [ 8999.8 IOPS]
Random Read 4KB (QD=32) : 228.869 MB/s [ 55876.3 IOPS]
Random Write 4KB (QD=32) : 254.977 MB/s [ 62250.2 IOPS]
Test : 100 MB [F: 0.0% (0.1/442.5 GB)] (x5)
Date : 2014/11/07 7:34:16
OS : Windows 8.1 Pro [6.3 Build 9600] (x64)
As my workstations are backed up to supplementary internal drives (primary), and to network (secondary), data loss is not really of concern.
Nobody here really keeps anything of importance on their workstation anyway with exception of myself (CAD\IT Admin), and the owners. I have a StarTech USB 3.0 device that supports HDD/SSD that I use to 'clone' drive images for a physical drive swap in the event of a catastrophic failure. If we don't have time for replication, the network SystemImage's .VHDx is mapped to a Hyper-V Virtual Machine, which can be used with Windows Server 2012 R2 RDS, or RDP from any other workstation in the domain.
Cheers