RAID slower than single drives!?!
MaximumPC's most recent issue has some interesting benchmarks; 2 & 4 drive striped RAID arrays running SLOWER than single drives. I just got a FastTrack 100 TX2 RAID, and according to Nero's (5.5) HD benchmark, the RAID sure enough is slower than the single drive. My single drive in an IBM 60GB 7200RPM, the RAID has 3 Maxtor 60GB's 7200RPM D740X drives. Benchmark numbers where the same with 2-4 WD 40GB 7200's, and 2-4 of the maxtors (yeah, I've got a bunch of hard drives ). Write speed on the RAID may be a little faster, but I want read speed. Am I better off ditching the RAID and putting back in my Promise Ultra100 card? (I may do that no matter what...when ever the RAID array is being accessed, the system will freeze for about 1 second every 30-60 seconds. Very annoying.) Anyone know why the RAID is slower now? I remember seeing some of MPC's own benchmarks awhile ago showing how much faster it was.
You tried both RAID 0 and RAID 5 arrays?
A couple of things to check:
1. Is ATA100 enabled.
2. Do you have the drive's installed on separate ports (primary, secundary).
3. Have you got the latest chipset and RAID controller drivers.
RAID 5 is useually slower then a single drive and RAID 0 is only faster in reading/writing larger files.
My case has so many fans that it hovers above the ground .
I was using RAID 0, striping across 2, 3, and 4 drives. Ultra100 was enabled (UDMA5), with all the latest drivers. Write speeds WERE up, but read speed was down (just like Maximum PC's benchmarks showed). The drives jumber setting for two drives where master/master, for 3, master/slave/slave (master/slave/master would not work at all, the RAID card didn't see any drives hooked up at all).
My guess is that IDE just sucks at accessing multiple drives at the same time, thus slowing down somewhat. Maybe the FastTrack TX4 card would be faster (since all drives hook up to it as master, 1 per channel) ? Maybe it's just an issue with the FastTrack. If I have the time, I'll put 2, 3 and 4 of the drives onto my Athlon systems on-board RAID and re-run the benchmarks. Still, RAID used to be faster at reading, wasn't it??
The RAID Controller is a Promise FastTrack TX2 Ultra100 card, with Ultra DMA 5 enabled. I always thought IDE raid was faster, but both my benchmarks and Maximum PC's show it to be slower for reading, faster for writing. Maybe it's just something with the Promise card...that's the one MPC used, too. I don't have the RAID installed anymore (I swapped it out for a normal Promise Ultra 100 card), and now get >40MG/sec on all 6 drives (4 maxtor 60GB, one IBM 60, one WD 40 GB, all 7200rpm). Now if I just had room for my other 3 WD drives...
This is defenitly the last time I get an IDE RAID setup. I will for sure go with SCSI in my next system, but probably not RAID, since the cards are about $1000. Oh well, a 15,000 RPM drive is nice