| System Hardware | |
|---|---|
| Hardware | Details |
| CPU | Intel Core i7-920 (45 nm, 2.66 GHz, 8MB L3 Cache) |
| Motherboard (Socket 1366) | Supermicro X8SAX; Revision: 1.1; Chipset: Intel X58 + ICH10R; BIOS: 1.0B |
| RAM | 3GB DDR3-1333 Corsair CM3X1024-1333C9DHX |
| System HDD | Seagate NL35 400GB ST3400832NS; 7,200 RPM, SATA 1.5Gb/s, 8MB Cache |
| Test HDDs | 3.5” 15,000 RPM (Fujitsu MBA3147RC) 2.5” 10,000 RPM (Toshiba MBF2600RC) |
| Power Supply | OCZ EliteXstream 800W OCZ800EXS-EU |
| Benchmarks | |
| Performance Measurements | h2benchw 3.12 PCMark Vantage 1.0 |
| I/O Performance | Iometer 2006.07.27 Fileserver-Benchmark Webserver-Benchmark Database-Benchmark Workstation-Benchmark Streaming Reads Streaming Writes |
| System Software & Drivers | |
| Driver | Details |
| Operating System | Windows Vista Ultimate SP1 |
| Intel Chipset | Chipset Installation Utility 9.1.0.1007 |
| AMD Graphics | Radeon 8.12 |
| Intel Matrix Storage | 8.7.0.1007 |
If the 3.5 inch drives have twice the number of platters as the 2.5 inch drives, shouldn't their performance be similar to that of two 2.5 inch drives in RAID? Or does the performance of multiple platters not scale that way?
Nice article.
"Two 2.5” SAS drives in RAID will outperform one fast 3.5” SAS drive in every workload."
I would like to see proof of this, as I don't believe the "in every workload" bit. RAID-0 and -5 don't always deliver performance as promised. As far as I know, it depends on the file sizes, among other factors. Applications that use thousands of little files (web servers come to mind) can't always get a good performance boost from data striping, as the overhead required to read/write many small files outweighs the speed gains achieved by spreading the load across multiple drives. In this case, one faster 3.5" drive might outperform two slower 2.5" drives.
SO ... We ALL DO have 5.25" drive bays as a common option ... Can you guess what I am gonna say next ??
Perhaps mechanical drive manufactureres should revisit the 5.25" form-factor for super-perpendicular, ultra-dense, high-performance desktop apps, such as video editing and graphics (and other HD+ media content apps).
UltraRaptor 5"x10K dual-platter short-stroked.
= or not =
lol, Alvin you beat me to it
but thats a whole other article.
Ive always wondered what if the latest hard drive technology was applied to the old 5.25 standard. I can easily see 300 MB/s transfer speeds and at least 1 TB platters with enough room for maybe 6 or so in one drive. Although realistically the price per gigabyte would be unmatched and the price per performance and efficiency would be hideously low compared to the trend I see on page 8... unless you add another actuator arm
try spinning a stack of 5.25" disks at 10k rpm.
"2.5” and 3.5” enterprise drives are based on the same platter diameter"
LMAO ...another drive "expert" who doesn't know ____ from shinola.
@Alvin Smith: I'm more interrested in high-density low-RPM 5,25" drives for desktops and fileservers. I would easily buy sine 10TB 5,25" 3600RPM (or 1800RPM?) drives for RAID-5 to store huge ammounts of media and other files. Given the price pr GB was right ofc. 50MB/s sequential read/write pr drive would be sufficient for mass storage.
" Two 2.5” SAS drives in RAID will outperform one fast 3.5” SAS drive in every workload"
...and what about a drive failure in your striped array? You should consider the whole, bigger picture, not just speed and capacity.
WHICH ONE IS BATTER 2.5” SAS OR 3.5” SAS