MOST RECENT SPEED TESTS

SPEED TESTS by Category

FireWire 800 Speed Mysteries Solved: Dock versus Enclosure? 3.5 versus 2.5 inch RAID PowerBook G4 versus Power Mac G5

Originally posted 03/26/04 by rob-ART morgan, mad scientist. QUESTION 1: Does Wiebetech's (FireWire 800/USB2) ComboDock perform just as well as a standard FireWire 800 enclosure? ANSWER: The Wiebetech ComboDocks are just as fast as a standard FireWire 800 enclosure. Why would I want to use a ComboDock instead of an enclosure? Because they provide a smaller form factor for a FireWire 800 drive system and make switching drives a snap... something I'm always doing. QUESTION 2: Some of you were excited about the idea of attaching a pair of Hitachi 7K60s to the FireWire 800 port of a PowerBook, creating a fast striped array for your mobile digital video workstation. If a striped pair of 2.5 inch drives go fast (70+MB/s) when connected to a PowerBook G4, will a pair of 3.5 inch drives go even faster? We used two Wiebetech ComboDocks connected to two Maxtor DiamondMax Plus 9s to test this scenario. ANSWER: Yes and no. LEGEND: int=built-in FW800, pc=cardbus FW800, 1chan=single controller, 2chan=separate controller for each drive In all cases, the test computer was an Aluminum G4/1.33GHz PowerBook The 3.5 inch 7200 drives we tested were faster than the "7K60" 2.5 inch 7200rpm drives when doing sustained READs -- especially when dual channels (one controller per drive) are used. But, to our surprise, 3.5 inch drives were NOT faster when doing sustained WRITEs. That's true whether you connect them to a PowerBook or Power Mac. The 3.5 inch drives have the advantage of greater capacity (up to 250GB currently). The 2.5 inch drives have the advantage that they are lighter, more compact, and run off of FW800 bus power. QUESTION 3: In a recent test, a PowerBook G4 "embarrassed" a Power G5 when running a FireWire 800 RAID set. Will the same sustained WRITE "gap" occur with the bigger 3.5 inch drives? Again, we used two ComboDocks to test this. ANSWER: Yes. LEGEND: int=built-in FW800, pc=cardbus FW800, pci=FW800 pci card, 1chan=single controller, 2chan=separate controller for each drive PM G5=G5/2.0MP Power Mac, PM G4=G4/1.42MP Power Mac, PB G4=G4/1.33GHz PowerBook The G4 Power Mac and G4 PowerBook posted READ times on par with the G5 Power Mac. Yet they both posted sustained WRITE times significantly FASTER than the "World's Fastest Personal Computer." Huh? Yes, even an older Titanium G4/800 PowerBook with CardBus FW800 and single channel FW800 striped RAID did sustained WRITES faster than a G5. If you need a RAID set for your G5 Power Mac, you better stick with SATA, PATA, or SCSI. QUESTION 4: What if we put four 3.5 inch drives on two channels? ANSWER: No difference. LEGEND: int=built-in FW800, pc=cardbus FW800, pci=FW800 pci card, 2chan=two controllers on four 3.5 inch drives PM G5=G5/2.0MP Power Mac, PM G4=G4/1.42MP Power Mac, PB G4=G4/1.33GHz PowerBook The G5 edges the others in READ speed, but lags far behind on WRITE speed. Scandalous! If you know anyone in the FireWire department at Apple or participate in Apple discussion groups, please make some "noise" about this G5 deficiency. If you don't, this "feature" might get passed on to the Rev 2 G5 Power Macs!   SOURCES FOR TEST PRODUCTS The ComboDocks were provided courtesy of WiebeTech. You can purchase direct from their site or from one of their resellers. You might also want to check out their Forensic ComboDock. If you you want a case that holds two 3.5 inch drives, check out their BayDock 800. The Maxtor DiamondMax Plus 9 3.5 inch 7200rpm drives we used in testing were provided courtesy of Maxtor. The Hitachi Travelstar 7K60 2.5 inch 7200rpm drives we used were provided courtesy of Trans International. The FireWire 800 Cardbus PCMCIA Card and FireWire 800 PCI card used was provided courtesy of LaCie. We were so shocked by the results, we tested on three different G4 PowerBooks and two different G5 Power Macs. Same results. The PowerBook G4/1.25 and Power Mac G5/1.6 were tested at CryWolf. The PowerBook G4/1.33 was tested at local Apple Retail Store. The G4/1.42MP Power Mac, G5/2.0MP Power Mac, and G4/800 PowerBook were tested in the Bare Feats lab. We used Intech's SpeedTools Test Suite -- particularly QuickBench X -- for capturing sustained read/write speeds. We specified Extended Test and check the "20-100MB" box. Has Bare Feats helped you? How about helping Bare Feats?