Slashdot Mirror
High-End, High-Capacity SATA-150 Roundup
Maxtorn writes This review is published to cover a "300GB Maxtor drive, but provides a roundup covering a few high end, high capacity drives from Maxtor, Seagate, and Hitachi. Synthetic / real world performance, thermal results, and noise output are all covered on drives ranging from 200-500GB in capacity and with 8-16MB of cache memory. A solid reference for those shopping for a new drive."
Now I like a drive I can use in more than way. I can use this on my current ATA aetup, and if I upgrade motherboards, I can just switch cables and move on.
You are not the customer.
Pros:
- Fastest SATA-150 drive tested to date
No issue with speed, it's good.- Several capacities available, with 300GB being the highest
Not unexpected from and industry leader.- Quiet operation
Weighty consideration for the home or office, a brace of noisy drives is unwelcome while trying to watch video or listen to music on the computer.- Supports Native Command Queuing
Fine.- Excellent value, only 48 cents per GB
Really this is a minor concern, unless you're building a storage rack and only care bang/buck. If I want cool and quiet, I'll pay extra for it.- 16 MB of cache memory provides a nice performance boost
The bottleneck isn't likely to be your cache it's your MB and OS, but always nice to have more cache.Cons:
- Runs a bit warmer than other drives
Might warrant an extra fan if running a brace or more, potentially negating and quiet running. I've got an old Quantum drive you could fry an egg on and the heat effectively is killing the bearing lubricant.- Three year warranty is good, but not the best
Really, what good is a warranty, other than it's DOA? Does anyone do backups anymore? How's that MTBF? A warranty is the least of my concerns if my drive dies in the first year.A feeling of having made the same mistake before: Deja Foobar
After you click through the first two ad-cluttered pages, you start to see some results. They're presented in a single bar graph with dark shaded gradients.
The graph uses the same X axis to compare three totally different quantities: CPU percentage, access time in milliseconds, and bandwidth in MB/sec. As a bonus, note that smaller values for CPU % and access time are good, but larger values of bandwidth are good.
Edward Tufte, where are you?
And I know nobody is impressed by hard drive space anymore, but 300GB for only $139 truly does boggle the mind. We're at the $500 = one Terabyte point. That's nuts.
If you wanna get rich, you know that payback is a bitch
Offtopic, but I've never gotten why everyone is so down on Maxtor drives. Maybe it's just me, but Maxtors have been the most reliable drives in my experience. I just got my first two Seagate drives about 3 months ago, so I can't claim to make a good judgement about them yet, but they're doing better than I would expect out of a Western Digital. In the past decade I've had at least 5 WD drives fail on me. Someone once told me that I had to be abusing the case that 3 of these WDs failed in, but in that same case I had a 540MB Maxtor drive that was running fine (and seeing more use than the WDs). Actually, that 540MB Maxtor is STILL running just fine to this day (over 10 years after I got it). I'll be buying Maxtor (and probably Seagate) drives for the forseeable future and I've sworn off WD.
I don't really see the advantage of a large cache on the disk. I would much rather have the cache on a 2GB/s connection to the CPU than a 150MB/s one. Bung an extra few hundred MBs of main memory, and you are likely to see more of an improvement than adding a small amount of RAM to the disk - particularly since the main RAM can be used for other things when you are not using it as cache.
I am TheRaven on Soylent News
I own an older 9GB IBM Ultra Wide SCSI hard drive. The drive boasts an access time below 5ms. As anyone serious about performance knows, a hard drive's access time is an enormous factor. Read and write performance tends to increase exponentially with lower access times (non-sequential operations). The difference between 9ms and 5ms is deceptively large. I haven't come across any hard drive as fast as my IBM for nearly a decade. To witness it in person, specs aside, is a miracle to behold. I shit you not.
When I sometimes come across articles about SATA or SATA peripherals, I keep reading about the "performance" that SATA brings to the table. Not much can be said about the differences between IDE and SATA regarding bandwidth, they are nearly the same. I still don't know what all the hype is about. Modern hard drives don't even come close to saturating an ATA-133 bus. Burst (cache) speeds don't count. Without RAID, you'll never hit the upper limits anyway. Modern hard drives don't even have access times to justify a lower latency. Sure SATA scales better, but who cares? For the time being, SATA is ATA with new clothes.
My love of SCSI aside, IDE is almost always faster in terms of raw performance. SCSI shines in RAID configurations or with multiple devices (five or more). If all you need is one drive on a Linux server, IDE wins hands down. IDE is also free from the nightmares of SCSI termination and ninety+ connector types. My attraction to SCSI comes from the availability of high performance hard drives. No self respecting manufacturer would release a high-speed drive to the budget market. In the 90's, the best drives were exclusively SCSI and they still are.
When SATA was announced, I hoped that it would offer the advantages of SCSI with the simplicity and cost of IDE... a replacement to both. How wrong I was. Sure, the bandwidth is higher and the connectors are much more sexy. I hate ribbon cables and 68-pin connectors just like anyone else. Even the technology behind the interface is sound, but the manufacturers haven't taken it seriously. The best drives are still exclusive to SCSI. The best servers don't have SATA. SATA is neither the absolute replacement to IDE nor the successor to SCSI. It's been positioned as some bastard to fill the gap between the two.
Now that digital photography, music, and video have finally become commonplace, the focus has been placed on increased storage capacity. Performance has taken a back seat and will for some time. There has always been a trade-off between the two, they are mutually exclusive. SATA solves this in no way. Low-end consumer hard drives that would normally be released with an ATA interface are simply offering SATA if they want to be seen as "high performance." Even the new Maxtor DiamondMax 300GB drive, is offered in a comparable ATA-133 model. Hitachi sells a drive that offers SATA-300, not because it can physically transfer data that fast, but because it sounds good.
We had this problem in the SCSI world too. There was SCSI, Fast SCSI, Wide SCSI, Ultra SCSI, Ultra Wide SCSI, Ultra2 SCSI, Ultra2 Wide SCSI, Ultra3 SCSI / 160, Ultra-320 SCSI, and Ultra-640 / Fast-320. The thing is, they are backward and forward compatible. The oldest drives work with the newest controllers and the oldest controllers work with the newest drives. The bus speed is very useful if you use RAID, made more feasible since the wide variations of SCSI support up to 15 devices per controller. SCSI advances have been more about performance and less about marketing (UDMA-33, UDMA-66, UDMA-100, UDMA-133).
Admirers of SATA should shut up already. It's only a nominal increase over IDE in performance. The hard drives you can buy are exactly the same as one would expect of the low-end / IDE market. Even though SATA may be technologically superior in every way to IDE, what's happening is no better than putting a SCSI interface on a slow IDE drive. I've learned that you can use SATA drives on SCSI controllers. Why? That's exactly the same stupidity I'm referring to... the combination of the extremely budget conscience with the high-end.
SATA will never be "high performance" unless SATA drives become "high performance."