They should be, but unfortunately many neglect to do them.They should be doing backups anyway.
They should be, but unfortunately many neglect to do them.They should be doing backups anyway.
I have a max'd out 2013 iMac 27 with a 3TB FD, which I use for professional video editing. It works fine. See attached performance benchmarks. I also have a 2013 MacBook air with 512GB SSD -- it has higher disk write performance than the iMac.
However the issue is *not* which has the highest benchmarks, but does it make an obvious difference in your workflow. And if there's a difference is it worth the cost in terms of smaller storage size, using slower external drives, etc.
If I was getting the iMac today instead of last year, I'd probably get a 512GB SSD or 1TB SSD, since I have a 8TB Thunderbolt Pegasus R4 for most of my data.
But I have no problems with the 3TB FD from a performance standpoint.
There are several important factors besides performance:
(1) You must have room for all your data, both now and future. If you work with large data sets, moving that off to slower external hard drives can be tedious and error-prone.
(2) You must be able to back up your data, both internal and external. The data on internal SSD, internal FD *and* external HDD must all be backed up. IOW if you can only afford a 256GB SSD and you have 500GB data, this forces you to buy external (slower) HDDs for this, then *that* must itself be backed up even though it resides on external HDD. The budget for a given computer configuration must include the backup hard drives, whether your data is internal or external.
If your workflow uses smaller data sets or it can be infrequently moved on and off the smaller SSD, then SSD would present less operational complexity.
BTW the Anandtech test of FD vs Samsung SSD may be flawed, because they forced the FD to 80% full vs an unstated capacity factor on the SSD. SSD write performance slows down dramatically as the drive fills up.
I'm hoping when I get my next iMac in 5 or 6 years time, that SSD is either a standard item (i.e. iMac's no longer use the traditional hard drive)
I wouldn't be surprised if that happens in 6 months.
I suspect within a year or so all traditional hard drives will have been removed from apple products. They are far and away the most unreliable component in a PC.
Does the SSD portion "fill up" and slow down the performance over time with regular use?
I will definitely have a time machine (it's own hard drive) plus a spare hard drive with a second copy of client files and I also backup both the computer and external with remote back up with Backblaze in case of fire or theft.
It doesn't matter how fast the disk is unless it produces noticeably faster real-world performance. The goal isn't to win the "Benchmark Olympics".Still nowhere near my pure SSD (256GB, with 40GB free).
It doesn't matter how fast the disk is unless it produces noticeably faster real-world performance. The goal isn't to win the "Benchmark Olympics".
In the above case your SSD was roughly equal in read performance to my FD, and the write performance about 2x. Here's the problem: on most systems reads outnumber writes by 10 to 1 or more. I invite anyone who doubts this to start Activity Monitor and compare global reads vs writes at the bottom of the screen. So in this case, SSD write performance is much better than FD but:
(1) FD write performance is not slow
(2) Overall what counts is read performance
Improving something you infrequently need doesn't help real-world practicality. If there is a cost to that improvement (money or diminished storage space), you've traded a real-world improvement for one which may not produce dramatic differences.
Most apps have a mix of bottlenecks. Last night I imported 70 GB of video material into FCP X. The average I/O rate was about 50-100 megabytes/sec, and that was from my 8TB Pegasus R4 Thunderbolt RAID array. An *infinitely* fast SSD would not have been much faster, because the import process was limited by multi-threaded CPU factors.
The important disk performance metric is *not* I/O rate on a benchmark, it's disk queue length on your real-world workload. If the disk queue (e.g, number of pending I/O operations) is not excessive, then making the disk faster won't help.
Disk queue length can be monitored using the Dtrace facility from terminal, e.g, "sudo impending". More info: http://dtrace.org/blogs/brendan/2011/10/10/top-10-dtrace-scripts-for-mac-os-x/
If I was getting an iMac today I'd probably get a SSD but not because it would transform system performance. It would be a little faster in certain real-world situations. If prospective iMac purchasers can afford the SSD size which fulfills their current and future data storage needs, then by all means get it. But if a limited budget causes them to scrimp on SSD size and then purchase slow external hard drives to move data around, they may regret the SSD.
No - the most important reason for getting an SSD is so that you have some really cool benchmarks to screenshot and post at every opportunity![]()
So how much do these SDD bench marks really play out in day-to-day use. Are we talking about a 2 second difference between the SDD and Fusion?
When I read reviews about SDD, it feels like I would be a fool to get a fusion because, fusion performance won't last, the speed slows down over time, or it's a temporary solution while Apple waits for the price of SDD to come down, so fusion won't stick around, or that SDD blows fusion out of the water performance wise. Yet for those that have the fusion, they seem pretty happy with it. It's hard for me to determine what is the best move. Being that I will be getting an Apple discount (15%), should I not pass up the 1 TB SDD? Or would that money be better spent elsewhere?
Arrrrggggg I can't decide!
So how much do these SDD bench marks really play out in day-to-day use. Are we talking about a 2 second difference between the SDD and Fusion?
When I read reviews about SDD, it feels like I would be a fool to get a fusion because, fusion performance won't last, the speed slows down over time, or it's a temporary solution while Apple waits for the price of SDD to come down, so fusion won't stick around, or that SDD blows fusion out of the water performance wise. Yet for those that have the fusion, they seem pretty happy with it. It's hard for me to determine what is the best move. Being that I will be getting an Apple discount (15%), should I not pass up the 1 TB SDD? Or would that money be better spent elsewhere?
Arrrrggggg I can't decide!
So how much do these SDD bench marks really play out in day-to-day use. Are we talking about a 2 second difference between the SDD and Fusion?
When I read reviews about SDD, it feels like I would be a fool to get a fusion because, fusion performance won't last, the speed slows down over time, or it's a temporary solution while Apple waits for the price of SDD to come down, so fusion won't stick around, or that SDD blows fusion out of the water performance wise. Yet for those that have the fusion, they seem pretty happy with it. It's hard for me to determine what is the best move. Being that I will be getting an Apple discount (15%), should I not pass up the 1 TB SDD? Or would that money be better spent elsewhere?
Arrrrggggg I can't decide!
No - the most important reason for getting an SSD is so that you have some really cool benchmarks to screenshot and post at every opportunity![]()
Seen that screenshot in about 6 different threads![]()
It doesn't matter how fast the disk is unless it produces noticeably faster real-world performance. The goal isn't to win the "Benchmark Olympics".
In the above case your SSD was roughly equal in read performance to my FD, and the write performance about 2x. Here's the problem: on most systems reads outnumber writes by 10 to 1 or more. I invite anyone who doubts this to start Activity Monitor and compare global reads vs writes at the bottom of the screen. So in this case, SSD write performance is much better than FD but:
(1) FD write performance is not slow
(2) Overall what counts is read performance
Improving something you infrequently need doesn't help real-world practicality. If there is a cost to that improvement (money or diminished storage space), you've traded a real-world improvement for one which may not produce dramatic differences.
Most apps have a mix of bottlenecks. Last night I imported 70 GB of video material into FCP X. The average I/O rate was about 50-100 megabytes/sec, and that was from my 8TB Pegasus R4 Thunderbolt RAID array. An *infinitely* fast SSD would not have been much faster, because the import process was limited by multi-threaded CPU factors.
The important disk performance metric is *not* I/O rate on a benchmark, it's disk queue length on your real-world workload. If the disk queue (e.g, number of pending I/O operations) is not excessive, then making the disk faster won't help.
Disk queue length can be monitored using the Dtrace facility from terminal, e.g, "sudo iopending". More info: http://dtrace.org/blogs/brendan/2011/10/10/top-10-dtrace-scripts-for-mac-os-x/
If I was getting an iMac today I'd probably get a SSD but not because it would transform system performance. It would be a little faster in certain real-world situations. If prospective iMac purchasers can afford the SSD size which fulfills their current and future data storage needs, then by all means get it. But if a limited budget causes them to scrimp on SSD size and then purchase slow external hard drives to move data around, they may regret the SSD.
As I've said earlier in this thread, I regret Fusion.
I think the real problem is my workflow; I'm a developer and I'm guessing the constant writing/rewriting of files during builds is too much for the little 128GB flash portion.