Speed on retina iMac seems slow ...?

[eroxx]

I turned on my new retina iMac and immediately restored from time machine backup using the migration assistant.

After it was done (14 hours later - 2.6 TB backup) I turned on the computer and ran blackmagic disk speed.

I was getting 200ish write and 500ish read.

The iMac is souped up - 4.0 cpu and upgraded gpu option. Also with the 3TB fusion.

Would love your thoughts thanks!!!

 

[carlob]

1Tb SSD

 

[eroxx]

Curious how different the ssd vs fusion is. Thanks for posting!

 

[yjchua95]

Curious how different the ssd vs fusion is. Thanks for posting!

A Fusion is never as fast as a pure SSD setup.

Even if all reads/writes were done on the 128GB SSD alone, it's slower, because smaller SSDs tend to be slower.

 

[eroxx]

But do those numbers seem right to you that I posted?

 

[rainking]

1tb fusion drive.

 

[eroxx]

Thanks I feel a bit better. I'm a little lower than that but my 3tb drive is 80 percent full - does that make a difference?

 

[rainking]

Thanks I feel a bit better. I'm a little lower than that but my 3tb drive is 80 percent full - does that make a difference?

Yes it does.

 

[eroxx]

thanks so much. I was hoping it would be better to help justify the purchase!

 

[hfg]

thanks so much. I was hoping it would be better to help justify the purchase!

What does BlackMagic DiskSpeedTest show if you set it for a 1GB test file size? This should keep the test on the SSD so you aren't seeing the hard disk speeds.

 

[joema2]

Thanks I feel a bit better. I'm a little lower than that but my 3tb drive is 80 percent full - does that make a difference?

Yes, all HDDs inc'l FD will slow down as they fill up. This is an advantage of SSD which retains full performance to about 97% full.

Recommendations for how much free space to maintain on HDDs vary from 20% free up to 50% free for video editing.

If you're at 20% free space you need to think about moving some data to an external disk. Just don't use a slow 5400 rpm bus-powered USB portable drive. The fastest bus-powered 7200 rpm portable drive I've found is the HGST Touro S. For continuous use, external AC-powered drives are better.

That said, don't freak out over one test. I personally don't like BlackMagic because it fluctuates so much. However it is commonly used and is free, so everybody has it. If you can, get QuickBench and run that with varying test sizes. Probably the most professional I/O test suite is DiskTestR, but it can be complex to use and interpret: http://diglloydtools.com/disktester.html

On my top-spec 2013 iMac with about 50% full 3TB FD, BlackMagic 2.2 gives about 300 MB/sec write, 650 MB/sec read. However it is quite variable.

A good app to monitor disk I/O is iStat Menus. However in the latest version 5, they have removed the ability to monitor I/O split between SSD and HDD portions of Fusion Drive. The older 4.22 version still does this.

 

[dyn]

Unfortunately the macperformanceguide/diglloydTools site is filled with too many false and inaccurate information. One of them is thinking that the "erase free space" function in Disk Utility clears memory cells where in fact it will fill them with data. Disk Utility will even tell you it will do this (it's about secure erasing data which is done by overwriting the old data with random new data or just 1's or 0's). The part about reconditioning is also nonsense due to the lack of understanding what TRIM and Garbage Collection do. Or something rather simple like how to turn off Wifi (by making the service inactive instead of simply hitting the "turn wifi off" button). And so on and so fort.

 

[joema2]

Unfortunately the macperformanceguide/diglloydTools site is filled with too many false and inaccurate information. One of them is thinking that the "erase free space" function in Disk Utility clears memory cells where in fact it will fill them with data...

Despite doing a site-specific search on http://diglloydtools.com/ I can't find any of the items you mentioned. Are you sure you didn't misunderstand any of these? Lloyd Chambers is an expert at Mac system software with decades of experience. I have no personal association with him, but having worked at that level myself I can tell he has a deep, detailed knowledge of system software, especially I/O and performance issues.

 

[dyn]

Try macperformanceguide.com instead. And do let other IT people have a read, my colleague had to laugh at some of his statements. He seems to be the "shout first, think and research later" kind of guy, he does update his posts with some new info. I just wouldn't blindly trust someone like that.

 

[joema2]

Unfortunately the macperformanceguide/diglloydTools site is filled with too many false and inaccurate information. One of them is thinking that the "erase free space" function in Disk Utility clears memory cells where in fact it will fill them with data...

OK I found the below statement from June 2009 (OS X Leopard), updated in Jan 2010. Is this what you're talking about? It sounds like he's accurately describing what happens.

http://macperformanceguide.com/Storage-SSD-Reconditioning.html

"Write over the entire capacity of the SSD. This can be probably be done successfully with Apple’s Disk Utility (“Erase free space”), but the fastest approach is a tool like DiskTester’s recondition command."

...nonsense due to the lack of understanding what TRIM and Garbage Collection do....

If the above article is what you mean, it was written about OS X Leopard in 2009. Did Leopard have TRIM support?

..."Or something rather simple like how to turn off Wifi (by making the service inactive instead of simply hitting the "turn wifi off" button)."

Are you talking about this article he wrote back in 2010 about disabling unneeded services? He wasn't talking about turning off WiFi -- in fact did not even mention WiFi. Rather he was describing how to disable unneeded services in general, and happened to depict an AirPort graphic as one example of various services among others.

http://macperformanceguide.com/Security-Network.html

 

[dyn]

For wifi it's the following: http://macperformanceguide.com/blog/2014/20141118_0958-OSX_Yosemite-disable-WIFI.html

The other one is a bit further back but not as far as 2010 I recall. It was in the beginning of his tool. Bit difficult to find, see if I'm able to. Anyway, that article also shows he doesn't quite understand what is going on. At first he uses the erase free space which doesn't resolve the problem. He then writes 1 large file to disk and probably erases it afterwards. What happens is that all those memory cells that were used for that single large file are now all marked as "reusable". The GC algorithm uses this to clear them out. In other words, it is more likely that this was just GC at play. Maybe I'm just nitpicking but it is these small things that make my alarm bells go off.