SSD speeds with Tempo

[beachmat]

We installed a Sonnet Tempo SSD in a 6-core Mac Pro 2010 with a Crucial M4 SSD and there didn't seem to be any speed increase over the SATA 2 SSD installed in one of the drive bays. So I tested both with XBench and indeed the Crucial came up with a lower score (430 vs 490), only beating the the SATA 2 drive on 1 measure. I'm wondering if anyone can suggest why this might be? As far as I can see there is no configuration needed for the Tempo. I tested a SATA 3 SSD in a 2012 Mini and the score was around 600, beating the SATA 2 Mac Pro SSD on every measure by some margin, which is kind of what I was hoping for with the Tempo.

Thanks

 

[DeltaMac]

The Crucial M4 is an older SSD technology, showing specs of 415 MBps in testing (from 2011)
Not sure why you would expect better than that. It is what it is, with tech from more than 6 years ago.
You can't force the SSD to provide speed faster than its design capability, just because you try a faster interface card. That won't happen.
Try your Sonnet with the more modern MX300. I imagine you will then see some difference.

 

[beachmat]

I will try it thanks. I still wouldn't expect it to be slower than a SATA 2 SSD though.

 

[DeltaMac]

You are seeing the limitation of the SSD device itself, and not the bus connection that you might be using.

If the controller on the SSD itself is relatively slow (like the Marvel in the M4), then SATA 3 is not much help, despite the possible speed capabilities of the Tempo card.

You have to compare the specs for each individual SSD that you might use.
You said that you tried an M4 SSD. What was the OTHER SSD that you tried out in the Tempo card?

 

[beachmat]

The SSD in the 2012 Mini I tested is a Samsung Evo 850. We also have 2 Samsung Evo 850s on a Tempo Pro in a Mac Pro 2008 which I don't have access to right now, but again I don't think the score was approaching that of the Mini. But then again it is an older Mac.

 

[DeltaMac]

Sorry, I guess I missed what you were asking about in your OP.
Your question is now not too clear (to me, anyway!)
The EVO 850 is also SATA 3.

Anyway, with the Tempo, in the 2010 MacPro.
Did you try a different PCI-e slot? Perhaps it is simply a matter of moving from a 4-lane slot to a 16-lane slot (?)
You might even try running the Expansion Slot Utility. You can run that manually from the System/Library/CoreServices folder. That MIGHT provide a suggestion for a better slot for you to use.
Might help (or not...), but worth doing.

 

[h9826790]

Sorry, I guess I missed what you were asking about in your OP.
Your question is now not too clear (to me, anyway!)
The EVO 850 is also SATA 3.

Anyway, with the Tempo, in the 2010 MacPro.
Did you try a different PCI-e slot? Perhaps it is simply a matter of moving from a 4-lane slot to a 16-lane slot (?)
You might even try running the Expansion Slot Utility. You can run that manually from the System/Library/CoreServices folder. That MIGHT provide a suggestion for a better slot for you to use.
Might help (or not...), but worth doing.

Expansion Slot Utility won't work on the 5,1.

We installed a Sonnet Tempo SSD in a 6-core Mac Pro 2010 with a Crucial M4 SSD and there didn't seem to be any speed increase over the SATA 2 SSD installed in one of the drive bays. So I tested both with XBench and indeed the Crucial came up with a lower score (430 vs 490), only beating the the SATA 2 drive on 1 measure. I'm wondering if anyone can suggest why this might be? As far as I can see there is no configuration needed for the Tempo. I tested a SATA 3 SSD in a 2012 Mini and the score was around 600, beating the SATA 2 Mac Pro SSD on every measure by some margin, which is kind of what I was hoping for with the Tempo.

Thanks

You better compare the details but not just the score. There are 8 measurements in the Xbench disk test, and at least half of them are not really SATA 2 bandwidth limiting in your case. Xbench can gives you a rough idea of how the disk perform, but not very consistent. You need to run the benchmark a few times to know the average / max score. Also, if there was some background disk activity when you running the benchmark, the score can be lowed quite a bit. Therefore, its totally possible that the score on the Tempo card is lower because those non SATA II bandwidth limiting items get lower scores (by chance) which affect the final score significantly.

In any case, if the raw data (those 8 measurements) looks normal, then both the SSD and the Tempo card should be good to go.

 

[beachmat]

Ok, thanks. As I say we also have a Tempo Pro with 2 Samsung Evos in a software RAID 1, and the score is 469, still nowhere near the 600 of the Evo in the 2012 Mini. The moral of the story seems to be that you may get some improvement with cards but it won't be as good as an onboard SATA port.

 

[h9826790]

Ok, thanks. As I say we also have a Tempo Pro with 2 Samsung Evos in a software RAID 1, and the score is 469, still nowhere near the 600 of the Evo in the 2012 Mini. The moral of the story seems to be that you may get some improvement with cards but it won't be as good as an onboard SATA port.

RAID 1 won't improve performance. In fact, may lower it a bit.

Tempo Pro simply do don't have enough bandwidth for 2x SATA III SSD at the same time. Therefore, RAID 1 will definitely reduce the overall performance in your case.

And yes, Tempo SSD should be a bit slower than native SATA III port, but not significant.