03-15-2012 06:52 AM
Yesterday I received the replacement with latest firmware 0309.
After I installed Windows 7 x64 with all updates, all the drivers I need, antivirus and some basic programs, I run AS SSD. Score: 823
Results are reproducable.
I will bench once a month from now on. It seems, that I am not the only one with that issue, but I hope it's exceptional.
Thanks for the great support.
03-16-2012 01:30 AM
Great, everything looks good now to you right?
Looks good to me, nice to see you had a fast turnaround on the replacement too!
03-17-2012 01:20 PM
We will have to wait and see if this fixes this problem. A secure erase and restore looks great too afterwards but performance will drop after a few months. This will be a good test to see if it was drive or system related.
03-18-2012 02:18 AM
Ahh OK, well keep us posted when you think it's been long enought to know!
Hopefully it'll all be sorted out now!
03-20-2012 05:06 PM
I was looking at sata 3.1 today and it seems it adds support for 'queued trim'.
It seems the trim command isn't currently compatible with NCQ and requires the queue to be emptied before the trim command can run and the queue allowed to build up again.
Maybe someone with this issue could try disabling trim for the duration of the benchmark?
03-20-2012 06:36 PM
A few days after my Secure erase my 4k 64 thrd are down to well over half their secure erase value.
I tried set DisableDeleteNotify to 1, rebooted, no change,
03-21-2012 02:53 AM - edited 03-21-2012 02:54 AM
You do not need to reboot, and I'm not sure if that resets it or not, so please test again without a reboot just to be sure.
It's a good idea, but we aren't sure it matters anyway, he was just thinking out loud I believe. Also, please add your current CMART Data, so Targetbsp can see if anything looks out of the ordinary, thanks!
03-21-2012 11:09 AM - edited 03-21-2012 11:09 AM
I thought I might not need to. It did not reset it. I don't know why this would make a difference in any case - it doesn't seem related. It would seem the issue described is TRIM not getting a chance to execute if the queue constantly has work.
But thanks for trying.
03-21-2012 11:17 AM
From what I understood, it's the other way around. The queue would be emptied so the trim command can execute. But I only read one article about it. And it was only a single paragraph. So I didn't look into it deeply!
04-12-2012 10:13 AM
I've just seen that there is a new firmware 000F.
So I thought: Right, I should test anyway, if everything is okay with my replaced SSD.
And what can I say... the same "problem" occured again.
Screenshot AS SSD:
You know how pedantically i test, so please don't suggest things like energy saving features, encryption etc.
I've seen in forums, that the acc. time will increase with the new firmware.
Will the new firmware fix the "problem" or am I just unlucky?
I hope the firmware fixes it, though it would mean, that I sent back the "faulty" SSD and waited for a new one for no reason...