I purchased a 500gb Samsung Evo and handed down my 256gb 830 to my daughter who in turn has handed down her 128gb C300 to my son. Her C300 was being particularly slow so I just thought it was down to the drive getting old but once I had cloned her OS onto the 830 and rebooted it's exactly the same. I then thought maybe its the OAS and it needs re-installing so that's what I was planning to do as I prefer to do full re installs anyway. This is where it got more confusing as I installed the C300 into my sons pc with her OS install intact with the intention of cloning his OS onto it but accidentally let it boot off the C300 and it was as quick as ever. So the question is why is my daughters PC not utilizing the SSD properly. She has a rocket 620 Sata 3 card as she doesn't have native sata 3. I know some people don't rate the rocket card but it's always been okay in the past.
Wierd... When I read the title (and most of your post) I thought Is the mobo in AHCI mode not IDE? Is is plugged into an Intel port rather than a different one Is the SSD failing But you answered all of those questions... Have you tried it not in the RAID card? Maybe that's the culprit - SATA 2 might not be that much slower than SATA 3 if her board supports that...
Yeh I did give that a thought to try, it would rule it out. I do have her 1tb Samsung F3 plugged into the sata 3 card as well but again it's always been like that.
Have you tried something like AS SSD to check the actual quantifiable speeds of each drive? Just for arguments sake could you try unplugging the F3? It shouldn't make a difference but you never know...
Whilst the C300 is highly likely to have slowed down d.t. lack of trim, this wouldn't necessarily explain it being quick on the first boot on your son's machine as there'd obviously have been no opportunity for it to have run. That said, it would both naturally be quicker on a proper 6Gb/s port, so you may just be noticing a comparative increase in speed, not the full speeds... ...&/or if you're looking at the speed after a while (so after some writes & deletes & stuff, not just the initial boot) then trim could have done its thing. The only other possibilities that immediately spring to mind are - 1. if the pcie slot that the card's in is only 1.0, not 2.0 - this would obviously halve the bandwidth 2. or if the pcie slot is one that's splitting lanes using an NF chip, as this would make it noticeably slower d.t. increased latency - though i've no idea what your daughter's machine is so these may, of course, be completely irrelevant.
It originally ran fine in her pc that's the confusing thing. Her board is a MSI 790 GD70 and it's in the 2nd pci-e 16x slot I think. Edit.............Actually I think it's in the bottom slot. Yes bottom slot.
Right, well being an AMD board with nothing but pcie 2.0 slots then it's clearly neither of the half ideas... it was worth checking though just in case. it could, however, quite feasibly still be the lack of trim with insufficient idle time for GC to do things - well, the C300 really didn't have the best GC (which was *a* reason why i went for V2s in R0 back in the day) & that could easily explain a slow down... ...with, again, the initial improvement in speed d.t. it being on a proper 6Gb/s controller - with trim then improving things subsequently. Naturally you could tell by SEing the drive & doing a quick fresh OS install to see if things are still as slow. Yeah, there's just some SSDs that really aren't fabulous in non-trim... Well, whilst i knew that they weren't great in non-trim when i bought them (though naturally that wasn't how i was originally using them), i've been surprised at how poor my old pair of 830s are on a proper raid card where, even used almost exclusively for sequential r/ws (batch converting many 100s of GBs of a/v & whatnot), some s/w crashes when writing once they've had too much written to them in too short a period &, without stopping the work process completely to give enough time to let GC do its stuff, i've got to quick format the things to reset the nand. i'm now umming about the downside of losing lots of the sequential speed & sticking them on the 3Gb/s intel ports (the 2 6Gb/s ones have 840 Pros on), which would, of course, increase the process times... ...vs putting up with having to wipe it periodically - at least until the new pcie SSDs appear as, assuming they're not stupid money, i'd then rearrange things again. Naturally though, that's not your daughter's usage, but, assuming you're proposing to use the card again this time around, i'd keep my fingers crossed that the increased capacity of the 830 provides more leeway for the drive to look after itself... ...& certainly be mindful that it is, similarly to the C300, an issue with the Samsungs so not to go mad with filling on & ensure that the sleep settings aren't too aggressive. (there's always extra OP as well)
I've moved the C300 into my sons board which only has 3gbs ports and it's way quicker than it was on the 6gbs rocket card (which I'm thinking is now the issue). She did fill it almost right up but I did have OP set to about 25gb. Sleep settings, hmm, it's set to S3 and gets put into sleep once a day.
i misunderstood that your son's machine didn't have 6Gb/s either. Anyway, it's only an S1 state that would retain power to the drives whilst in the sleep state... ...though, naturally, it then means that all & everything, other than the processor, stays fully powered. What i was more meaning though was both the amount of time before it goes to sleep - well, it may only be going to sleep once a day, but if the machine was only on for an hour in total then the sleep setting would, naturally, be pretty short. Generally it's better to be set to at least an hour - though if speeds are suffering then it's obviously worthwhile increasing it. Otherwise, extra OP's great, but it shouldn't be done at the expense of free space. Well, an issue with SSDs is that of block fragmentation (ie blocks with dirty pages) - & if there are too few blocks as free space then the SSD has to prioritise this; in the extreme, having to do it on the fly... ...data being written in pages, but only erasable in blocks of course. Now, block combining is part of what GC will do during idle time, however if the drive's already struggling d.t. a lack of trim, this is all going to compound... ...GC will have to erase blocks so that there's somewhere to combine fragmented ones to, then combine, then erase the blocks that were combined from & then rinse & repeat - plus there's wear levelling to be done... ...& if the GC isn't great in a non-trim environment to start with then it'll naturally be even worse having to do too much, & you'd need loads of powered idle time for it to maintain speeds & even more to recover when it's been backed into a corner. Anyway, whilst there's not been a discussion about free space & OP for a while (& i'm not attempting to start another one as it got somewhat dull), i always recommend maintaining *at least* 15% of the formatted area as free space before thinking about extra OP on an OS drive; ideally aiming for 20% free space + ~28% extra OP... ...though naturally adapting it to the usage - so, as kind of extremes - - unless they were semi-constantly being installed & uninstalled, a games only SSD would need a very minimal amount of free space & no extra OP - whereas a usage with a very high random write (& erase) load would need far more - hence the very large amount of native OP in many Enterprise drives.
Just wondering if the partitions are 4k aligned or not, it's possible that if they weren't then cloning might have copied the misalignment.
put the 830 in the sata 2 port on your daughters pc also enable AHCI on the C300 (if your on wndows 7 - it`ll use trim as the C300 does support it)
i had alignment issues on a patriot wildfire drive on a sataII port and the performance was not poor. I'm not sure there has been any qualitative statements as to what constitutes poor performance though.
