By what you have said it sounds like your drive is failing.
250Gb Seagate, they don't make those anymore.
What's the manufacture date on that puppy, early 2000 probably.
HDD making loud wongwongwong sound when loading game
- Thread starter Dorray
- Start date
250Gb Seagate, they don't make those anymore.
What's the manufacture date on that puppy, early 2000 probably.
Shouldnt it make the struggling sound with other games or under a high work load? Played BF4 an hour ago. MY PC is dead silent. I got 8 fans in here.
Played ED after that, same funny struggling noise. Game was just fine. Temps all good.
I'm loving this!
What percentage does windows say is empty on the hard drive?
I'm assuming a defrag did nothing.
The only thing I got is its not contiguous and was saved last on the drive?
What percentage does windows say is empty on the hard drive?
I'm assuming a defrag did nothing.
The only thing I got is its not contiguous and was saved last on the drive?
Depends on how much program storage he needs. 500GBytes for an SSD is small. You don't want to fill an SSD too much past 50% full because it will lose efficiency and lag. Throw in the windows dynamic page file (if it's on the same drive) that you you really can't see during run time and you can kill the advantage of an SSD.
From 250Gb to 500Gb thats double and not too heavy on the wallet.
I did check the spec the OP quoted and it is quite modest so my suggestion was based on that.
I have a similair spec machine at home and I wouldn't be spending any sort of big money on upgrading it.
Currently has a 256Gb SSD boot with 2Tb HDD for storage.
Anyway just my thoughts on a potential solution for the commander.
You're thinking you can cram 90% full data on an SSD drive like a regular drive without a corresponding efficiency loss. You can't, they work differently. A spindle drive's head seeks out and cubby holes data onto the old drives and they work pretty much up to efficiency as long as they are defragmented. An SSD drive has to move data around to accomodate new data in blocks. Once they are loaded past 50%, they begin to suffer efficiency losses moving that data from one place to another and the fuller the SSD drive gets the slower it gets. It won't be as slow as a spindle drive, but the speed efficiency advantage that you bought the SSD drive for is dwindling quite a bit as you get it loaded significantly beyond 50%.
edit> Your efficiency loss comes with the extra data writes necessary to move blocks around to accomodate new data... Best analogy... think of a warehouse that all the shelves are full at 50% and the other 50% is the aisle space. Once you fill all the shelves up and you have to move data to/from those shelves, the more crap stored in the aisles, the more stuff to move to get to the shelf space. As long as you're OK with your SSD not really operating up to its full potential, then go ahead and cram it full... My own opinion is that's a waste of money/potential.
Last edited:
Deleted member 38366
D
A live 250GB HDD?
That is very old and now deep in the realms of possible mechanical failure.
If anything, I'd strongly recommend making a Backup of all relevant Data, just to be prepared for the worst.
That is very old and now deep in the realms of possible mechanical failure.
If anything, I'd strongly recommend making a Backup of all relevant Data, just to be prepared for the worst.
You're thinking you can cram 90% full data on an SSD drive like a regular drive without a corresponding efficiency loss. You can't, they work differently. A spindle drive's head seeks out and cubby holes data onto the old drives and they work pretty much up to efficiency as long as they are defragmented. An SSD drive has to move data around to accomodate new data in blocks. Once they are loaded past 50%, they begin to suffer efficiency losses moving that data from one place to another and the fuller the SSD drive gets the slower it gets. It won't be as slow as a spindle drive, but the speed efficiency advantage that you bought the SSD drive for is dwindling quite a bit as you get it loaded significantly beyond 50%.
edit> Your efficiency loss comes with the extra data writes necessary to move blocks around to accomodate new data... Best analogy... think of a warehouse that all the shelves are full at 50% and the other 50% is the aisle space. Once you fill all the shelves up and you have to move data to/from those shelves, the more crap stored in the aisles, the more stuff to move to get to the shelf space. As long as you're OK with your SSD not really operating up to its full potential, then go ahead and cram it full... My own opinion is that's a waste of money/potential.
A mechanical HDD will suffer more from being 90% full, even if rigorously defragmented (which can take forever if it is nearly full, or be outright impossible if too full), than most modern SSDs will.
Most any remotely recent SSD is going to have at least 7.37% (NAND is built and sold in SI binary size units, while drive capacities are in metric units, which means a 128GB SSD almost always has at least 128GiB of NAND) and usually closer to 15% over-provisioned spare area (a 480 or 500GB SSD will typically have at least 512GiB of NAND) out-of-the-box (some will have much more, especially if they have a fraction of physical capacity reserved for SLC caching as well) and nearly every vaguely reputable SSD controller/firmware has very mature TRIM support, garbage collection, and wear leveling at this point. Performance will depend more on avoiding long periods of sustained writes that fill up pseudo SLC caches than on total capacity used (though dynamic caches do depend on how much space the drive has free), and both issues can be mitigated if the drive is allowed time to garbage collect (a few minutes here and there where it's powered on and not being written to, almost a given in client use). Yes, using most of a modern SSD's capacity is still going to reduce performance and increase write amplification, but it's not likely to be a dramatic difference in most uses, and certainly won't bring an SSD down to mechanical drive levels of performance.
With a mechanical HDD, having a drive near full automatically means that a large portion of the data on it is closer to the spindle, where performance will be about half of what it is at the outer tracks, and significantly increases positional latency when having to seek for data that, perforce, cannot all be on the outer tracks.
As you mention, there are very good reasons not to fill either type all the way if you don't have to, but most any SSD already has a significant fraction of it's "shelves" set aside so they cannot be entirely crippled when filled to near design capacity with user data, and ultimately, drives exist to hold data. If someone's budget is tight, running an SSD closer to capacity than ideal is still far better than doing without an SSD. The cheapest ~500GB SSD I can buy on Amazon or Newegg today ($45), filled to 90% capacity, will be vastly faster, in almost every real-world test, than the fastest 15k RPM SAS drive I've ever seen, freshly defragged and at 20% capacity.
Why solid-state drive (SSD) performance slows down as it becomes full
sums it up in a nutshell... here they're saying 70% full vs 50% full...
sums it up in a nutshell... here they're saying 70% full vs 50% full...
While the author explains the mechanism behind performance degradation, he fails to substantiate his claims (which fly in the face of everything I've experienced since the first few generations of SSDs), neglects to mention the myriad of mechanisms in place to ensure performance is rarely crippled, and doesn't offer any clear rationale for his recommended cut off of capacity used.
Most remotely modern SDDs, at well over 70% capacity, will not lose enough performance to fuss over and essentially none of them will lose as much performance as a mechanical HDD will.
Just ran some tests myself on my seven year-old ADATA SP900 128GB. It's one of the smallest and oldest SSDs I still have, and has some of the lowest factory overprovisioning I've ever seen (was originally 0%, but they bumped it up a bit with updated firmware about five years ago).
I benched the drive at 18% full (I'm mostly using it for a documents/image editing/swap folder), then copied about 80GB of my video recordings to it, then benched it again at 84% full.
CrystalDiskMark is only a quick test, but it writes a considerable amount of data in a short time and hammers the drive much more than general use will. As you can see, there was no loss of performance...if anything it's slightly faster (probably because throwing a huge amount of writes at the drive triggered it's garbage collection).
Every other SSD (unless it's some thumbdrive in a box scam) I've used in at least the last five years or so behaves similarly, or better, with next to no performance degradation until they are near completely full, because they all have more factory overprovisioning and newer controllers/firmware manage NAND better than this drive.
This seven-year old ADATA drive, which is worth about twenty bucks, is, even at 84% full a faster general purpose drive, than any rotating platter drive ever manufactured, barring those that have their own SSD cache's built in (SSHDs or similar).
Interesting results. Thx for taking the time to run the tests. Wonder if there's a way to test overall system performance with the Windows Swap file being located on the main SSD drive vs. being located on a separate dedicated SSD drive. Something tells me that there would be a difference in that case because the Swap file read/writes would be happening independently/simultaneously from the read/writes going on the primary drive, but I'd love to see the performance difference.