Full HDD affects game performance?

PentiumD No. 1 Fan of PentiumD's! New user 7 posts Posted: 08/11/2008 9:11 PM	I was just wondering if there is some sort of relationship between a full hard drive and decreased game performance. I have a 80 gig hard drive with 3 gig free and games that should play really well on my Pc are as slow as anything! Can someone tell me how much free space I should leave on my hDD to get top performance in games?
	permalink Digg Slashdot send to a friend

jake :) New user 69 posts Posted: 08/11/2008 10:11 PM	it may decrease your game perfomance, a cluttered hdd slows down your computer because the hdd has to go through more data to get o your files needed for the game, de-fraging your hdd may speed things up a little. with 3 gigs i would just re-format the drive ans start from scratch.
	permalink Digg Slashdot send to a friend

LostBenji Radio & Computer Technician New user 196 posts Posted: 10/11/2008 6:11 AM	Full drive is no good for just about anything unless its purely bulk storage. Yes it will be slugging your game and yes, its time to clean it out. Run disk cleanup first as your bound to have plenty of crap on it.
	permalink Digg Slashdot send to a friend

GoughLui I plug into everything like USB. New user 112 posts Posted: 15/11/2008 2:11 PM

Yes there maybe a 'link' between the full drive and slow performance - but I think this warrants a LOT MORE EXPLANATION than what is given here. Many people have the misconception that a full drive runs slower and that by pure deletion, their computer regains the speed. WRONG, for the most part. There are a few things to keep in mind (most of which will not apply to SSD's) - physically, the beginning of the drive is faster than the end of the drive due to your rotating platters and their varying circumference - which means that data stored at the end of the drive will be read back at about 2/5ths the speed on the data at the beginning. This can't really be helped. But what happens to full drives is, as files get deleted and inserted, a phenomenon called fragmentation comes in - files are 'broken up' into little bits to fill these gaps in order to fit. And, people making more space, will delete other files, making other new gaps which will be filled with fragments of other files. This DEFINITELY ADVERSELY AFFECTS LOAD TIMES for larger files spread over a few gaps due to the overheads of seeking (moving the head to a new position) and I/O overhead (which is the commands themselves, and time taken to execute the command). Hard disks can only seek so fast, and then they have to wait for the patch they want to arrive under the head - all up, this is called "Access time" and can be 13ms per seek. Files can be fragmented into 4000+ fragments (my hard disk did this a few times) - that's 52 pure seconds waiting for the bits of data essentially. This is an extreme example. That's what defragmenting is about. Essentially, given enough room, we want to re-assemble all of the fragmented bits back into one contiguous file, if at all possible. Defragmentation is hard without at least 10% of the drive free, as large files are difficult to reassemble and lots of moves are required - taking more time and being less efficient. For SSD's without a seek-mechanism, the dominant problem is the I/O overhead - with SSD's typically requiring 0.13ms of access time, things are much much faster where fragmented - and defragmenting is not recommended because of flash wear. There is one advantage of defragmentation - file recovery without a file table becomes easier if the files are contiguous. If they aren't - there's a fat chance you'll find the other 4000 bits... But defragmenting using normal tools still poses a problem - an often used file is not defragmented - this is the PAGE FILE. A RAM substitute when you run out of RAM. Most Windows users allow the page file to expand and contract as needed - Under your Advanced System Settings somewhere, there is the option to change this. When you leave it on this setting, there can be cases where you suddenly need more page file, and the file expands and becomes fragmented. Then it keeps fragmenting on ... Your slow virtual-RAM gets slower! And defragmenting, unless using special tools, won't do the trick here. You can check the pagefile fragments in the report of the normal defragmenter in XP, or using the command line defragmenter under administrator on Vista. I came to a shock when I realised, even with 4Gb RAM on my machine, my pagefile was 2Gbish and it had 1400 fragments. Solution? Disable the pagefile, defragment, make a permanent one. This has a disadvantage of mounting the pagefile at the end (slower) part of the hard disk, but at least if you set a permanent size, it won't break into bits. But the interesting part is this - unless the games are loading and reading stuff from the drive all the time, in big chunks, there is no reason for this to slow down the game EXCEPT for load times. Most of the time, given enough RAM, a game would load, and load ahead of time, all the stuff into the RAM. Your drive would have to really be performing like a dog to fall behind, or your RAM is insufficient. Otherwise, maybe your swapfile is fragmented. There are other reasons, of course, which involve Background Processes - those things that startup at the beginning and stay in memory (TSR, it used to be known) which take up memory and CPU time that are slowing you down. That is much more likely if everything is continuously slow. Likewise, a throttled CPU due to overheat, incorrect underclocked BIOS settings can do the same, as does a bad overclock when the buses aren't in synchronism, and suffer bottlenecks during data transfers to the video card for example. At least, I've hopefully given a more complete picture of this. To everyone who might read this, thanks.

permalink Digg Slashdot send to a friend

NetR@nger CEO Home Computer Systems(HCS) New user 50 posts Posted: 21/11/2008 9:11 PM	A cluttered and almost full hdd will effect the performance of everything.How do u survive with only an 80gb drive??,i have more than that in my music collection.
	permalink Digg Slashdot send to a friend

The Big Baboo Penguin In Training New user 27 posts Posted: 30/12/2008 7:12 AM	Wow "GoughLui" That's the first time anybody's ever explained that to me in language that I can understand :):) Thanks for that reply and A Happy New Year to you and all your family and friends.
	permalink Digg Slashdot send to a friend

AK8 New user 9 posts Posted: 05/01/2009 12:01 PM

GoughLui wrote: Hard disks can only seek so fast, and then they have to wait for the patch they want to arrive under the head - all up, this is called "Access time" and can be 13ms per seek Good post, but correction required - "access time" is the time it takes to write to the drive, not the time it takes for the head to reach the specified location on the platter (which is seek time). Access time is defined as seek time plus rotational latency, which is why a drive with higher RPM's is far more likely to be faster overall than a drive with lower RPM's. For OS drives, access time is also a better determinant of a drive's performance than any seek time value. Most publications (unfortunately including APC) tend to place emphasis on seek time or STR (sustained transfer rate). GoughLui wrote: But defragmenting using normal tools still poses a problem - an often used file is not defragmented - this is the PAGE FILE. Not only that, but also MFT data. One of the main purposes of the MFT data is to index the location of all files and their fragments. If the MFT data is fragmented, then (as with fragmented files) the drive heads have to travel from one location on the drive to another just to find all of the fragments of a particular file. The more data on a drive, the greater the amount of MFT data and the greater potential for the MFT data to become fragmented. Also, a drive full of fragmented data is also more likely to cause the system to further fragment the MFT data due to a lack of contiguous free space around the MFT zone. My own experience is that fragmented MFT data has the same or greater negative performance impact on a system as a defragmented paging file - particularly with smaller drives. Windows Defragmenter doesn't defragment MFT data. For this you need a third party defragmenter that specifically supports MFT defragmentation. I only know of four utilities that do this - Raxco PerfectDisk, O&O Defrag, Executive Diskeeper Pro and Norton SpeedDisk. These are also capable of defragmenting other protected files such as paging files. If you're unable/unwilling to purchase one of these utilities, you can at least download PageDefrag from MS Sysinternals to defragment the paging file. MS recommendations for the paging file include adding a paging file to a seperate drive. One paging file per drive (not per partition). By having one on the system drive and another drive, the system will choose the paging file on the partition with the "least activity". There is a caveat though. If you have drives of different speeds (eg. 10K RPM system drive and 7200RPM storage drive), the system is likely to use the paging file on the slower drive because of the lesser amount of activity. The 10K RPM drive may be able to better handle paging operations on top of other operations than allowing the system to use the 7200RPM drive for paging operations. Other factors such as the number of heads and cylinders a drive has, or quality of the drive controller can also affect this. As an example, a drive with one or two platters will likely be faster than a drive with four platters simply because it doesn't have to switch heads and cylinders as often as those with more platters. As GoughLui suggests, it's wise to set the maximum and minimum size of the paging file to prevent fragmentation. It's a common misconception that paging file optimisation means moving the paging file off the system drive to another drive. Not so. Windows requires a paging file on the boot drive for doing memory dumps. Memory dumps can be disabled, but this is not recommended since they are a valuable part of diagnosing system faults.

permalink Digg Slashdot send to a friend

alan7486 New user 7 posts Posted: 25/01/2009 7:01 PM	Thanks Very MUCH for the interesting insight
	permalink Digg Slashdot send to a friend

Not signed in
	Anonymous user
Login here
Sign up now!

	May  APC out now!
read more