I need to add an addendum to this thread. RAM Caching done by PrimoCache Beta v0.9.1 which can be found at this link: I do not have an SSD to give you performance numbers with a RAM Cache, but I can assure you they eclipse the numbers of my HDD by a steep margin, especially Random 4K reads/writes.Īs a precaution I must note, If you wish to have RAM cache accelerated writes, because the data is held in RAM, you run the risk of data corruption if you loose power to the system unless a UPS is implemented.Īll tests were done with Crystal Disk Mark v3.0.2 圆4 and can be found at this link:
My personal opinion, is that most users would be ok with 16GB of RAM and a 6GB RAM Cache as most users do not transfer files larger than 6GB that often. This will require more RAM in your system to get a large enough RAM Cache so that you don't run into a problem of your Cache filling up with part of the file then having to flush that stored data before it can read more data during a transfer. Files up to a maximum size of your RAM Cache will transfer between drives at a much faster rate than a normal HDD or SSD can.The downside to RAM Caching, is if you have to transfer file sizes that are extremely large, like 10GB or larger. Ok, so what does this mean in terms of real world performance? Programs load much faster and web pages within web browsers load faster. Pretty pathetic from a SATA II generation green drive, eh? (This test took 3 times as long as the Cached test) These are my control tests with the same data set without a RAM Cache. The interesting part is when you get to Random 4KB reads/writes. I would ignore the Sequential reads/writes and Random 512KB reads/writes as those are pure RAM data transfer rates.
#Diskmark for ram windows 7#
OS : Windows 7 Home Premium Edition SP1 (圆4)
Read and write performance from my OS partition of my RAM Cached 2TB Seagate Barracuda LP (This is a SATA II device):ĬrystalDiskMark 3.0.2 圆4 (C) 2007-2013 hiyohiyo I used a 16KB block size to save on overhead and tested performance using Crystal Disk Mark with a 4000MB Data set and ran 4 iterations of each test. In my experiment I setup my system, a Seagate 2TB Barracuda LP, 16GB Mushkin DDR3 2400 MHz RAM running at 2133 MHz, with a 6144MB RAM Cache. After a certain amount of time has passed (that you are allowed to specify) it will attempt to write all of the remaining deferred blocks to the SSD or HDD, if they are not relevant anymore they are dumped to trim and deleted. This algorithm also has a Cache flush feature.
#Diskmark for ram software#
The software utilizes a math algorithm to defer blocks of data before they get written back to the SSD or HDD, then systematically re-writes the data back to the SSD or HDD. The larger area of space you give a software program access to makes your RAM Cache more effective with large file transfers. A true RAM cache is a software selected area of your system's RAM that is set aside and can only be used by that program as a means of reading and writing data to an SSD or HDD. Hello everybody, I wanted to share my experience with RAM Caching and to give people more of an understanding on what it is and what it does for your Computer.įirst off I need to explain a little about what RAM Caching actually is.