New
#31
The "Sequential" numbers that are advertised mean absolutely nothing for most users. If you are transferring a big file from the SSD to a HDD, you are only going to go as fast as the HDD will let you. I wish they would just stop advertising those numbers to be honest.....most of the drives out today are up against the SATA III wall anyway. The numbers that actually mean anything are the Random 4K Read/Writes, not the 32/64 thread either, and the access times.
To be perfectly honest, the most important things to consider today in an SSD are manufacturer and controller. Size isn't even much of a concern because the standard today is 128GB and if you want more than that, get 256/512GB. It's not like it was a few years ago when you had to settle on a small capacity because the 120/128GB drives are several hundred dollars.
Thanks WHS. i think i will let the software do the migration, see you in a week or so ... LOL
Thanks
Yes. What else?
Or you can go here and form your own conclusion:
AnandTech | Bench - SSD
Unfortunately, the advertised speeds are pretty irrelevant. They show you the Sequential Read/Write speeds and the 4KB Random Read/Write speeds. Sequential speeds, as I mentioned above, mean nothing to 99% of users, and the 4KB Random Read/Write speeds mean nothing to probably 95% of users. What they are showing you is Random 4KB Read/Write speeds with a queue depth of 64. That would be mean something for say, a server, which works in high queue depths. A home computer will normally be in the range of QD 1-4.
Here is a snippet of my Samsung 830 128GB with the speeds that are actually tangible highlighted.
Note: The 4K-64Thrd is what manufacturers advertise.....for my Samsung 830, the specs say 80000 IOPS Read, and 30000 IOPS Write. That would be great IF I normally worked in a queue depth that high.
I wasn't commenting on the sequential speeds, though they are important for certain applications. I was referring to the 4k write speeds and yes, I know these are at queue depths greater than one and 3x is an exaggeration, but they are significant in many applications and not just servers - on the order of 15% I reckon on light loads. That is the order of % of a generation of processor or graphics card or the order of magnitude we overclock to.
And here is some more performance measurements at lower queue depths (3) that show some significant differences:
http://www.anandtech.com/show/6337/s...250gb-review/6
http://www.hardocp.com/article/2012/...5#.UUpoLFfSCsM
http://www.hardocp.com/article/2013/...0#.UUpprVfSCsM
Last edited by GeneO; 20 Mar 2013 at 21:01.