Reg Hardware first reviewed a Samsung SSD - the South Korean giant's first, and a device aimed at computer makers rather than upgrade-hungry punters - way back in March 2009. Rather keen we were on it too. But that was two years ago - how has Samsung managed to improve matters in the meantime? Samsung Series 470 Samsung's …
I have been moving all of my kit to Samsung
In the last few years I have been moving most of my kit on the home network to Samsung and frankly it is on par with the comparable Seagate and definitely better than the comparable WD.
I am almost sold on this one. The 64G will be making my purchase list for a boot + client NFS root drive.
I am somewhat confused by the comment that "The closest rival to Samsung’s new toy is Intel’s 320 series".
I'd have said that for the 256GB part, the Crucial M4 (which is approximately £330) would kill this new samsung part stone dead. The M4 is sata-III, built on a 25nm process (so doesn't chuck out lots of heat), and it totally destroys the new samsung SSD in terms of performance: well over 400MB/s on a read, >300 on a write.
Price of flash?
So it's £20 cheaper for the 256GB one than it was more than two years ago? Is that not worthy of a raised eyebrow?
re: moores law
the higher capacities do seem to be permanently expensive but I have noticed some prices coming down. or rather, getting more for the same money.
This samsung 64gb model is around £110, 2 years ago I bought a 32GB corsair for the same price.
Considering getting one - could use it as a boot drive or to copy battlefield onto prior to playing.
i have it on good authority that doing so can shave 10 seconds or more off the load time for multiplayer games.
plenty of time to make a break for the best sniping point on the map :D
Moron's law more like
128GB is less than twice the 64GB, but 256GB is more than double the 128. Why? Is it the laptop sizing sweet spot so they fuck you over?
I am don't care much about minor differences of price or performance.
4 Gb Kingston costs 1.25 per Gb . So do these.
Review all the sizes please
Why assert about the 64GB one when only reviewing the 256GB one? Please review all the sizes so that we can see how the more reasonably priced ones fare compared to the top end models.
Flash drives continue to be expensive, small and fast, but not that fast.
120GB OCZ Vertex 3 with Sandforce 2200 (500MB/sec read) - £185
90%? I don't think so.
I'll stick with my M4, it's faster and cheaper.
I don't quite get why anyone would want to release a SATA II drive now that SATA III is becoming common, and I really don't quite get why you would award this drive a score of 90%, given that it's based on ye olde inferior SATA II technology.
Maths is hard
Perhaps it's SATA II because 3Gb/s > 250MB/s
Unless you're talking one of the 500MB/s drives there's no point in putting a SATA III interface on a drive that isn't capable of filling a SATA II bus. Except that middle management types might be willing to pay more for the same performance...
Those 4K random read/write numbers look quite low to me - you may want to invest in a real controller for benchmarks (I'm fond of the 6Gbit LSI SAS HBA myself,) and secure erase the drive before each test (assuming you want to print the 'fresh drive' results.) Also, I always have trouble trusting storage benchmarks run on Windows... just so much unpredictable behaviour there.
I guess this is why I run all of my own SSD benchmarks...
Re Distrusting storage benchmarks run on Windows.
Given that the vast majority of punters installing this kit will be installing on a Windows system benchmarking it on a Windows based rig would seem to be reasonably logical, hmm? Benchmarking it on a setup that a majority, even amongst El Reg readers, will not be installing on would seem a touch eccentric? If, for example, CrystalDiskMark cannot be relied on to read accurately in Windows that would imply that any software running in a Windows setup on this drive would not deliver the read/write speeds recorded by the benchmarking tool which is itself only another software package. In other words if you benchmarked this drive on, let us say, a Linux based setup the figures you would obtain could not readily be related to the disks performance in a Windows pc. So, even if you are correct in believing that benchmarking tools run in Windows cannot be relied on (a contention that I regard as somewhat dubious) running them in any other system to test the drive could not be related to real performance with the OS that a majority of us use. To sum up, if you are correct your point is meaningless and if you are not correct then you have posted b****x.
Its is a well known phenomena that the smaller SSD will be slower than the large capacity ones. It is generally to do with the number of chips installed (8,16 or 32) So the figures for the 256gb drive can in no way be relied upon to be the same for smaller drives, especially 64gb. Its generally the write performance that suffers, often dropping by 50% thus affecting any benchmarks. And yes, these drives are still too expensive 25nm was supposed to change all that, they traded longevity for price. And we got ...shorter longevity, so where's the price?