Discussion in 'PC Hardware' started by ForrestGump, Oct 29, 2018.
You need to login to view this posts content.
You need to login to view this posts content.
Samsung 860 EVO 1Tb
I don't think it makes much sense using different settings and/or even different versions of CDM.
Furthermore it makes no sense to compare apple with oranges (SATA drives with NVME)...
People should use the very same version of CDM with default settings and explicitly compare same protocols...
yes i know that NVMe interface gives much more speed than normal sata drives that's why i specified to mention the SSD model.
A laptop is good if your always on the move, but I can't understand why peeps have a laptop and do not move around with it
Do not compare things that can not be compared or measured with these tools.
Do not forget that this is an electronic process, not mechanical process.
And one more thing - don't forgot that there are at least 4 different types SSD drives and they are a bit different, the working pricipe is different and also speed is different.
But as You all know (at least I think so that all), even different light (by color) has different speed, but how it afects You or your life? So also here is no matter, has this SSD 599 MB/s or it is 999 MB/s. But honestly say, I don't beleve none of those numbers, what You have published. Simply these parameters are not measurable with these methods.
Here the speed determines the CPU and settings of your computer.
Both Yen and kaljukass are correct. You're comparing apples with oranges.
The numbers are only useful when you're comparing drives of the same size.
And look at the big differences in speed between NVMe and SATA.
It is nice to see what hardware everybody has though.
My new Samsung 860 PRO
i know that speed depends on the type of SSD that you have that's why i specified that SSD model with its capacity have to be mentionned in the first post.
Maybe the real speed cannot be measured but anyway this gives a global idea on the SSD read/write speed and may help users to make their choice when buying SSD.
One SSD and two results.
Actually nowadays are those speed results not suitable to make a decision which SATA SSD one should buy...
- SATA III sequential speeds are close to bus speed limit either way
- results are dependent on the tools themselves, even the same tool CDM 5 to CDM 6 delivers different results depending on kind of blocks which are read/written and way how to determine the speed at all.
Those variations on results exceed the difference of brands itself.
- depends on cache usage / amount of cache / way of caching. For instance RAPID mode installs a RAM disk for caching and bench mark results (write of small blocks) change a lot.
- bench mark tools are synthetic bench marks. How far they compare to one's real usage is another question.
What's more interesting is the durability and infos about that come from the manufacturer themselves and are based on predictions since there are no real facts about new technologies such as 3D V-NAND.
There is general experiences already. For instance SLC / MLC / TLC....BUT actually nobody knows how long the own SSD really lives.
And there are different values we only get from the manufacturers themselves and we actually do not know how to categorize them:
- MTBF = Mean Time Between Failure (My old Vertex II is rated here 2 million whereas the Samsung EVO is 1.5 only)
- TBW (Tera Bytes Written)
- DWPD (Diskful Writes Per Day).
MTBF makes IMHO not much sense. It's a pure extrapolation....
TBW makes more sense, although it should be related to total capacity and hence DWPD makes more sense....
Finally it also depends on usage (R/W ratio).
We choose relying on what the manufacturer says about their products and reputation of the former model and general reputation....
But if a certain SSD is really good or bad only time can tell.
And one extra remark:
I can clearly notice that when I make a clean installation of Linux the first few boot times are the slowest.....
I presume that the first optimizations of the SSD take place during the first idle times of the fresh installed OS.....
SSD tester always advertise their standardized condition ("tested on a fresh / clean installed OS")...
So finally I'd say when buying a new SATA SSD you can leave speed tests alone to make a proper decision.
Are you sure that relying on the model and on what is mentionned by the manufacturer is enought to make a choice to buy a good SSD ? because every manufacturer can say what he wants to attract consumers to buy their products ?
That's even the point. We only can rely on what they say about their own products and sure they do not say something negative. And we can read at forums about established (old) models....but nobody knows in advance how long a new model will live.
When I had chosen my first SSD I really did not know which brand and there is no experience before since I never had a SSD before.
I loved the Samsung Spinpoint HDD series, but they did not make SSDs that time yet....
I had chosen OCZ. They convinced by their advertising and bench marks. I also had no other sources to make a decision. (It's a SATA-II device and it's around half speed of recent SATA models, but already 8 years old).
OCZ is gone and Samsung makes now own SSDs since they have canceled their HDD manufacturing.
My second SSD for Linux was then a EVO 850 many years later.....
It's actually still guesswork to buy a 'good' SSD when it comes to lifetime.
I am very surprised that my old one is still working.
But that's not NAND technology alone. Another one I am very curious about is OLED technology at screens...
yeah, running os with SSD already allows people to use their machine with certain confort as they will notice huge difference while they are working on it but i bet having SSDs running in RAID 0 will give much more speed so i'm curious to know which is exactly the boot time under the configuration that you have.
which os are you running on the machine using RAID 0 ?
could you please do a boot time test and tell me the timing that you get between the moment when you push on power button and the moment when login screen appears?
then another test when you time from "restarting" on windows session until login screen.