It depends on how old:
http://www.overclock.net/t/1571271/...rt-on-any-ami-uefi-bios-with-an-intel-chipset
Is it possible to get the 950 drive booting on an older mobo with an m2 pcie card?
Doesn't look like load times are different between an EVO and the 950Pro
http://techreport.com/review/29221/samsung-950-pro-512gb-ssd-reviewed/4
Doesn't look like load times are different between an EVO and the 950Pro
http://techreport.com/review/29221/samsung-950-pro-512gb-ssd-reviewed/4
in case anyone's interested, someone over on Tom's posted a promo code fm newegg good on the 950 PRO that = a $35 discount off samsung $349 price on the 512 GB variant -
i suspect it will also give a 10% on the 256 GB variant, just haven't tried it
Promo code AFSAMSSD1023B
The problem I find with most of these games is just like many review sites that try to differentiate 2GB video cards with the equivalent 4GB versions is that they rarely use games that can actually make good usage of more than 2GB.
These games are fairly lightweight loading wise, I do monitor drive speeds regularly with Gkrell.
While there is merit in blind testing popular games, it would take more than 5 or 6 games to represent the thousands of games that are out there..
I do expect the majority of games to be just as quick with Sata3 as with NVMe but I don't see any review sites going out of their way to find games that might show a difference.
Problem is that if you have to look for that one particular game out of the dozens you play with that one specific configuration, then how often does a customer hit that scenario in the real world? Seems like if it made an appreciable difference, we should see it just about every game.
2gb and 4gb video cards do show a difference, when you crank up the resolution/go multi monitor, in just about every game.
The 2-4GB divide tends to be with midrange cards like the 960 or 280 that aren't really suitable for modern games at very high resolution and most reviews I've seen don't test multi monitor setups.
What I tend to see is review sites that have already chosen their set of benchmark games so they can compare results between many cards without having to retest them. They may be a good choice of well known games but they aren't always suitable for tests like these.
Take this for example.
http://www.eurogamer.net/articles/digitalfoundry-2015-nvidia-geforce-gtx-960-2gb-vs-4gb-review
I have Rivatuner Statistics server running a few graphs on my keyboards screen, one of which is Vram usage and I can say from experience that even on Ultrawide 2560x1080 neither Tomb Raider (even with SSAA, let alone FXAA) or the Witcher 3 will ever go above 2GB Vram usage.
Battlefield 4 is known to cache data in Vram that can use above 2GB but thorough testing has shown it makes no difference in performance from a 2GB setup, It just saves reloading the textures.
Crysis 3 and Ryse are both happy with 2GB. Shadow of Mordor is tested with medium textures for both cards which it states in game only uses 2GB.
I haven't played Far cry 4 yet but Assassins creed is known to be a Vram hog and is the only game in the whole test that shows any difference.
My feeling is that a lot of the games I see being tested are chosen in the same vein and so they can be compared cross generations.
I would like to see how Dragon Age: Inquisition, Total War games and GTAV are affected in particular. Can't think of any more off the top of my head right now.
I do agree the majority of current games won't be affected, but I expect in the future as they become more demanding we will notice more of a difference even if it's only a few seconds here and there.
I also want to see what happens when firmware and drivers are matured, the disparity between these and the SM951 seem to indicate there are currently issues.
I don't think games are going to get any demanding even in the future or at least during this console generation. PC versions are only used to show off during e3 and other such events. By the time the game is launched its seriously watered down. This is the trend now and a lucrative one for publishers. VR might 'demand' a little more because it needs high fps to enjoy it.
Probably not, we've yet to see if the reduced latency with NVMe will result in a more responsive snappier system but I expect for those purposes you'll be very happy with the 850.
Samsung Magician 4.8 out, now just waiting for them to release the Samsung NVMe driver they have been working on. No links on the Download page yet.
newegg.com just changed the release date of the 950 Pro from the 1st of December to the 1st of November.