Samsung 960 NVMe Evo

>>61264522
only super fast for continuous reading/writing.

>>61264522
M.2 drives are PCIe storage, normal SSDs are SATA storage.

Boot initialization for SATA devices happens before PCIe devices in the POST process.


tldr; SATA devices get booted into first

>>61264522
Your gaems don't load any quicker either.

>>61264566
Except that's wrong.
http://www.tomshardware.com/reviews/samsung-960-evo-nvme-ssd-review,4802-4.html

It clearly still performs well under random read and write, and the queue depth performance is good.

>>61264658
What are you talking about pajeet?

>>61264522
SSDs don't make device enumeration and driver startup any faster. dmesg

>>61264522
Because processing needs to be done.
The bottleneck isn't your storage anymore.

>>61264636
Will that ever change?

>>61264522
>what is bottleneck

File: 1488750318133.jpg (90KB, 500x423px) Image search: [Google]

90KB, 500x423px

>>61264522
>mfw ~18 seconds from grub to login screen on a 10 year old laptop with a 7200rpm drive running sysvinit
lel@systemd hindus

>>61264522
Sounds like a bad case of a systemDisease

systemd-analyze blame

Post the output, let's see which service might be the culprit.

>>61265224
>$ ls /etc/rc[35].d/S*
>$
gjm8

>>61265202
I suppose a manufacturer could create their own POST initialization that started with PCIe devices instead of SATA, but it's currently not standard.

I think if you have a laptop that doesn't have SATA devices at all it will go straight to PCIe initialization, but idk about desktops.

Is there any way to get rid of the nvidia bloat? I swear all that crap slows my nvme m.2.

>>61265282
It would certainly be something to look at.

File: Screenshot_2017-07-08_01-20-55.png (886KB, 1659x494px) Image search: [Google]

886KB, 1659x494px

Alright, so I was able to cut off most of the userspace stuff but the kernel still takes suspiciously long. I had a look at dmesg as >>61265050 suggested and it turns out ACPI is a main contributor. There are these two 2s bumps in pic related. Does anybody have an idea why? Can I safely disable ACPI in grub?

>>61265571
looks like at least one of those is the boot order thing anon mentioned.

>>61265644
If it's M.2 it's still PCIe even if it's SATA speeds moron.

I'm going to try here instead of making my own thread.
I recently built a new pc and i have to manually go and boot the SSD in bios every time i start up my computer. It is not in the list of boot prios for some reason. What do ?

>>61264522
How did you partition the drive?

Your partition should be properly aligned for ssds to squeeze maximum speed.
I once didn't do it and it fucked me up.

To automatically properly align your partitions you should use 'parted' with '-a optimal' flag. For example:
parted -a optimal /dev/nvme0n1 mkpart primary 0% 100%

Note that this is a single command, you don't open parted interface.

>>61265704
Go to the boot menu in your bios' advanced settings and select the drive there. If it's still not there, unplug all other drives, boot to desktop once, then reconnect drives.

>>61265763
>>61264522
Yeah, that time is ridiculously slow, something is wrong. I have an nvme drive too, which is not even Samsung (it's Toshiba, almost 1k slower) and my system boots in like under 2-3 seconds after I press enter in Grub. Unfortunately, I don't have access to my computer right now, so I can't check the exact time and post 'systemd-analyze'.

>>61265787
Thanks anon

>>61264522
>autisming over boot times
amazing

 5.882s NetworkManager-wait-online.service
          4.799s apt-daily.service

wut do

>>61267927
update, they fixed some of that

You're booting userspace in six seconds dude, that's what the SSD does. You've got hardware that's slow to initialize

>>61267998
what package?

>>61267927
static ip

>>61268129
NetworkManager
check it against the dev's version

>>61268165
This, DHCP service is slow as fuck.

>>61264522
Can't you move /boot and the boot manager to a real SSD but keep your OS on that?

disk speed isn't the only factor, if it's waiting on other things like mounting raids, connecting to a network, etc, then it doesn't matter how much faster root is

>>61265763
This does however rewrite my partitions and erase my data right?

>>61265202
You should reorder few lines of UEFI code. Should be fucking easy, if you have sources. But we don't and manufacturers don't care much, so we're stuck with this crap.