m.2 905p optane

oleNBR

Limp Gawd
Joined
Dec 9, 2016
Messages
371
anyone know how this will work in it's m.2 format? normally 905p is 12v but m.2 is only 3.3v, which means current involved is well over 4-5 Amps. I can see if intel changes controller to that similar of 800p to reduce power consumption greatly and uses way less, would that work?
 

Abula

Gawd
Joined
Oct 29, 2004
Messages
961

Attachments

Last edited:

oleNBR

Limp Gawd
Joined
Dec 9, 2016
Messages
371

oleNBR

Limp Gawd
Joined
Dec 9, 2016
Messages
371
probably $500-550 for the 380gb m.2 version. or about $600 if they charge a premium which I think they will, m.2 22110 is meant for enterprise.
 

MickeyBailey

Limp Gawd
Joined
Jun 5, 2018
Messages
142

Abula

Gawd
Joined
Oct 29, 2004
Messages
961
Maybe an m.2 stick with a SATA power connector?
its very unlikely, the link that oleNBR posted has Intel rep saying it will be 3.3v so its like any other m.2 drives, my only doubt is how good will it be, thinking that both the 2.5/PCIe require a more massive size, but maybe its scale down in everything, size and performance, but will keep my eye on the Q3 release to see if its performs as its big brothers.
 

oleNBR

Limp Gawd
Joined
Dec 9, 2016
Messages
371
its very unlikely, the link that oleNBR posted has Intel rep saying it will be 3.3v so its like any other m.2 drives, my only doubt is how good will it be, thinking that both the 2.5/PCIe require a more massive size, but maybe its scale down in everything, size and performance, but will keep my eye on the Q3 release to see if its performs as its big brothers.
i wouldnt doubt that performance will be a cut below. maybe lower sequential or random performance, more so on sequential read/write those prob takes most power imho. problem is this also comes with 7 channel so its meant for full 7 channel sequential performance like the u.2 or PCIE version. im not sure where they can really cut power to make it work with 3.3v, imho the highest m.2 SSD i've seen uses like 3 Amps which is crazy, that comes to around 10w from m.2 at 3.3v.

if they can reduce the 14-16w U.2 version to just 10w then its doable.
 

MickeyBailey

Limp Gawd
Joined
Jun 5, 2018
Messages
142
i wouldnt doubt that performance will be a cut below. maybe lower sequential or random performance, more so on sequential read/write those prob takes most power imho. problem is this also comes with 7 channel so its meant for full 7 channel sequential performance like the u.2 or PCIE version. im not sure where they can really cut power to make it work with 3.3v, imho the highest m.2 SSD i've seen uses like 3 Amps which is crazy, that comes to around 10w from m.2 at 3.3v.

if they can reduce the 14-16w U.2 version to just 10w then its doable.
Perhaps the controller is a new model that consumes less power. Or they might just throttle the device when power consumption hits the threshold.
 

MickeyBailey

Limp Gawd
Joined
Jun 5, 2018
Messages
142
so far September, ending Q3 and still not m.2 905p..
There is not a big market for Optane right now so I’m not surprised no new products are releasing. They need gen 2 to roll out with increased density before the chips are remotely profitable.
 

brentsg

[H]ard|Gawd
Joined
Mar 8, 2003
Messages
1,694
905p is a NVME SSD, for the PCIe version it grabs power from the PCIe slot, and the 2.5 i seen two versions (maybe there are more),

Intel 480GB Optane 905P U.2 Internal SSD SSDPE21D480GAX1 - Comes with u.2 cable

Intel 480GB Optane 905P U.2 Internal SSD SSDPE21D480GAM3 - Comes with m.2 cable adapter

IIRC both grab power from a SATA POWER connector.
I know this is old, but I just wanted to mention that Intel sells the U.2 to M.2 adapter cable separately. I grabbed the U.2 version on sale, then picked up the cable.
 

oleNBR

Limp Gawd
Joined
Dec 9, 2016
Messages
371
I know this is old, but I just wanted to mention that Intel sells the U.2 to M.2 adapter cable separately. I grabbed the U.2 version on sale, then picked up the cable.
where can u buy this cable? also they said there will be a m.2 22110 for 780gb 905p, possibly next year at CES
 

Abula

Gawd
Joined
Oct 29, 2004
Messages
961

crmb

n00b
Joined
Oct 4, 2011
Messages
38
Why there is a such big heatsink on the 480GB model when the 380GB model is bare ?

Price are not good in EU now (and it is not even distributed in my country). I have to pass.
I like the 480GB model because it remind me of my WD Velociraptors.
 
Last edited:

Kaos_Drem

Gawd
Joined
Oct 16, 2004
Messages
960
Interesting, specially the larger length, maybe doesn't loses much performance, that said, pricing is really close to the standard PCIe/U2. Only downside i see atm is that length, some motherboard might not have a m.2 slot with that requierement, and its very likely that laptops will not.
You would not want to put something that pulls 11W into a laptop, sounds like a recipe for battery life disaster.
 

oleNBR

Limp Gawd
Joined
Dec 9, 2016
Messages
371
You would not want to put something that pulls 11W into a laptop, sounds like a recipe for battery life disaster.
my laptop wont be meant for battery use. it'll mostly be plugged in. battery is served as a purpose similar to that of UPS.
 

oleNBR

Limp Gawd
Joined
Dec 9, 2016
Messages
371
When are we going to see benchmarks on these?
picking mine up in a few days.. so we shall see. they said its full 7 channels so performance wise not much different from U.2 905p with exception that laptop M.2 go through chipset/dmi so theres that additional latency wheres desktop you have the ability to use PCIe lanes directly from the CPU.
 

Brahmzy

Supreme [H]ardness
Joined
Sep 9, 2004
Messages
4,955
^^ Pretty freakin awesome random reads. Still pissed about Smeltdown patches. Kinda neutered my 3 900ps a bit. I can feel the difference from pre-patch days.
Fighting the patches is kinda pointless at this point.
 

oleNBR

Limp Gawd
Joined
Dec 9, 2016
Messages
371
^^ Pretty freakin awesome random reads. Still pissed about Smeltdown patches. Kinda neutered my 3 900ps a bit. I can feel the difference from pre-patch days.
Fighting the patches is kinda pointless at this point.
can you not revert the patch? the security flaws are most likely not going to impact consumers especially the issue has been around forever.

also 905p is thorugh chipset, could get another maybe 8-10% if it's direct PCIe from CPU.
 

Brahmzy

Supreme [H]ardness
Joined
Sep 9, 2004
Messages
4,955
At some point we’ll all have to be on a version of W10 that includes it in the OS and the BIOS includes it...
 

oleNBR

Limp Gawd
Joined
Dec 9, 2016
Messages
371
At some point we’ll all have to be on a version of W10 that includes it in the OS and the BIOS includes it...
thats probably when intel has 10nm finally. i dont plan to do that cause zen2 will have 64 cores for their wx chips for the same price intel will have it for only 20 cores
 
Joined
May 23, 2017
Messages
45
I'm waiting for an enthusiast review site to get one, and compare to the U.2 version and other high end M.2 drives. No purchase until then.
 

oleNBR

Limp Gawd
Joined
Dec 9, 2016
Messages
371
I'm waiting for an enthusiast review site to get one, and compare to the U.2 version and other high end M.2 drives. No purchase until then.
the performance is pretty much exactly the same as 905p U.2 version, cause i own both.

both are also better than 900p cause i owned 900p as well, 905p in general (U.2 or m.2) are around 5-10% faster in terms both sequential and random read/write so it is a nice boost, im still hoping the sequential can be faster though, but problem is heat which will not work with m.2 if it goes any faster. intel needs to have the controller on a better node or just a new controller all together to reduce power by 30%.

only downside to m.2 version is if the m.2 slots on motherboard goes through chipset then its added latency.
 

JargonGR

Limp Gawd
Joined
Dec 16, 2006
Messages
492
the performance is pretty much exactly the same as 905p U.2 version, cause i own both.

both are also better than 900p cause i owned 900p as well, 905p in general (U.2 or m.2) are around 5-10% faster in terms both sequential and random read/write so it is a nice boost, im still hoping the sequential can be faster though, but problem is heat which will not work with m.2 if it goes any faster. intel needs to have the controller on a better node or just a new controller all together to reduce power by 30%.

only downside to m.2 version is if the m.2 slots on motherboard goes through chipset then its added latency.
What motherboard are you using? The U.2 version of 905p is on my purchase list but I have found not enough data about its performance vs the PCI-E version. A user on ROG forum reports that after trying everything it is slower than the PCI-E 905px and he is quite experienced on this since the launch of x299 - I am talking about the Rampage VI Extreme motherboard and the U.2. set-up through the CPU.

There should be no performance difference when the U.2. is going through the CPU but I need to confirm that before buying this to upgrade my 960Pro. I am only interested in it Random 4K of course and if not performing to the max it would be a waste.


BTW - Do you feel the difference between a Samsung 960 Pro and this when launching applications like Adobe Bridge (cache) / Illustrator / Photoshop / Premier / After Effects / Autocad, etc? Is it noticeable under heavy multitasking or when using Virtual Machines?
 

Brahmzy

Supreme [H]ardness
Joined
Sep 9, 2004
Messages
4,955
4x-6x 4K low-qd random read performance is very noticeable on an OS drive. Opening apps, surfing, mail, Windows updates, general usage etc. this is where my 900ps shine.

It is not noticeable for large block sequential workloads, whereas my 970 Pros are. 970’s sequential performance trumps the Optane, but cannot begin to touch the Optane in the area I mentioned.

Only you know what workloads you’re likely going to see the benefits with which storage.
 
Top