M.2 NVME on Maximus X Hero

[SherardG]

Hello experts,

I am looking to fully utilize my M.2 NVMe on my Maximus X Hero. Currently, I have a 960 EVO 500GB which is slotted in the M.2_1 slot (with heatsink) and the max I get before setting everything to gen3 PCIe is 2176 MB Read (can't remember write), and after enabling everything in "advanced options" to Gen3, and explicitly stating for the M.2_1 slot to use PCIe from Auto, I am getting a mere 3000 Read / 1600 Write (using CrystalDisk Mark 6)

Am I doing anything wrong? What other settings should I change? I must mention that I am running XMP profiles and am unsure if anything was changed by that.

Only the M.2_2 slot has option to choose x4 mode from x2, should my card be here instead?

I would greatly appreciate your help.

[MasterC]

HI SherardG,

What's the capacity of your 960 EVO?
The 250GB max out at 3200/1500MB/s, while the 500GB at 3200/1800MB/s. and the 1TB at 3200/1900MB/s.

Are you comparing to Raid 0 setups?

How good is the airflow, even with a heatsink on the M.2 it can throttle, especially when you keep running benchmarks.

[Starscream]

After focusing on my CPU speeds, I had to revert back to default BIOS settings. I re-enabled XMP, but did not set anything else. (I forgot about the settings you mentioned.)

So you made me curious. I just ran Samsung Magician, and it says performance results 3137 / 2010. It also says the interface is PCIe Gen. 3x4. Also noteworthy was a firmware update applied today, after just updating when new last week.

Did you peel the backing off of the thermal tape that was pre-installed on the heatsink? I was surprised to learn that the blue tape was indeed thermal tape. I removed the label from my SSD and the backing from the blue tape and installed.

Edit* I have a pro, but hope this is helpful. Perhaps mine is under performing as a pro since I'm not sure what the performance should be.

[SherardG]

MasterC@ASUS wrote:
HI SherardG,

What's the capacity of your 960 EVO?
The 250GB max out at 3200/1500MB/s, while the 500GB at 3200/1800MB/s. and the 1TB at 3200/1900MB/s.

Are you comparing to Raid 0 setups?

How good is the airflow, even with a heatsink on the M.2 it can throttle, especially when you keep running benchmarks.

Hello MasterC,
My EVO has 500GB on it. I don't get those figures running Samsung Magician; I read on a thread some threads here that it doesn't report the correct figures, in fact, I am only able to get the 3000mb+ figures from CrystalDisk; I pretty much stopped doing benches since they eat up the lifespan of my SSD.

No, I am not comparing to a RAID0 set up; and I have no plans of running RAID as well. The airflow has a lot of negative air pressure on it. On idle, sensors report them as 37C. Under load after 2 hours of gaming, I get 51C max. My case has a lot of negative airflow which I am trying to turn into positive a/f. I'm currently running 3x120mm fans as intake at full RPM, I have 1 140mm exhaust fan and 2 140mm push-exhaust fans for my radiator. They run on a curve and at idle runs at 39% speed (700rpm).

Starscream wrote:
After focusing on my CPU speeds, I had to revert back to default BIOS settings. I re-enabled XMP, but did not set anything else. (I forgot about the settings you mentioned.)

So you made me curious. I just ran Samsung Magician, and it says performance results 3137 / 2010. It also says the interface is PCIe Gen. 3x4. Also noteworthy was a firmware update applied today, after just updating when new last week.

Did you peel the backing off of the thermal tape that was pre-installed on the heatsink? I was surprised to learn that the blue tape was indeed thermal tape. I removed the label from my SSD and the backing from the blue tape and installed.

Edit* I have a pro, but hope this is helpful. Perhaps mine is under performing as a pro since I'm not sure what the performance should be.

I might try updating my firmware as well. I did peel off the adhesive protector for the themal tape on the heatsink. What I did not do however, was peel off the label on my SSD. Thanks for your inputs; firmware might have resolved the magician issues, since a lot of people running a 960PRO get numbers which aren't far off from mine using the bundled software, but are blazing fast and hit their rated speed with CrystalDisk.

[Silent_Scone]

Figures seem fairly ballpark to me. Don't forget the max theoretical speed is exactly that, and you're close to it. Try updating the firmware as you say.

[SherardG]

Silent Scone wrote:
Figures seem fairly ballpark to me. Don't forget the max theoretical speed is exactly that, and you're close to it. Try updating the firmware as you say.

Hello Silent Scone, everyone,

Thanks for all your feedback. I've done one more bench after upgrading firmware and results stay more or less the same, though the results in Samsung Magician are worse of from what I've got. Since I really have no complaints (was just trying to get my money's worth!) I've decided against running any more tests to keep my TBW's to actual usage rather than from bench marking.

Just one more thing; should I also remove the label on the stick itself and directly stick the thermal pad to the stick? I just removed the blue thermal pad leaflet.

[Menthol]

Leave the sticker on, you would need it in case of RMA, it also has a thin copper layer as heatsink

[SherardG]

Menthol wrote:
Leave the sticker on, you would need it in case of RMA, it also has a thin copper layer as heatsink

Just saw your reply now, but thanks! I did think that it would void warranty but wasn't sure.

[Starscream]

Menthol wrote:
Leave the sticker on, you would need it in case of RMA, it also has a thin copper layer as heatsink

There is another label on the the other side of the SSD which I think has the info related to RMA.

I'm undecided if the copper label style heatsink residing between the thermal tape for the MB heatsink and ICs would be a benefit, since there is another layer of paper or some other material on top of it. It appears to me to that the copper label heatsink is best used in standalone applications only. But, I'm no expert on these matters!

[Menthol]

I am sure removing the sticker will improve contact between the thermal pad and chips by a small amount, whether it's a measurable amount can only be determined by testing with and without removing it.
You would see the biggest temperature difference with active cooling