Z170 Extreme4, m.2, etc. troubles

Edited by AlexF - 28 Feb 2016 at 9:31pm

AlexF wrote: I am having the same problem Christer has.  My read speeds are too slow (especially random read).  The speed jumps remarkably for a seconds when I put the computer to sleep and then wake it back up and immediately run the Samsung Magician benchmark test -- but then it drops to being slow again.  I installed all the latest chipset drivers and the Samsung NMVe driver.  My memory is running at 3200, I have a SATA SSD and SATA HDD plugged in as well. My setup is: GIGABYTE GA-Z170N-WIFI (rev. 1.0) LGA 1151 Intel Z170 HDMI SATA 6Gb/s USB 3.0 Mini ITX Intel Motherboard + Intel Core i5-6500 6M Skylake Quad-Core 3.2 GHz LGA 1151 65W BX80662I56500 Desktop Processor Intel HD Graphics 530 + G.SKILL Ripjaws V Series 16GB (2 x 8GB) 288-Pin DDR4 SDRAM DDR4 3200 (PC4 25600) Intel Z170 Platform Desktop Memory Model F4-3200C16D-16GVK + SAMSUNG 950 PRO M.2 256GB PCI-Express 3.0 x4 Internal Solid State Drive (SSD) MZ-V5P256BW

Thanks for this information. Sorry you have this problem, but that it happens on another manufacture's mother board makes this issue a potential problem with the 950 Pro itself. I'm NOT saying I know this is true, but if we see more cases of this on other boards, that puts the drive itself into doubt.

I am having the same problem Christer has.  My read speeds are too slow (especially random read).  The speed jumps remarkably for a seconds when I put the computer to sleep and then wake it back up and immediately run the Samsung Magician benchmark test -- but then it drops to being slow again.  I installed all the latest chipset drivers and the Samsung NMVe driver.  My memory is running at 3200, I have a SATA SSD and SATA HDD plugged in as well.

Edited by AlexF - 28 Feb 2016 at 11:38am

netjrb wrote: Hi Parsec, What BIOS version are you using? Thanks.

Sorry, I meant to reply to your question a day or two ago...

I'm still using 2.50, with my Z170 Extreme7+ board.

I've never had the random hangs in Windows you mentioned, and have run the Magician software many times, no problem. I've used almost every UEFI version our board has, including some of the no longer available Betas. I have an early production board that came with 1.40.

Hard for me to say which version is the "best", since I never thought I had problems in Windows related to the UEFI version.

But, I would suggest version 2.10 due to its update of the Intel RAID Option ROM and EFI driver. The EFI driver is used by PCIe SSDs, and the RAID Option ROM update has bug fixes for PCIe SSD RAID support.

The IRST driver (version 14) that provides RAID support for PCIe SSDs is in its infancy, and while IMO it is not as good as it could/should be, I can't criticize it too much. It's a one of its kind software for PCs, and manages to work with multiple protocols (SATA and NVMe) at the same time. Quite an accomplishment by Intel, and provided free to us, if we purchase the appropriate platform.

We cannot forget ASRock in this as well, since they provide most of the few boards that let us use this capability.

Christer703 wrote: Okay so i have found the source of the problem for me.  After a lot of testing i found out that by removing 2 of the 4 ram sticks (HyperX Fury 2666MHz 4x4GB), my ultra m2 port runs at PCIe 3.0 mode with full speeds all the time working exactly as its suppose to do.  So either it is something with the bios that causes this problem or it is something wrong with the motherboard. I have tested the Ram so this is not the issue alone, however there might be a problem with this exact type of ram in combination with this motherboard. I am in contact with Asrock Support about this and will post the answer as soon as i get it.

Thanks for posting what you discovered. That is quite strange that your memory caused the problem, I don't have an explanation for it... yet. The memory controller is in the CPU, and the M.2 slot uses the DMI3 lanes from the Z170 chipset. Whatever interaction (or bug) is involved is more complex than meets the eye.

I tried creating a RAID 0 array of SATA SSDs on my Z170 Extreme7+ board to see if that caused the problem. It doesn't seem to make any difference on my board, and I am using two M.2 SSDs, a 950 Pro and an SM951.



This is a very special RAID 0 volume, composed of a Samsung 840 Pro and a Crucial M4, two completely different SSDs that share nothing besides being SSDs. The result surprised me, I ran the benchmark twice with the same general results:



The 4K write speed is surprising, and does not seem to be a fluke. The M4 limits the sequential write speed, but otherwise this is a good result, with excellent 4K write speed.

Don't let anyone tell you that mixing two different SSDs in RAID 0 is not a good idea. Frankly I've never done this before now, I'm still not sure how I feel about it...