That's strange. Can you try switching the highpoint to a different PCIe slot ie. 1 or 3? How about trying to install highpoint's macos drivers with the RX 580 installed to see if there is any difference?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Can a Mac Pro boot up to Mojave with a non metal card?
- Thread starter HyperX13
- Start date
- Sort by reaction score
That's strange. Can you try switching the highpoint to a different PCIe slot ie. 1 or 3? How about trying to install highpoint's macos drivers with the RX 580 installed to see if there is any difference?
I bought a heat sink from Amazon and we will see how it works. But here is my issues. I used to build PCs for over a decade and I know about thermal paste and CPU coolers, etc. Well these Samsungs have stickers on top of the chips that say if you remove them you void warranty. So how would a heat sink work with a sticker on there? There would be hardly any heat transfer. Anyway - I will still stick it on..
As an Amazon Associate, MacRumors earns a commission from qualifying purchases made through links in this post.
It's a metallized sticker made for interface with a thermal pad.
I’ve played musical slots already. Made no difference. I’ll look into the drivers, but I doubt that is the problem. No one else is using them that I’m aware of. It’s worth a shot though.
Do you have DriveDx installed? Do you know what firmware your 970 is running? Apparently there is newer firmware installed on some brand new 970 models, but it's not available as a firmware download/update (for Mac or Windows). Don't know if this would play a factor into your issue.
I have the magician software installed in Windows, it reports Latest Firmware is installed. Where can I find out mor information about this?
Wouldn't be available through magician software and nothing has been released by Samsung as an update. Just appears to be shipping on SOME new purchases. Noticed two different versions on DriveDx when troubleshooting a boot issue on a client machine. They were 500GB blades and thought were identical units, but they were obviously not. Did not write down the exact numbers. One blade was presenting an issue with booting (High Sierra) on a PX1 and the other was not. Boot drive was cloned via CCC to each blade. Believe the machine was 4,1>5,1 and that was done by someone else. Firmware was 140.0.0.0.0.
Within DriveDx it will tell you the firmware version in the General Information section, usually under model. My personal 970 EVO 1TB reports as firmware 2B2QEXE7.
Within DriveDx it will tell you the firmware version in the General Information section, usually under model. My personal 970 EVO 1TB reports as firmware 2B2QEXE7.
Interesting, do you remember if the boot problem was with the older firmware or the newer?
Regarding the boot issue. Did it always fail, or was it inconsistent?
Not certain which one was the "fix" and was semi-inconsistent issue. Similar to the "failure to appear" for Apricorn Velocity Duo X2 issue where occasionally on cold boot one or both of the SSDs would not appear on desktop, but are accessible in finder/disk utility... but happened slightly more frequently than Apricorn issue. Just tried another blade as part of troubleshoot. Initially thought they installed incorrectly or failed to wipe/format the physical disk vs. container/volume.
So neither drive was being used as a boot drive then. They just didn't show up for the party on the desktop after cold booting?
In case FW version turns out to be a significant factor and more samples are useful, both my 1TB 970 EVO and 256GB 970 EVO currently have firmware 2B2QEXE7. No problems.
No, machine would randomly not start from cold boot with 970/PX1. Could not replicate each time. Total kill power, wait and restart "fixed" a few times. Then 5 or 6 times later would happen again.
[doublepost=1551213050][/doublepost]In case it's helpful (no need for DriveDx just for this) - System Report and under Hardware > NVMExpress > Apple SSD Controller > Samsung SSD EVO 1TB it shows the firmware version under REVISION.
I have experienced the issue on my own machine with Apricorn Velocity Duo X2 as external/media drives, but have not experienced this happening with the 970 NVMe/PX1 boot drive. Impacts Apricorn Duo drives only. System always seems to boot fine. My machine is an authentic 5,1 and setup is with GPU in slot 1 (GTX 1080 FE right now), PX1 in slot 2, Apricorn Duo in slot 3, and (right now) USB3 in slot 4 with HS 10.13.6 17G5019.
Do not recall the Apricorn Duo issue before the 138/140 firmware. Did not happen more than a fluke (once or twice) when Apricorn Velocity Solo X2 was boot drive with SATA SSD. Instances have absolutely increased since 138/140 firmware, but not more than a dozen or so times and always "fixes" itself with a power down and cold restart. NVMe/PX1 was installed near time of 140 firmware.
Suspected it could be a power on distribution issue, but I have nothing to verify that. Also cannot pinpoint firmware actually causing the issue since it's not repeatable. The near randomness makes it almost impossible to troubleshoot. Worse than describing a sound to a mechanic and saying "fix this" and it not happening for them to observe.
Will likely install the 141 firmware when it is available publicly in case it fixes anything. I'm sure it will lack release notes.
No reports of the issue reoccurring on client machine since the blade swap. Will ask them to let me know, but no guarantees they actually will. It's basically a freelancer machine and overall gets treated poorly.
Hmm, this is the revision I have with my 2 970 Pros. I've never had a booting issue with the 970 Pro.
BTW I use SuperDuper.
Lou
Good news! I got it working and booted with the Samsung 960
It would not work at first so I ran this command
disktuil apfs updatepreboot disk1s1
I then rebooted it and it works!!!
Great speeds too
I also used a thermal heat sink and I am running at 49 - 50 degrees.
I am using Lycom dt120 PCIe card, I suspect the startech one would work too. I have it in slot 4 (furthest away from my video card). I just rebooted 5 times in a row and all is good. I still left my original 512 SSD drive in there in case of issues etc. I used Carbon Copy cloner to make the clone.
Do you have a heatsink on this one?
Oops... Sorry, just re-read your post and see that you are.
Oops... Sorry, just re-read your post and see that you are.
Last edited:
My 970 Pro 1TB has the same firmware revision as yours so it's definitely not the firmware. What could possibly be the link between the RTX 580 and nvme blade or highpoint card? Hmmmm.
As an added data point, I've experienced this a number of times the past month or so. Probably 3 times total. Start the computer, power light comes on, but hangs up before the monitors get powered. It's right before the fans spool up as I recall. When I press the power button the machine instantly powers off. Then wait... then reboot and all is normal. Instead of a hangup... the fans spool up and down and the monitors instantly come to life like normal.
I don't have an NVMe drive yet but I do have an RX580, Apricorn Solo X2, on Firmware 140.x.x.x and Mojave. So I wonder if it's related more to something else than the nature of an NVMe drive being present.
Curious - did you ever send your Apricorn in for firmware flash/update? Apricorn offered this several years ago for people who were running both Solo X2 and Duo X2 in the same machine. There was a conflict that firmware flash corrected. This was NOT a user installable update and card physically needed to be sent to Apricorn for the (free) service. These cards are discontinued and firmware update/flash services are no longer offered.
No never did. I just run the Solo X2. The startup-hangs definitely started since I upgraded to Mojave. But, there were so many changes... RX580, 140, Mojave. And very intermittent.
I'm on HS 10.13.6 with GTX 1080 FE, 140 firmware, Duo X2 and NVMe/PX1 boot drive... only common thing here is an Apricorn product present and the 140 firmware. Think this is an issue that started with either 138 or 140 firmware. May start to look more seriously into replacing Apricorn with NVMe RAID.