Yeah, it still restarted after I uninstalled the nvidia drivers. The weirdest thing is... usually if your computer restarts due to a power issue or force restarting or pulling the plug, it displays a message on the screen when it starts back up saying the computer wasn't shut down properly. That doesn't happen. The computer boots back up like as if nothing happened.
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.
Mac Pro 5,1 with Sapphire RX 580 random restarts?
- Thread starter Jayson A
- Start date
- Sort by reaction score
Yeah, it still restarted after I uninstalled the nvidia drivers. The weirdest thing is... usually if your computer restarts due to a power issue or force restarting or pulling the plug, it displays a message on the screen when it starts back up saying the computer wasn't shut down properly. That doesn't happen. The computer boots back up like as if nothing happened.
Doubt it's tied to the NVIDIA web driver, BUT if you're uninstalling, make sure you clear all the NVIDIA files in system LaunchAgents and LaunchDaemons. These are not always removed correctly.
I'm just wondering if you ever did that clean install on a test drive you were advised to do and considering?
Also, if the thing worked perfectly when your AMD card was installed alone, I suspect there could be a case where the problem is with the video card itself. It may be unlikely but perhaps macOS is tickling something in the hardware that doesn't work quite right, whereas Windows doesn't. It wouldn't be the first time a nix based OS has exposed a hardware failure that WinX was oblivious too.
For that matter, it may have exposed a failure elsewhere that only occurs under very specific conditions.
Just thinking aloud...
Also, if the thing worked perfectly when your AMD card was installed alone, I suspect there could be a case where the problem is with the video card itself. It may be unlikely but perhaps macOS is tickling something in the hardware that doesn't work quite right, whereas Windows doesn't. It wouldn't be the first time a nix based OS has exposed a hardware failure that WinX was oblivious too.
For that matter, it may have exposed a failure elsewhere that only occurs under very specific conditions.
Just thinking aloud...
Last edited:
Yeah, funny thing is. I was playing Rocket League in macOS and it didn't freeze, but doing something like browsing the web or very little usage was making it freeze.
Tonight, I'm going to do a fresh install just in case. I usually keep my OS install pretty light, so I'd be surprised if something is screwed up with my software but anything is possible I guess.
So weird that Windows doesn't have an issue. It's also weird that there are lots of people with the same exact setup as me running the Sapphire PULSE RX 580 without issues on macOS and Windows, so I guess it could be a faulty card... who knows.
[doublepost=1527873254][/doublepost]Can I use the card in slot 2? Would there be any difference in performance?
Correct, but for open fan type card. The card will run quite a bit warmer if installed in slot 2. Seems the PCIe compartment's air flow is really decided to handle the GPU in slot 1.
I tested this myself. The GPU run few degrees C cooler in slot 1 with lower fan speed.
I can certainly try it. As long as my computer doesn't decide to shut itself off, I don't care if the GPU runs a little bit warmer. First I'm going to format and reinstall macOS though. One thing at a time ;-)
Since you mentioned about "Would there be any difference in performance?". Warmer means less boost, or more throttling. So, there will be some difference in performance, but most likely not noticeable.
have you tried a clean install of osx yet?
any one know how to read crash reports?
sounds like it's, not GPU (as problem with both) and no problem in osx
i still think as you have no problems in windows it's a software problem.
it may be ram so worth trying like you are
ie pull all but 2 sticks, see if any problem
if no problem add two more sticks and see if any problem
if no problem add two more sticks
and so on until you get a shutdown, if shutdown with first two sticks swap them and try agen
etc..
but may just be software, and as no problems on windows im not shore if it is the ram or not but may be.
iv had a background app/demon/task that i had installed (then it had updated it'selff) that gave me random shutdowns, spent weeks till i found out what did it.
any one know how to read crash reports?
sounds like it's, not GPU (as problem with both) and no problem in osx
i still think as you have no problems in windows it's a software problem.
it may be ram so worth trying like you are
ie pull all but 2 sticks, see if any problem
if no problem add two more sticks and see if any problem
if no problem add two more sticks
and so on until you get a shutdown, if shutdown with first two sticks swap them and try agen
etc..
but may just be software, and as no problems on windows im not shore if it is the ram or not but may be.
iv had a background app/demon/task that i had installed (then it had updated it'selff) that gave me random shutdowns, spent weeks till i found out what did it.
Clean install of 10.12.6 didn't work. Took out RAM, still crashed. Tried PCIe slot 2 and still crashed. Updated to high Sierra, so far no crashes on HS. I'll keep watching it.
Have you ever blown the dust out of your cMP? Especially the power supply, blowing it out from the CD caddy area. IMO this sounds like dust is Causing an issue or something is not getting enough power.
You could also flash a custom firmware for your rx580 to lower its power requirements.
You could also flash a custom firmware for your rx580 to lower its power requirements.
The freezes always happen when I'm barely doing anything. I don't know how the card would be drawing too much power when it's basically at idle.
So far High Sierra has been solid. Haven't had a restart in a while. Maybe Sierra wasn't as compatible with the Sapphire PULSE RX 580 as I initially thought.
So I have been watching this thread because I have a similar story. Thought I would share my experience too. I installed a RX580 Pulse 8GB in slot#1 and moved my stock GT120 to slot #2. This worked fine for several uses across 3 weeks. Then the first time I did any elongated use, just using the Dolphin emulator, for longer than like 30mins....I started getting restarts out of nowhere. The audio and graphics would act funny for a few seconds before the restart happens as a bit of a warning. I was concerned that it was overheating as I thought I smelled and felt that my case was warmer than normal. I brushed it off as a weird one time event because no message or prompt came up stating anything once it booted back up. After the second time it happened I went into the log reports and did have something there from each event. If I were to post it would anyone be able to interpret this kernel panic or whatever the term is to describe what happened? I cannot decipher what the log is able to tell me about the issue.
When I installed the RX580 I noticed the fan blades would catch on the back corner to the GT120. I was thinking like what the hell, how are you supposed to use two GPUs like this then? I had to stick a small piece of bamboo stick, just something a couple millimeters wide to prop the two GPUs apart, so slightly pushing the RX580 down and the GT120 up. This allowed the RX580s fans to spin freely but due to what happened I believe it was not good enough for acceptable air flow and cooling. So after I got the two random restarts I put the GT120 in slot#4 and have not had a problem since. I did not have any monitoring software when the restarts happened so I am not positive it was overheating. I have not seen the system go over 98*F since making the change with the GT120 going to slot #4. I only use OSX so I cannot speak to how it worked with Windows. I am on a MP 4,1 to 5,1 running OSX 10.13.4 for the record too
So I do think I have currently resolved my issue, but if anyone could help with interpreting the log I would feel much better after knowing what exactly caused the issue in the first place so I can take any preventative measures from having repeat events.
When I installed the RX580 I noticed the fan blades would catch on the back corner to the GT120. I was thinking like what the hell, how are you supposed to use two GPUs like this then? I had to stick a small piece of bamboo stick, just something a couple millimeters wide to prop the two GPUs apart, so slightly pushing the RX580 down and the GT120 up. This allowed the RX580s fans to spin freely but due to what happened I believe it was not good enough for acceptable air flow and cooling. So after I got the two random restarts I put the GT120 in slot#4 and have not had a problem since. I did not have any monitoring software when the restarts happened so I am not positive it was overheating. I have not seen the system go over 98*F since making the change with the GT120 going to slot #4. I only use OSX so I cannot speak to how it worked with Windows. I am on a MP 4,1 to 5,1 running OSX 10.13.4 for the record too
So I do think I have currently resolved my issue, but if anyone could help with interpreting the log I would feel much better after knowing what exactly caused the issue in the first place so I can take any preventative measures from having repeat events.
Blocking the GPU's fan is a big big matter.
I tested this when I have two HD7950. If the 1st one block the 2nd one's fan. The slot 1 card can easily run 20C warmer with much higher fan speed. We should really avoid this. Just a few mm gap for the fan to suck air in is not enough.
https://forums.macrumors.com/thread...mac-with-2-d700s.1732849/page-5#post-21722712
Sorry to burst your bubble, but I have been using my pulse rx580 for over 9 months in sierra without any problems.
It is good news though that hs upgrade seems to have resolved your issue...
yea I was worried about that and my initial response was to move the GT120 to a different slot to allow for better air flow for the RX580. I am a noob with GPU upgrades though so I wasn't 100% sure if that was the issue or not. To clarify, the whole RX580 was not covered. The rear fan was like 80% exposed and that is why it was able to catch on the rear corner of the GT120 (holy cow did that make some serious racket when it did!). Still it seemed sketchy to me but not knowing any better I thought a x16 GPU should be in a x16 PCIe slot. I got thinking the GT120 is just being used to power a monitor and not being used for any real GPU tasks so going to a x4 slot wasn't going to slow down any gaming needs for me. I do appreciate the link from your own overheating experience! That is interesting that going from a x16 slot to a x4 slot has such a minimal performance penalty. If I knew that I would have put it in slot #4 right from the start. So thanks for clarifying I was likely going to fry my GPU if I left it there.
Sorry to burst your bubble, but I have been using my pulse rx580 for over 9 months in sierra without any problems.
It is good news though that hs upgrade seems to have resolved your issue...
Yeah it was the weirdest thing. I even formatted the drive and installed a fresh version of 10.12.6 on it. I hadn't even started to move my files to it when BOOM I got an auto reboot.
I was prompted for a firmware update right before I installed High Sierra though, could the firmware update be why it's not crashing now? Should I try Sierra again or will the firmware be downgraded when Sierra is installed?
It must be the 580 doing something because I never had a problem before it, and it's never been a problem in Windows... even when I ran the benchmark games.
I tried taking the ram out one pair at a time, I pulled all hard drives out except for my boot drive. I put the 5770 back in and it didn't restart (although that could've been a coincidence). I tried a fresh macOS install. I tried resetting the NVRAM and the SMC. I tried running disk repair on the drive and it said it was OK.
I watched my temps and they never really went very high. I have the same exact model computer at work and I was getting the same temps and that one never shuts down on me.
My last thing to try was installing a different OS so I went with High Sierra. I haven't had a shutdown since. Weird right?
By the way... is there a small chance that my PULSE RX 580 has a newer firmware on it than yours and perhaps there's just something about it that Sierra doesn't like?
Before getting the RX 580, I made sure that it would be 100% compatible with Sierra and everyone said it was, but for some reason not mine. Perhaps it's defective, but I can't return it because High Sierra fixed it and Windows doesn't have a problem, so I can't rule out software being the cause of the restarts.
I'm thinking about re-installing 10.12.6 again. I just don't understand why High Sierra doesn't have an issue and Sierra does.
Do you have a spare disk, that would be the easiest way.
FYI I have the new fw installed on my mac also, I installed HS a while back on a different drive to have a look, but have stayed on sierra.
The fw will not be removed...
What's the new firmware version?