Not to my knowledge.Has there been any confirmation of where the gpu hotspot is?
Yeah I haven't seen it either. I guess it's hard to say where that is, unless that is maybe a VRM temp? Are all your VRM's cooled well?Not to my knowledge.
I was wondering if we can read the vrm temp?
Yes that can be the problem.Yeah I haven't seen it either. I guess it's hard to say where that is, unless that is maybe a VRM temp? Are all your VRM's cooled well?
Another thing, did you use the stock "X" backplate when mounting the morpheus? Or is there another backplate provided? That can help in increasing the mounting pressure.Yes that can be the problem.
I mean it looks weird that its the viscosity and pressure that should be the reason when so many people have similar problems with hotspot. Hmm.
When i set the target at 65 and max 85 i have perfect stability so its not really a problem so i will just wait until someone at reddit or ocuk finds the reason.
Tried both the stock x backplate provides by far most pressure. Running the morpheous provided now.Another thing, did you use the stock "X" backplate when mounting the morpheus? Or is there another backplate provided? That can help in increasing the mounting pressure.
Meh. Its all not working yet imo.Anyone figured out how to undervolt the HMB on Vega yet? As far as I can tell, only core voltage can be changed.
EDIT: NM, seems it's not possible to reduce HBM voltage at the moment. Lowering the Wattman value just makes it possible to lower the GPU voltage.
Check this out:Tried both the stock x backplate provides by far most pressure. Running the morpheous provided now.
Thanx mate. Will try it. Distributed the last time using fingers. Credit card is a good idea. Take a card with a lot of cost ! then try lower viscosity at the same time. Not more noctua anymore. LolCheck this out:
https://www.reddit.com/r/Amd/comments/74087c/i_lowered_the_hot_spot_temps_on_my_vega_56_by/
Not sure how much of a difference it makes, but I always put thermal paste by spreading it out over the whole core with a credit card, as the guy in that post did.
Have you tried that?
Check this out:
https://www.reddit.com/r/Amd/comments/74087c/i_lowered_the_hot_spot_temps_on_my_vega_56_by/
Not sure how much of a difference it makes, but I always put thermal paste by spreading it out over the whole core with a credit card, as the guy in that post did.
Have you tried that?
I have instability even when i reach hotspot of 95c.According to the developer of GPU-Z himself, the "GPU Hotspot" temperature is located directly inside of the GPU die. It is not a VRM temperature sensor.
- https://www.reddit.com/r/Amd/comments/716edj/rx_vega_56_morpheus_ii_hotspot_temp/
- https://www.techpowerup.com/forums/threads/devs-what-is-gpu-temperature-hot-spot-on-rx-vega.236843/
- https://cgit.freedesktop.org/~agd5f...hwmgr/vega10_thermal.h?h=amd-staging-drm-next
Heh, I saw that reddit post too, and I was about to link to it until I saw you did.
While the "GPU Hotspot" sensor is still a bit mysterious on this card, there have been multiple posts online where people have improved their Hotspot temperature simply by re-applying their thermal paste. These appear to be the general guidelines for achieving a good reading from this sensor:
Me personally? I used Gelid Extreme (which I've realized isn't 100% ideal), and I also used the "blob" style of application, where I allowed the pressure from the cooler to spread the compound by itself. If I had to guess, this is probably why my Hotspot temps still aren't that great. Under load, this sensor will creep over 80, maxing out at 85 or so. However, according to people who installed their cooler "properly", the temperature should be well beneath that.
- Use a runny, low-viscosity thermal paste. (Something like Arctic Silver 5 or Gelid Extreme is a bit too thick; use something like Arctic MX-2 instead.)
- Use a lot of it.
- Use the "spread" style application, where you manually spread it across the die first before mounting the cooler.
- As always, make sure you have proper airflow and cooling inside your case.
Granted; last time I mounted the cooler, I used a very generous amount of the thermal compound, as people from elsewhere have instructed. I used one considerably sized blob on the main GPU die, and two more smaller (but still pretty big) blobs on each HBM module. They're much bigger than the normal "pea" sized applications that people usually talk about. I think this has helped me a little. I also installed an additional front intake fan on my case, which may have also helped.
In my opinion: If your hotspot sensor is reaching 90 degrees C, then that's too high. And in my experience during the trial and error of installing this cooler, going over 100 will lead to instability.
Maybe eventually, I will remount my MORPHEUS cooler using the correct thermal paste and application method. Since the hotspot doesn't go over 90 though, I'm gonna leave it for the time being. Maybe next year when summer comes around and ambient temperatures go back up, I'll have to do it.
My other temperature readings are fantastic. GPU and Memory temperatures are only 60 and 70 degrees under load, respectively. It's weird that the "GPU Hotspot" reading is so much higher, but as other people have theorized, the Vega64 has multiple sensors around the card, and the Hotspot is simply showing the hottest part of the card, most likely on the bottom level of the die.
Anyway, a while ago I promised that I'd post pictures of the cooler. Sorry for the wait... I've been really lazy lately.
How is hotspot and stability going?According to the developer of GPU-Z himself, the "GPU Hotspot" temperature is located directly inside of the GPU die. It is not a VRM temperature sensor.
- https://www.reddit.com/r/Amd/comments/716edj/rx_vega_56_morpheus_ii_hotspot_temp/
- https://www.techpowerup.com/forums/threads/devs-what-is-gpu-temperature-hot-spot-on-rx-vega.236843/
- https://cgit.freedesktop.org/~agd5f...hwmgr/vega10_thermal.h?h=amd-staging-drm-next
Heh, I saw that reddit post too, and I was about to link to it until I saw you did.
While the "GPU Hotspot" sensor is still a bit mysterious on this card, there have been multiple posts online where people have improved their Hotspot temperature simply by re-applying their thermal paste. These appear to be the general guidelines for achieving a good reading from this sensor:
Me personally? I used Gelid Extreme (which I've realized isn't 100% ideal), and I also used the "blob" style of application, where I allowed the pressure from the cooler to spread the compound by itself. If I had to guess, this is probably why my Hotspot temps still aren't that great. Under load, this sensor will creep over 80, maxing out at 85 or so. However, according to people who installed their cooler "properly", the temperature should be well beneath that.
- Use a runny, low-viscosity thermal paste. (Something like Arctic Silver 5 or Gelid Extreme is a bit too thick; use something like Arctic MX-2 instead.)
- Use a lot of it.
- Use the "spread" style application, where you manually spread it across the die first before mounting the cooler.
- As always, make sure you have proper airflow and cooling inside your case.
Granted; last time I mounted the cooler, I used a very generous amount of the thermal compound, as people from elsewhere have instructed. I used one considerably sized blob on the main GPU die, and two more smaller (but still pretty big) blobs on each HBM module. They're much bigger than the normal "pea" sized applications that people usually talk about. I think this has helped me a little. I also installed an additional front intake fan on my case, which may have also helped.
In my opinion: If your hotspot sensor is reaching 90 degrees C, then that's too high. And in my experience during the trial and error of installing this cooler, going over 100 will lead to instability.
Maybe eventually, I will remount my MORPHEUS cooler using the correct thermal paste and application method. Since the hotspot doesn't go over 90 though, I'm gonna leave it for the time being. Maybe next year when summer comes around and ambient temperatures go back up, I'll have to do it.
My other temperature readings are fantastic. GPU and Memory temperatures are only 60 and 70 degrees under load, respectively. It's weird that the "GPU Hotspot" reading is so much higher, but as other people have theorized, the Vega64 has multiple sensors around the card, and the Hotspot is simply showing the hottest part of the card, most likely on the bottom level of the die.
Anyway, a while ago I promised that I'd post pictures of the cooler. Sorry for the wait... I've been really lazy lately.
It's a RM650x. Less than year old but never had any issues with it before. And Vega's more efficient than my old HD7950 when idle.^PSU?
Dont have any sorts of that problems in 2 Vega rigs. 56 and 64. And that's when pushing cheap 3000 ram to it's specs or above in 1700 and 1600x rig.Anyone have memory instability after installing Vega? I put a new Vega 56 into my system a few days ago and now all of my RAM is testing unstable. It's a ryzen build with a 1700, crosshair vi hero, and Samsung b-die yet it's unstable down to 2400, and even on a 2nd pair of ram that was stable before but not after putting in my Vega 56.
No crashes yet, it's just failing HCI memtest and google's stressapptest.Dont have any sorts of that problems in 2 Vega rigs. 56 and 64. And that's when pushing cheap 3000 ram to it's specs or above in 1700 and 1600x rig.
Sounds weird. Can you get info about the crash?