• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

The RX Vega 64 Owners Thread

kundi said:
Clearly that card is Liquid Cooled. What settings/drivers are you running? Several of us are having constant crashes(especially with 17.8.2) with our 64 AIO cards while gaming on anything above balanced mode while using good PSUs(mine is a Seasonic 1000w Prime Platinum).


How do you know your HBM is 1050mv? Wattman and other tools show 950mv for my AIO 64.
Click to expand...

I'm using the Vega 64 AIO.

Settings and drivers used mentioned in the YouTube description. Core is overvolted by 0.050mv, so running at 1.200v.
 
kundi said:
Thanks. Not an option for me. Gaming at +50% causes a system crash within 5-10 minutes. I've already filed an RMA but since my retailer doesn't have this card in stock, it will just be refunded.
Click to expand...
Could be faulty, my first card crashed when using anything but Power saver. Had two cards since then and both okay.
 
carlpls said:
IMO you should return it.

The way I think of it is like this: I'm paying good money for a graphics card, and I want it to work perfectly out of the box, if it doesn't, it should be fixed. If it were some cheap whatever thingy, I wouldn't really care, but this is quite a bit of money we're talking.
Click to expand...

I think I will, don't want take any risk and as you said for the price paid you would expect that it works normally straight out of the box.
 
@Matt.

Using hwinfo64 with Rtss for overlay (I wanted to make sure my core voltage was applying) to which all this time it looks like it's not. No matter what change I make to core voltage, the overlay doesn't budge from 1.356v which I'm guessing is down to the +50% power. Obviously at that high voltage I'm not to happy and need to figure out why it's not altering.

Can you help Mate? Any ideas?

Something not right at all here. No matter if I chose Power Saving, Balanced, Turbo, Custom and change my core voltage, nothing makes any difference to the GPU Core Voltage VDDC, just stays at 1.356 no matter what is going on.

 
Last edited:
kundi said:
Clearly that card is Liquid Cooled. What settings/drivers are you running? Several of us are having constant crashes(especially with 17.8.2) with our 64 AIO cards while gaming on anything above balanced mode while using good PSUs(mine is a Seasonic 1000w Prime Platinum).


How do you know your HBM is 1050mv? Wattman and other tools show 950mv for my AIO 64.
Click to expand...

Sorry, meant that the GPU was stock, but the HBM was not. I manually changed it to [email protected] within Wattman. It does start at 950mV for AIO
 
iihavetoes said:
Sorry, meant that the GPU was stock, but the HBM was not. I manually changed it to [email protected] within Wattman. It does start at 950mV for AIO
Click to expand...
Thanks for the info. My card crashes if I set the memory to 1050mv. I have a new 64 AIO inbound anyway since my current card crashes when setting it to anything above balanced mode. On 17.8.2 even balanced mode crashes the card.
 
Last edited:
TonyTurbo78 said:
@Matt.

Using hwinfo64 with Rtss for overlay (I wanted to make sure my core voltage was applying) to which all this time it looks like it's not. No matter what change I make to core voltage, the overlay doesn't budge from 1.356v which I'm guessing is down to the +50% power. Obviously at that high voltage I'm not to happy and need to figure out why it's not altering.

Can you help Mate? Any ideas?

Something not right at all here. No matter if I chose Power Saving, Balanced, Turbo, Custom and change my core voltage, nothing makes any difference to the GPU Core Voltage VDDC, just stays at 1.356 no matter what is going on.

Click to expand...

Memory and core voltage are the wrong way around, it's a bug which i reported to HWINFO64. Start paying attention to HBM voltage as thats core voltage and vice versa.

I'm using AMD internal software which can verify voltages and core voltage is working, but current available reporting tools are buggy.
 
AMDMatt said:
Memory and core voltage are the wrong way around, it's a bug which i reported to HWINFO64. Start paying attention to HBM voltage as thats core voltage and vice versa.

I'm using AMD internal software which can verify voltages and core voltage is working, but current available reporting tools are buggy.
Click to expand...

How is there 1.356v going to the HBM memory if in Wattman it's set to 1.050mv?

And also how come I can drop my core voltage to 800mv mid bench but not crash out?

So let's get this straight. In Wattman the core and memory voltage is the correct way around but in hwinfo64 it's the other way around?

Can you take a pic of your Wattman please?
 
Last edited:
TonyTurbo78 said:
How is there 1.356v going to the HBM memory if in Wattman it's set to 1.050mv?

And also how come I can drop my core voltage to 800mv mid bench but not crash out?

So let's get this straight. In Wattman the core and memory voltage is the correct way around but in hwinfo64 it's the other way around?

Can you take a pic of your Wattman please?
Click to expand...

I'm referring to HWINFO, it reports HBM voltage as actual Core voltage.

You can find my Wattman settings in some of my YouTube videos, but i only increase core voltage to 1.200v.

If you can drop your voltage that low and not crash then it's working.
 
AMDMatt said:
I'm referring to HWINFO, it reports HBM voltage as actual Core voltage.

You can find my Wattman settings in some of my YouTube videos, but i only increase core voltage to 1.200v.

If you can drop your voltage that low and not crash then it's working.
Click to expand...

Ok makes sense. One thing im not getting though. If HWinfo is displaying things vice versa, that the 1.356v that's being displayed in HWinfo is actually the memory voltage. My question as above is, how is there that much voltage going to memort when in Wattman it's only set to 1050mv?
 
I don't think that memory voltage in Wattman is the actual voltage for thé HBM chips, i think it's something else which has to with memory. Would love to get an explanation on what it is exactly and how does it work since in my case increasing it is the only thing which makes my card stable.
 
Memory voltage appears in Wattman on Polaris it does nothing.

I suspect on VEGA it does nothing.

HBM1 was supposed to be 1.2V but I guess as yields/quality was subpar AMD needed to use 1.3V for it for all 'samples' to hit 500MHz. When I undervolt HBM1 at stock clocks, say knock off 25mV in lengthy testing of several hours it will have issues.

HBM2 again is same story IMO. 1.35V is used by VEGA FE and RX VEGA 64. RX VEGA 56 is lower, as clocked lower, GN/buildzoid measured as 1.3V on DMM but in VBIOS it's 1.25V.
 
Last edited:
Thanks guys explains a lot.

I'm amazed I can do 1650mhz core clock with an 800mv undervolt.

How come also on the AIO cards, the memory voltage defaults at 950mv where as its 1050mv on the air versions? Now I'm on water is it worth me dropping the HBM voltage to 950mv down from the 1050mv?
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom