RX Vega Owners Thread

I undervolted my WC/LC Vega64 20-37mv and that lowers temps a touch and helps maintain boost clocks. The 20-37 part is the difference between Wattman and Afterburner. Afterburner seems to be a bit more stable using lower clocks hence the lower 37mv part.

Wattmann is a true piece of junk. Absolutely can not believe AMD even has it. Shit never holds clocks. Always goes to balanced mode after computer restarts. Might as well stop wasting time on wattman and let third party do the job. AMD has wasted a lot of time on features that are not needed and they can actually focus on other improvements. Wattman is one of those features with wasted time and effort. Let afterburner and trixx do their thing and just eliminate wattman.
 
650w should be good enough. but you guys with the sff need to realize that not all sff cases can support the absolute best/hottest/most power hungry parts. there needs to be a trade off somewhere...

I've run on overclocked Intel quad and up to two GPUs on my 650w Seasonic, which is currently happily plugging away at the system in my sig (4.5GHz 6700k, OC'd 1080Ti). I don't even know what the wattage requirements are and I don't care; this 1080Ti only has an 8-pin and a 6-pin for power.

That mean that the GPU can only pull- according to spec, remember when AMD got in trouble here?- 75w from PCIE + 75w from the 6-pin + 150w from the 8-pin. That's all of 300w! Scary!
 
You run a little girl's specs. Why don't you keep it quiet already. Real man run 1KW and Vegas. :D

Lol- well, I did once run a pair of HD6950's on this PSU. Probably the highest draw it's seen in its 7 year life- honestly haven't really seen the need to go higher. Maybe when 4k120 panels come out, and I need a second 1080Ti?

:D
 
Why don't you stop?

Please, please tell us that there's no difference between a solid Seasonic and a trashy Rosewill.

Why are you talking about a trashy Rosewill when the very person that started this debate is using a brand new silverstone powersupply?

You seem to think that just because you are able to do to, everyone should be able to, completely ignoring the fact that each computer system, each person has different variables involved. Oh wait..... You you are running a 1080ti. Do you even have any Real experience with running a Vega 64 in your system, or are you just here to troll with little real long term life experience on the subject other than "I can so everyone can" mantality?
 
Why are you talking about a trashy Rosewill when the very person that started this debate is using a brand new silverstone powersupply?

You seem to think that just because you are able to do to, everyone should be able to, completely ignoring the fact that each computer system, each person has different variables involved. Oh wait..... You you are running a 1080ti. Do you even have any Real experience with running a Vega 64 in your system, or are you just here to troll with little real long term life experience on the subject other than "I can so everyone can" mantality?
look, this seems to be getting a little heated. everything i looked at (beside amds page - that was a mistake evidently) suggested that 650w would have been fine to run vega. if that isnt the case then it isnt the case and ill have to look at getting an sfx-l. Thank you guys for at least providing some insight
 
look, this seems to be getting a little heated. everything i looked at (beside amds page - that was a mistake evidently) suggested that 650w would have been fine to run vega. if that isnt the case then it isnt the case and ill have to look at getting an sfx-l. Thank you guys for at least providing some insight

Sorry. I just know that white paper numbers verses reality are two completely different things. White paper is under perfect conditions. Reality we don't live in perfect condition. Some people are lucky that get by with white paper requirements. Some don't.


Power supply requirements are not listed higher than what the white paper shows because of quality of power supply. They are listed high to cover real world variables that effect power draw and/or power output. Clean power/dirty power, heat, various hardware configurations etc.

It is not AMD's or any other manufactures responsibility or fault if some one chooses to go low quality or high qualility, and they do not set requirements based on that.

You may very well live in an area, or it could be your fuseboxes that is delivering dirty power which will make even the best power supplies work harder to deliver the rated wattage. If you are pushing it to the limit and this happens, it can't work any harder to keep up because it is already at it's maximum capacity.

I have dirty power where I live, and why I have a business class ups, to keep the power stable and clean.

Anyhow, yes there are people who like to go with the bare minimums and get by.. as in your case, it doesn't work.

Heck, you could have a power cable that is flawed causing more resistance than it is supposed to pulling more power.. God only knows. Again, every single person has different variables that can effect the results.

Anyhow I wish you luck.
 
Except the very reason the discussion came up is it isn't, when a person with a 650 watt power supply (high quality, brand new) has to run his card in power saving mode to avoid system shut down and other crashes/issues, making it clear it is a power issue,, and tries to blame the card and AMD, all while ignoring their minimum recommended requirements.

But.. each to there own. If a person wants to cripple his investment by running it in power saving mode and not being allowed to use it at it's full potential because the person doesn't want to admit it is a power issue, or others trying to argue it couldn't possibly be the power supply, so be it. Not my problem.

I have never had an issue do to power, but then again, I never try to get by with what should be enough, and always get above the minimum requirements per the manufactor. Becuase what "should be enough" does not take in account for all the variables that can effect the draw on a power supply, or the output of the power supply. All it takes for a "should be enoigh" power supply to be a " not enough" power supply is live in an area with a bad power grid that is not effectively supplying clean power.

Not to drag down this thread, but I think you're mixing it up - I'm the one running in power saving mode on my Vega and I've got a 450w SFX PSU. I think the guy with the 650w has a Vega 64
 
Not to drag down this thread, but I think you're mixing it up - I'm the one running in power saving mode on my Vega and I've got a 450w SFX PSU. I think the guy with the 650w has a Vega 64

You are correct, I did get you and NuclearLeamons confused. But it still does not change the fact that if you are running a below recommended powers supply, and having issues, that the power supply is the most likely the cause, even if others are succeeding. (I am pretty sure I have explained in enough detail the reasons why already).

Thanks for letting me know I got the two of you mixed up. :D
 
Well I just want to chime in to say that I appreciate each and every one of you'll. I hope you'll get along and find the time to write a Vega 56 / 64 OC guide that we can sticky at the top of the AMD subsection. :)
 
Anyone here flash the Vega 56 with a 64 bios? I like the idea of an overclocked Vega 56 without having to play with Wattman.
 
I did it on two cards

it worked on one, and didn't on the other.

What happened with the one that did not work? Did it simply fail to flash or did it fail during the flash and you had to do a recovery of some sort? Thanks.
 
What happened with the one that did not work? Did it simply fail to flash or did it fail during the flash and you had to do a recovery of some sort? Thanks.
The good 56 Vega card flashed to 64 and back without issues on either bios.
The Vega card that was bad flashed to 64 and got worse, so I flashed it back -- but it was acting up long before I flashed it. (screen corruption, crashes, system hangs, etc).

I'm still in the process of RMA'ing the bad card. Before I ever flashed it I occasionally got red screens of death, screen corruption, and system hangs, and crossfire profiles (with the two 56 cards) giving me like 10FPS on games that were supposed to support crossfire. With the flash the card was completely unusable for anything 3d.

Generally speaking - my understanding is you can flash it to 64 bios without concern of actually hurting the card - you should be able to flash back without issue assuming your card isn't bad to begin with. I don't know that I would flash it to 64 and then try to overclock as well, because the 64 version of the bios is already a overclock and overvolt of sorts, and I don't know if AMD did binning on the Vega chips and you probably shouldn't push your luck -- until more information is widely available on if it's safe or not. I did not overvolt/overclock my cards in the 64 bios version for that reason.

You can get the various BIOS files from Tech Power Up.
https://www.techpowerup.com/vgabios...om/vgabios/?manufacturer=AMD&model=RX+Vega+64

Then you flash with ATI Flash utility
https://www.techpowerup.com/download/ati-atiflash/

You MAY need to use the ATI pixel patcher from there to ensure the driver is recognized:
https://www.monitortests.com/forum/Thread-AMD-ATI-Pixel-Clock-Patcher

Of course run all of these as Administrator.

P.S.
DO NOT USE GPU-Z TO SAVE OR FLASH YOUR VEGA BIOS
 
Last edited:
The good 56 Vega card flashed to 64 and back without issues on either bios.
The Vega card that was bad flashed to 64 and got worse, so I flashed it back -- but it was acting up long before I flashed it. (screen corruption, crashes, system hangs, etc).

I'm still in the process of RMA'ing the bad card. Before I ever flashed it I occasionally got red screens of death, screen corruption, and system hangs, and crossfire profiles (with the two 56 cards) giving me like 10FPS on games that were supposed to support crossfire. With the flash the card was completely unusable for anything 3d.

Generally speaking - my understanding is you can flash it to 64 bios without concern of actually hurting the card - you should be able to flash back without issue assuming your card isn't bad to begin with. I don't know that I would flash it to 64 and then try to overclock as well, because the 64 version of the bios is already a overclock and overvolt of sorts, and I don't know if AMD did binning on the Vega chips and you probably shouldn't push your luck -- until more information is widely available on if it's safe or not. I did not overvolt/overclock my cards in the 64 bios version for that reason.

You can get the various BIOS files from Tech Power Up.
https://www.techpowerup.com/vgabios...om/vgabios/?manufacturer=AMD&model=RX+Vega+64

Then you flash with ATI Flash utility
https://www.techpowerup.com/download/ati-atiflash/

You MAY need to use the ATI pixel patcher from there to ensure the driver is recognized:
https://www.monitortests.com/forum/Thread-AMD-ATI-Pixel-Clock-Patcher


Of course run all of these as Administrator.


I got my vega 56 for 469.99 and had some amazon credits and I had the same issue. The card just wasn't stable. I was kicking it with drivers but seriously unstable and crashes running benches. Saw the vega 64 for 499.99 and seeing that it would only cost me 30 more at the end I ordered that. And it has been rock stable. actually i an down volt it a little and still maintain close to 1600 boost with power limit increase. For the price i love it. I don't game much so power isn't an issue for me. I got it because of freesync 4k monitor I have and I can say that I haven't had any free sync issues either. Power wasn't a concern for me for few hours of gaming a week. But i wanted the best experience for the price and it fit the budget.
 
Well, the flash worked, thanks. Now I just have to get the time to mess around and see how everything goes, this is one of the reasons I like AMD reference cards.
 
Well, it appears that now that I have flashed my Vega 56 with the 64 bios, I can run the HBM at 1GHz without issue. Also, I do not really need to run the fan at 4000 rpm's because it appears the 3000 rpm's nets me the same score with 3D Mark Firestrike and 3D Mark 11. I lose maybe 1fps average in the benchmark for Gears of War 4 and the fan noise is not bad at all, just a whoosh type hush sound. (Power limit set to 50+ and the power draw is the same as my Vega 56 overclocked settings.)
 
Well, it appears that now that I have flashed my Vega 56 with the 64 bios, I can run the HBM at 1GHz without issue. Also, I do not really need to run the fan at 4000 rpm's because it appears the 3000 rpm's nets me the same score with 3D Mark Firestrike and 3D Mark 11. I lose maybe 1fps average in the benchmark for Gears of War 4 and the fan noise is not bad at all, just a whoosh type hush sound. (Power limit set to 50+ and the power draw is the same as my Vega 56 overclocked settings.)

That's pretty good news for the IHV's putting out custom cards!
 
Except the very reason the discussion came up is it isn't, when a person with a 650 watt power supply (high quality, brand new) has to run his card in power saving mode to avoid system shut down and other crashes/issues, making it clear it is a power issue,, and tries to blame the card and AMD, all while ignoring their minimum recommended requirements.

But.. each to there own. If a person wants to cripple his investment by running it in power saving mode and not being allowed to use it at it's full potential because the person doesn't want to admit it is a power issue, or others trying to argue it couldn't possibly be the power supply, so be it. Not my problem.

I have never had an issue do to power, but then again, I never try to get by with what should be enough, and always get above the minimum requirements per the manufacturer. Because what "should be enough" does not take in account for all the variables that can effect the draw on a power supply, or the output of the power supply. All it takes for a "should be enoigh" power supply to be a " not enough" power supply is live in an area with a bad power grid that is not effectively supplying clean power.

Hey guys, I am finally back from my trip out west (well Ive been back for 2 weeks but was out of town for a few more poker tournaments) and am getting caught back up. I am currently using the latest driver's without a hitch, although I haven't played anything other then Shadow of Mordor (maxed @ 4K) with a buttery smooth FreeSync enabled 65-70FPS...I am working my way through the game so I can get the second one on sale during the Steam Winter sale.

I want to address NWRMidnight and his concern about power draw. You are doing something wrong with your power calculations. If you are reading the usage at the wall as you claim, then the meter is beyond inaccurate, as most Kill-a-watt type devices are. Hell most UPS are not very accurate.

I am running my sig system below, and a VEGA 56 flashed to 64 @ 1750(sustained)/1100Mhz HBM @ 1.15V and 75% Power Limit. With my CPU loaded with 7 threads of an AVX load (Prime95) and the last thread driving the GPU stress program, I have never seen any draw above ~565W. This is with a CPU that draws 235W with Prime95, a 28W pump running full speed, 2 SSDS, 1HDD, 4 GT 120s running full tilt and 2 20mm fans running full tilt. If i turn my fans off, I cut 19W off that mentioned 565W. Let's subtract the 28W from the pump, and we are now at a total 519W (at the wall measured with a $850 professional meter).

Now, before you start lecturing me on the differences in silicon variance etc please realize I have a Bachelor's in Electrical Engineering, I just play poker for a living instead of working in an office. I am running my 56 @ ~35C under a sustained load, and I realize that the air cooled cards are going to draw more power due to their higher operating temperatures.

I think it is safe to say that there isn't anyone out there besides a few other hardcore OC'ers running their VEGA's with the unlocked Power Limit I am, thus my power draw should be a worst case scenario. I believe you are just trying to help, but in this case there is zero need for a single GPU system with a modern quadcore (or a super efficient Ryzen 6/8 core) to need more then 650W even with an OC on CPU and the GPU. If you are the type to worry, then a quality 750W would give you plenty piece of mind.
 
Hey guys, I am finally back from my trip out west (well Ive been back for 2 weeks but was out of town for a few more poker tournaments) and am getting caught back up. I am currently using the latest driver's without a hitch, although I haven't played anything other then Shadow of Mordor (maxed @ 4K) with a buttery smooth FreeSync enabled 65-70FPS...I am working my way through the game so I can get the second one on sale during the Steam Winter sale.

I want to address NWRMidnight and his concern about power draw. You are doing something wrong with your power calculations. If you are reading the usage at the wall as you claim, then the meter is beyond inaccurate, as most Kill-a-watt type devices are. Hell most UPS are not very accurate.

I am running my sig system below, and a VEGA 56 flashed to 64 @ 1750(sustained)/1100Mhz HBM @ 1.15V and 75% Power Limit. With my CPU loaded with 7 threads of an AVX load (Prime95) and the last thread driving the GPU stress program, I have never seen any draw above ~565W. This is with a CPU that draws 235W with Prime95, a 28W pump running full speed, 2 SSDS, 1HDD, 4 GT 120s running full tilt and 2 20mm fans running full tilt. If i turn my fans off, I cut 19W off that mentioned 565W. Let's subtract the 28W from the pump, and we are now at a total 519W (at the wall measured with a $850 professional meter).

Now, before you start lecturing me on the differences in silicon variance etc please realize I have a Bachelor's in Electrical Engineering, I just play poker for a living instead of working in an office. I am running my 56 @ ~35C under a sustained load, and I realize that the air cooled cards are going to draw more power due to their higher operating temperatures.

I think it is safe to say that there isn't anyone out there besides a few other hardcore OC'ers running their VEGA's with the unlocked Power Limit I am, thus my power draw should be a worst case scenario. I believe you are just trying to help, but in this case there is zero need for a single GPU system with a modern quadcore (or a super efficient Ryzen 6/8 core) to need more then 650W even with an OC on CPU and the GPU. If you are the type to worry, then a quality 750W would give you plenty piece of mind.

Why push it though? It is your choice but, I have found that pushing up against the upper limit of any power supply is not always the best idea. I went from an 850 Watt to 1KW back when I was running an FX8300 at 4.5 Ghz and 2 x Sapphire Furies so now, I am on that 1KW Seasonic power supply with a 1700X at 3.75GHz and 1.25v and a Vega 56 flashed with the Vega 64 bios and power limit set to 50% and I am all set to go for years to come.
 
Why push it though? It is your choice but, I have found that pushing up against the upper limit of any power supply is not always the best idea. I went from an 850 Watt to 1KW back when I was running an FX8300 at 4.5 Ghz and 2 x Sapphire Furies so now, I am on that 1KW Seasonic power supply with a 1700X at 3.75GHz and 1.25v and a Vega 56 flashed with the Vega 64 bios and power limit set to 50% and I am all set to go for years to come.

I'd drop a second 1080Ti into my build below without a second thought. 650W? No worries.

[I mean, I *might* worry, PSU is actually pretty old, but you get the point, 750W and I wouldn't- if you read up, I had a pair of HD6950's on this thing :D ]
 
On a HTPC I built years ago, with a 6870, I got a 500W PSU because it was just enough and I was trying to go cheap with the rig.

When it came around to upgrading, when the 200 series came out, I ended up not being able to get the card I wanted due to not having enough power.

I ended up with a 280X, which was still a great card and I got some use out of it, but I really wanted a 290.

After that experience, I never go cheap on the power supply. I will get at least 1000W for my main rigs (also because of multi-GPU) and 750W for my HTPC/SFF rigs.
 
Hey guys, I am finally back from my trip out west (well Ive been back for 2 weeks but was out of town for a few more poker tournaments) and am getting caught back up. I am currently using the latest driver's without a hitch, although I haven't played anything other then Shadow of Mordor (maxed @ 4K) with a buttery smooth FreeSync enabled 65-70FPS...I am working my way through the game so I can get the second one on sale during the Steam Winter sale.

I want to address NWRMidnight and his concern about power draw. You are doing something wrong with your power calculations. If you are reading the usage at the wall as you claim, then the meter is beyond inaccurate, as most Kill-a-watt type devices are. Hell most UPS are not very accurate.

I am running my sig system below, and a VEGA 56 flashed to 64 @ 1750(sustained)/1100Mhz HBM @ 1.15V and 75% Power Limit. With my CPU loaded with 7 threads of an AVX load (Prime95) and the last thread driving the GPU stress program, I have never seen any draw above ~565W. This is with a CPU that draws 235W with Prime95, a 28W pump running full speed, 2 SSDS, 1HDD, 4 GT 120s running full tilt and 2 20mm fans running full tilt. If i turn my fans off, I cut 19W off that mentioned 565W. Let's subtract the 28W from the pump, and we are now at a total 519W (at the wall measured with a $850 professional meter).

Now, before you start lecturing me on the differences in silicon variance etc please realize I have a Bachelor's in Electrical Engineering, I just play poker for a living instead of working in an office. I am running my 56 @ ~35C under a sustained load, and I realize that the air cooled cards are going to draw more power due to their higher operating temperatures.

I think it is safe to say that there isn't anyone out there besides a few other hardcore OC'ers running their VEGA's with the unlocked Power Limit I am, thus my power draw should be a worst case scenario. I believe you are just trying to help, but in this case there is zero need for a single GPU system with a modern quadcore (or a super efficient Ryzen 6/8 core) to need more then 650W even with an OC on CPU and the GPU. If you are the type to worry, then a quality 750W would give you plenty piece of mind.
well, as the person who started the whole debacle (my apologies) , I believe I have found the cause of my power issues and it was neither the psu or gpu. In fact, it turned out to be the supply to the room itself. The dorm was built back in the 60s and the whole grid gets overloaded during peak hours. I tried the same system at a different local and haven't had a single shutdown. I do, unfortunately have another fault that has crept up and id hope the community here could lend a hand. Once or twice a day, the fan on my Vega 64 spins up and i loose all HDMI signal. reconnecting it to the gpu or monitor does nothing and i have to restart the pc in order to get the signal back. Anyone faced a similar issue? i haven't had such problems on that monitor previously with my older graphics cards.
 
I do, unfortunately have another fault that has crept up and id hope the community here could lend a hand. Once or twice a day, the fan on my Vega 64 spins up and i loose all HDMI signal. reconnecting it to the gpu or monitor does nothing and i have to restart the pc in order to get the signal back. Anyone faced a similar issue? i haven't had such problems on that monitor previously with my older graphics cards.

well.. I've having somewhat similar issues regarding the outputs of the cards, the cause may be similar and the solution too so, everything mentioned below to avoid all the writing again:

i've been having weird issues with these drivers, gona try olders, but from time to time my R9 390X and Vega64 machines refuse to display anything from the GPU when starting up. I have to plug the output from my mobo to another monitor and wait until windows start entirely and just press reboot it then will work again however that's really Annoying. it has happened now three times on each machine.

found the issue.. both cards are dropping both PCI-E Lanes and PCI-E Generation, Found that when the card doesn't boot up after restart/turn on it's because it is in a somewhat "disabled" state, with the second monitor plugged in Mobo Port I can see that the PCI-E 1 slot it's recognized as PCI-E 1.1@x0, forcing the mobo to set the PCI-E x16 slot to Gen3 fix part of the issue, the machine it's able to display from the GPU and up to the moment it have restarted just fine, did it several times, and power off/power on several cicles too, also tested turning off the PSU several times simulating a power shortage (even If I have UPS) and the machine has booted just fine, however the PCI-E Lanes are just a mess sometimes it boot as PCI-E X16@x4... sometimes x8, sometimes x1, sometimes full x16, I've tried resetting the cards and changing from PCI-E Slots and the behavior still persist.

Ran DDU in safe mode, reinstalled previous drivers and they has been working so far without issues.. I can live with the issue as much as the cards can stay at full X16 Gen 3 as I normally keep my personal machine turned on sometimes for weeks, however one of those machines is used by my brother and he always turn off the PC when go out of the house and at nights so the issue may be exponentially worse for him.
 
well.. I've having somewhat similar issues regarding the outputs of the cards, the cause may be similar and the solution too so, everything mentioned below to avoid all the writing again:
hmm, similar but not quite the same issue as its been happening during use not at startup
 
Hey guys, I am finally back from my trip out west (well Ive been back for 2 weeks but was out of town for a few more poker tournaments) and am getting caught back up. I am currently using the latest driver's without a hitch, although I haven't played anything other then Shadow of Mordor (maxed @ 4K) with a buttery smooth FreeSync enabled 65-70FPS...I am working my way through the game so I can get the second one on sale during the Steam Winter sale.

I want to address NWRMidnight and his concern about power draw. You are doing something wrong with your power calculations. If you are reading the usage at the wall as you claim, then the meter is beyond inaccurate, as most Kill-a-watt type devices are. Hell most UPS are not very accurate.

I am running my sig system below, and a VEGA 56 flashed to 64 @ 1750(sustained)/1100Mhz HBM @ 1.15V and 75% Power Limit. With my CPU loaded with 7 threads of an AVX load (Prime95) and the last thread driving the GPU stress program, I have never seen any draw above ~565W. This is with a CPU that draws 235W with Prime95, a 28W pump running full speed, 2 SSDS, 1HDD, 4 GT 120s running full tilt and 2 20mm fans running full tilt. If i turn my fans off, I cut 19W off that mentioned 565W. Let's subtract the 28W from the pump, and we are now at a total 519W (at the wall measured with a $850 professional meter).

Now, before you start lecturing me on the differences in silicon variance etc please realize I have a Bachelor's in Electrical Engineering, I just play poker for a living instead of working in an office. I am running my 56 @ ~35C under a sustained load, and I realize that the air cooled cards are going to draw more power due to their higher operating temperatures.

I think it is safe to say that there isn't anyone out there besides a few other hardcore OC'ers running their VEGA's with the unlocked Power Limit I am, thus my power draw should be a worst case scenario. I believe you are just trying to help, but in this case there is zero need for a single GPU system with a modern quadcore (or a super efficient Ryzen 6/8 core) to need more then 650W even with an OC on CPU and the GPU. If you are the type to worry, then a quality 750W would give you plenty piece of mind.

Hahaha.. you are so full of it. The fact that your first response on the matter, you have to throw out your degree and credentials is just a ploy to make it appear you know what you are talking about, shows you don't have a clue. If you did, you wouldn't need to throw out your "credentials" to back it up.

Lets use your very own number to show that your information is either wrong, or you are not being straight up with your performance and/or clock speeds, and/or you are not being honest about your power draw.. or maybe your $850 piece of equipment is broke. (BTW, throwing that out is also another ploy to make people believe your false information)

You state, after taking off the pump wattage and fans, you draw 519 Watts total.... and your cpu under full load pulls 235 watts.. lets subtract that, which leaves 284 Watts you claim your Vega is pulling (actually less if you remove hard drive, motherboard, memory, etc power draw) under prime 95, and graphic stress test (btw, I seriously doubt 1 thread (half a core) will put your Vega under full load, it will substantially bottleneck your Vega, even at your clock speed). Considering you are running it at the same clock speed as what LQ Vega 64 (1750 Mhz), which btw only averages 1663MHz (not the 1750 you claim your sustains) and pulls 350+ watts, air cooled averages 1513 Mhz and pulls 290 watts, and stock Vega 56 pulls 210 watts averaging 1433 Mhz, all on default settings and directly measured. ( https://www.pcper.com/reviews/Graph...4-Vega-64-Liquid-Vega-56-Tested/Clocks-Power- ) you can bet your ass your card is pulling north of 350 watts when under full load, for your card alone, specially with +75 power limit and if substaining 1750 Mhz as you claim, which means you have at minimum 66 watts watts of power coming out of thin air, and that is before you account for hard drive, motherboard/memory, etc. power draw.

I am also running 4 hard drives, 2 SSD, 10 fans cooling 3 radiators, 2 pumps, 6 core/ 12 thread Xeon overclocked from 3.06 to 4.34.. blah blah blah.. point is, your claims are bs. Your own numbers show it to be impossible. I suspect that GPU-Z (gpu power draw only) at the speeds/settings you claim will verify you are pulling 350+ watts alone under full load.

Go run Tom Clancy's rainbow six, it uses all my cores/threads, and puts the cpu at 70% usage, my card runs at 99% at 1663 Mhz, pulling 320+ watts.. You will pull less cpu power, but a substantially more GPU power than your prime 95/1 thread gpu stress tess.
 
Last edited:
Hahaha.. you are so full of it. The fact that your first response on the matter, you have to throw out your degree and credentials is just a ploy to make it appear you know what you are talking about, shows you don't have a clue. If you did, you wouldn't need to throw out your "credentials" to back it up.

Now, you state, after taking off the pump wattage and fans, you draw 519 Watts total.... and your cpu under full load pulls 235 watts.. lets subtract that, which leaves 284 Watts you claim your Vega is pulling (actually less if you remove hard drive, motherboard, memory, etc power draw) under prime 95, and graphic stress test. Considering you are running it at the same clock speed as what LQ Vega 64 (1750 Mhz), which btw only averages 1663MHz (not the 1750 you claim your sustains) and pulls 350+ watts, air cooled averages 1513 Mhz and pulls 290 watts, and stock Vega 56 pulls 210 watts averaging 1433 Mhz, all on default settings and directly measured. ( https://www.pcper.com/reviews/Graph...4-Vega-64-Liquid-Vega-56-Tested/Clocks-Power- ) you can bet your ass your card is pulling north of 350 watts when under full load, for your card alone, specially with +75 power limit and if substaining 1750 Mhz as you claim, which means you have at minimum 66 watts watts of power coming out of thin air, and that is before you account for hard drive, motherboard/memory, etc. power draw.

I am also running 4 hard drives, 2 SSD, 10 fans cooling 3 radiators, 2 pumps, 6 core/ 12 thread Xeon overclocked from 3.06 to 4.34.. blah blah blah.. point is, your claims are bs. Your own numbers show it to be impossible. I suspect that GPU-Z (gpu power draw only) at the speeds/settings you claim will verify you are pulling 350+ watts alone under full load.
Dude back it down a notch. If you back off the voltage, which he did, he will in fact be using less watts. I have mine mining @100-120W. When I game I am prob ~350W but I haven't changed my voltage there. And I can definitely maintain 1750+ and have so your assessment is innaccurate.
 
Dude back it down a notch. If you back off the voltage, which he did, he will in fact be using less watts. I have mine mining @100-120W. When I game I am prob ~350W but I haven't changed my voltage there. And I can definitely maintain 1750+ and have so your assessment is innaccurate.

Vega does not work that way, when you under volt, it decreases your max clock you can hit when you under volt because even if your card can hit higher clocks, it won't pull more power than what you set in the P states + power limit. Mining does not stress the card and uses substantially less power then gaming, and he is talking about his VEGA 56 with bios flashed to Vega 64 sustaining 1750 Mhz.. cough cough.. yea. You are also not claiming you are pulling 286 watts during during stress test at 1750+, and even admit you probably pull 350 watts. I even gave the direct measured power consumption of stock cards. My vega 64 pulls 392 Watts (GPU only) per afterburner, GPU-Z, and AMD's new monitoring that came out today in the newest drivers (I am pretty sure it's accurate) running 3dmark ultra stress test sustaining 1663 Mhz, mind you, mine is a air cooled vega 64 converted to water and flashed. It can't go above 1720 Mhz and stay stable no matter what (temps never go over 45 C and hot spot hits 78 C under full stress), the reason I have it set the way I do, because of how the Vega controls clock speeds, if I give it enough power, it will go spike to 1782 core in rainbow six siege and crash. It hits 1720 core in rainbow siege the way I have it set, and pulls during gaming 330 watts, which is less than stress testing, because it is not always under load. Of course, if you don't have Thermal head room, volts don't matter if you become temperature throttled.


I also just did the "stress test" in the same manner he claims (prime95 on 11 threads, and gpu stress on 1 thread - superpostion benchmark set for stress test).. my card bounces all over for core clock from 1000 mhz to 1600 averages about 1250 mhz and 74% utilization running prime 95 and gpu stress test in that manner, limiting it to 1 thread like that, and it is a slide show stuttering mess. The minute I stop prime 95, or allow the GPU stress test to use all threads, it shoots up to 1679 and 100% load. And what do you know, power measurements at the wall is is less when testing in this manner. Maybe because the card isn't even coming close to it's full power draw because it isn't able to hit it's max clocks because of the CPU/Thread limitation when testing in this manner.

Also, something I did not state in the previous reply: IF a UPS is not accurate, how can one trust it to protect your equipment properly? It has to be accurate to properly protect your equipment. The exception would be if you bought a cheap, bottom line, no name brand.. That's why I chose a business class top of the line one.
 
Last edited:
Vega does not work that way, when you under volt, it decreases your max clock you can hit when you under volt because even if your card can hit higher clocks, it won't pull more power than what you set in the P states <snip>

<snip>

Also, something I did not state in the previous reply: IF a UPS is not accurate, how can one trust it to protect your equipment properly? It has to be accurate to properly protect your equipment. The exception would be if you bought a cheap, bottom line, no name brand.. That's why I chose a business class top of the line one.
A)Vega works just like any other card. Better silicon, which he may have, could have plenty of headroom at lower voltages. Lower voltages uses less power so you wont hit your power limit even if you manage to max out the clocks that are stable at that voltage. Pascal usually never throttling max boost clocks is a perfect example of this concept in action
B)A UPS having stable voltage regulation has jack to do with how good a ammeter it has on the AC lines. They need one just good enough to provide a reasonable estimate of remaining battery runtime and what not so they wont spend too much on them
 
Well,

Just loaded up hellblade with my new ups running. CPU is a ryzen 1600 at 3.7GHz with 1.2v, memory is at 2933. Vega is sustaining 1550 at 1025 voltage with hbm at 1070 and 920 +50%. Have an old mcp655 powering the water loop with 2x 120mm fans, and a swiftech all in one with a 140mm on the cpu. Also have two SSDs and two 140mm fans in the front of the case.

UPS says 540 watts lol.
 
Ramble Ramble.


You have been a member here how long? I don't need to prove anything to you, there are plenty of members on this board that can vouch for me in a personal and professional capacity. In your rush to prove me wrong, you forgot a very important point in that wall of text. I am running a 75% PowerLimit, which combined with the lower voltage and the fact my card stays under 35C under load (and I believe its a very good sample just as JustReason has) allows my card to stay at 1750Mhz all day every day.

I've actually been able to sustain 1775Mhz clocks but I have to bump the voltage up the point it is useless for a less then 1% improvement. If you don't believe me, look at my FireStrike Ultra and Superposition scores, they are in VEGA Owner's thread. Then go find some results on the web showing any other VEGA significantly out performing my card. I will be here waiting ;).
 
Last edited:
Just retested with sustained 1500 1025 voltage tbm at 1070 and 920 and -10% and it was consistently at 390watts with some spikes up to 415.
 
Back
Top