Gamers are ditching Radeon graphics cards over driver issues

Status
Not open for further replies.
Well, than the answer is not enough airflow in the case. Had all these cards with no issues :D
 
Anyone with a Radeon VII should really be using The Pro Enterprise Software drivers. Im on 19Q4 version and it plays all the games without any problems or "thermal runaway"
The strange thing about Navi drivers is that not everyone has issues. ;)

That is why I tend to think it might be more about cooling then real software issues at least for a good chunk of the reported issues. No doubt some are real software issues with specific games... every GPU vendor runs into that.

AMD went with a blower on their first cards cause they where clearly worried about heat issues.... I have a feeling most of the "software" issues with black screens, or crashes are related to heat. I know the 5700 I have if I where to unplug my 3 case exhaust fans... I have no doubt the inside of my case would turn into a nice warm fall day in hell. After playing Odyssey for a few hours... my top exhaust feels like a space heater. (and I know its cool to shit on AMDs software right now... but their game tailored fan curves work great. Nice option when only a couple games really push the card... nice to ramp them down for games where I have the frame rate locked anyway instead of driving 300 FPS for nothing)
 
People always complaining, AMD makes solid cards that have decent drivers (mining).
What are people only playing benchmarks these days. Had my fair share of nvidia driver issues that seemed bananas, for instance an online game on old drivers would start to lag, no not ping, the game would lag to hell and artifact etc... on a 780. Why who knows but it happens.
 
  • Like
Reactions: ChadD
like this
That doesn’t matter unless the PSU has multiple rails for the PCI-E connectors. Those cables can more than handle the wattage of two 8 pins going through them.

Last I checked... ATX spec still says 150 watts per rail. Granted the spec is old... and I have no doubt most PSUs these days go above.

Still 5700 xt is a 220 watt card... I would not use one 8 bin cable with a splinter.

Perhaps most modern power supplies will be fine... but if your running something a couple years old. Or your just barely meeting the 750 watt min most MFGs state as a requirement I would for sure use 2 rails unless I had zero choice.
 
I hope AMD figures it out. Once you lose a customer for something like this and the other side works... you’ll have a really really hard time winning them back. The average joe.

I didn’t have any blackscreens on my Radeon VII but that’s Vega not Navi...

Last I checked... ATX spec still says 150 watts per rail. Granted the spec is old... and I have no doubt most PSUs these days go above.

Still 5700 xt is a 220 watt card... I would not use one 8 bin cable with a splinter.

Perhaps most modern power supplies will be fine... but if your running something a couple years old. Or your just barely meeting the 750 watt min most MFGs state as a requirement I would for sure use 2 rails unless I had zero choice.

I personally buy single rail PSUs. The idea of multiple rails feeding one device never felt right to me. I also believe it caused all sorts of issues with people using 2080tis and multi rail, corsair PSUs.

I agree about the splitter... it should be fine but it’d bug me. One well made connector in reality can handle 400W-ish. If you have a loose connection though... the risks aren’t worth it.
 
I hope AMD figures it out. Once you lose a customer for something like this and the other side works... you’ll have a really really hard time winning them back. The average joe.

I didn’t have any blackscreens on my Radeon VII but that’s Vega not Navi...



I personally buy single rail PSUs. The idea of multiple rails feeding one device never felt right to me. I also believe it caused all sorts of issues with people using 2080tis and multi rail, corsair PSUs.

I agree about the splitter... it should be fine but it’d bug me. One well made connector in reality can handle 400W-ish. If you have a loose connection though... the risks aren’t worth it.

I use to get black screens with my R9 290... and just about every computer I built for friends for a year would get them too. I lived in Japan when they came out and we could buy them below MSRP because a store in Akihabara sold them at MSRP but the yen rate was in my favor quite a bit. I jumped ship when people were still buying them for mining and bought a used GTX 970 cheap that Ive been wanting to upgrade for at least a year or two. Was going to get a 5700 XT but all this BS just means Ill be waiting longer.
 
Yes, my anecdotal luck is more credible simply because I own these cards vs you repeating shit you read :D
Not to mention your long sad history of bashing AMD around this forum...;)
 
I'm a web developer and I have to fix bugs all of the time. I recently started playing Borderlands 3 because it came free with my 3800X. My 290X is old though almost all games are solid.

I was encountering issues with the game crashing whenever I did an Alt+Tab which was infuriating. I change the display type (some fullscreen mode) and the problem stopped.

I did have another issue with the game though it seems to have stopped.

If you want to provide truly useful information you must know things like what keys you're pressing, when does the issue occur, etc. Is it when the game starts? Is it just Radeon or is it a game-specific issue? If you find the right wording people will usually say which GPU they're using. You can't fault Radeon for game-specific issues.

That being said the developers need good information to follow up. If you do encounter a crash you need to figure out what is going on. What are your temps? Is your GPU memory full before launching and it's not finding enough memory leading to a crash? Are you using 18GB of memory when you only have 16GB installed because you never ever close your browser and absolutely every tab is open? Are you using the latest drivers?

My issues have all been game-related, not driver related. I'm not saying the drivers don't suck and yes my GPU is a bit old though most of the threads I come across for the issues I do encounter are absolute trash. "Hey, try A and let us know what happens!" "Fixed, won't say what I did and just say fixed!" useless garbage threads.
 
I remember way back when I had AMD's last fastest single GPU (7970's) and all the massive driver issues getting Crossfire and Eyefinity working. The amount of hours I spent was insane... AMD has always had a terrible GPU driver team.
Yep. I skipped that one for GTX 680s. Cheaper and a little faster.
 
Last I checked... ATX spec still says 150 watts per rail. Granted the spec is old... and I have no doubt most PSUs these days go above.

Still 5700 xt is a 220 watt card... I would not use one 8 bin cable with a splinter.

Perhaps most modern power supplies will be fine... but if your running something a couple years old. Or your just barely meeting the 750 watt min most MFGs state as a requirement I would for sure use 2 rails unless I had zero choice.
Your PSU knowledge is a decade out of date, at minimum. Also, those recommendations are massively overblown. My computer doesn’t even consume 750 watts and it’s a hell of a lot more power hungry than what most people are using.

Yes, my anecdotal luck is more credible simply because I own these cards vs you repeating shit you read :D
Not to mention your long sad history of bashing AMD around this forum...;)
Anecdotal evidence is worthless. AMD acknowledged that there are problems. Stop acting like a fanboy, riding a company’s dick does no one any good, especially the customer. Being “loyal” or a fanboy and blatantly ignoring obvious problems only serves to give companies a free pass to screw people over.
 
I personally haven’t owned an AMD card in a long while, but I use them at work in specific machines for specific workloads and they are fine as long as I avoid doing updates in any way shape or form once they are stable. I have found them to be far more troublesome in that respect, that said they need airflow, dedicated power, and a clean system I have found too many programs that use GPU acceleration that just mess them up too easily.
 
Last edited:
That is why I tend to think it might be more about cooling then real software issues at least for a good chunk of the reported issues. No doubt some are real software issues with specific games... every GPU vendor runs into that.

AMD went with a blower on their first cards cause they where clearly worried about heat issues.... I have a feeling most of the "software" issues with black screens, or crashes are related to heat. I know the 5700 I have if I where to unplug my 3 case exhaust fans... I have no doubt the inside of my case would turn into a nice warm fall day in hell. After playing Odyssey for a few hours... my top exhaust feels like a space heater. (and I know its cool to shit on AMDs software right now... but their game tailored fan curves work great. Nice option when only a couple games really push the card... nice to ramp them down for games where I have the frame rate locked anyway instead of driving 300 FPS for nothing)
Something I've noticed is that AMD started to heavily monitor GPU temps since the introduction of Vega. For example the Vega 56/64 have HBM and Hotspot temperature sensors which show drastically different values from the core temp. I know this carried onto RX 5700's which now the Hotspot is called junction temp, and it can get stupidly hot. My Vega56 before a thermal paste swap would have a hotspot temp of 104C running Furmark. RX 5700's can achieve 110C in junction temp, which AMD considers within spec. There is no way that's normal to have a temperature of over 100C. To give you an idea the R9 Fury cards didn't measure HBM or any sort of hotspot or junction temperature.

So temperature could be a serious problem, unless the black screens happen when you run a web browser as well, which was my problem. For a while I was certain this was an issue with FireFox and used Chrome for over a month without any issues, but the problem came back eventually and nearly everyday the driver would crash and I'd get a black screen. I'm not sure how hot Navi cards get when using a web browser but for me my Vega56 doesn't go above 33C when using Firefox. BTW the hotspot temp on my Vega56 hits 76C when playing a game, which is fanatically better than the 104C it got before. Thermal Grizzly Kryonaunt is really good thermal paste, which is what I'd recommend Navi users to try if they do reach insane junction temps. I don't think it'll solve the issue but I wouldn't want to use a card that reaches 110C.
 
  • Like
Reactions: ChadD
like this
Your PSU knowledge is a decade out of date, at minimum. Also, those recommendations are massively overblown. My computer doesn’t even consume 750 watts and it’s a hell of a lot more power hungry than what most people are using.

Anecdotal evidence is worthless. AMD acknowledged that there are problems. Stop acting like a fanboy, riding a company’s dick does no one any good, especially the customer. Being “loyal” or a fanboy and blatantly ignoring obvious problems only serves to give companies a free pass to screw people over.

The spec hasn't changed.... perhaps most MFGs surpass the standard. But I wouldn't count on it... every major PSU MFG I'm aware of recommend GPUs with 2 power connectors, get their own plug. Sure it may work not doing that but... why the hell would you do that, if you buy a 200+ watt GPU get a proper PSU and plug it in as recommended.

MSI 8-pin x 1+ 6-pin x 1 750w recommended.
Asus 2 x 8-pin 600w recommended
Power Colour Two 8-pin PCI Express Power connectors 700w recommended.
Gigabyte 8 Pin*1, 6Pin*1 600w recommended

I agree 750w probably isn't required to run a 5700 xt... I know from AMDs driver that my card pulls around 220 watts... and I'm sure even counting for a bit of inefficacy 350 or so watts for everything else is fine. Having said that... I have no doubt there are people trying to run these things on 450 watt power supplies. Cause AMD has hit a sweet spot on price where a lot of people that are probably lower end GPU buyers might splurge a bit to get a 5700 or a XT. Especially when all the review sites come out and say its a killer value. Knowing its going to pull 250 watts at least on most peoples bronze and less then rated PSUs... I would say 700w is really the min I would consider for a 5700 class navi.

As for being out of date.... not really. The spec hasn't been updated since 2013... the last revision was 7 years ago. I really don't think many MFGs are going much further above and beyond the current spec... and the voluntary industry stuff like the bronze/gold/platinum efficiency badges don't exactly mean all that much.
https://web.archive.org/web/2017071...ocuments/guides/power-supply-design-guide.pdf
 
Man, I remember hearing about AMD's (well, ATI's) bad drivers way back. Sad to see that's still a thing. I guess they hadn't really been relevant enough for awhile, so it didn't come up as much recently.
 
Is the hardware actually good or really just broken that they can't figure it out with just software?
This, if you are having this much trouble writing stable drivers for your HW it is an indiciation of the quirkiness of the HW.
Just because the HW has competitive computation power doesn't mean it's good, which we saw so many times with Radeon.
 
Something I've noticed is that AMD started to heavily monitor GPU temps since the introduction of Vega. For example the Vega 56/64 have HBM and Hotspot temperature sensors which show drastically different values from the core temp. I know this carried onto RX 5700's which now the Hotspot is called junction temp, and it can get stupidly hot. My Vega56 before a thermal paste swap would have a hotspot temp of 104C running Furmark. RX 5700's can achieve 110C in junction temp, which AMD considers within spec. There is no way that's normal to have a temperature of over 100C. To give you an idea the R9 Fury cards didn't measure HBM or any sort of hotspot or junction temperature.

So temperature could be a serious problem, unless the black screens happen when you run a web browser as well, which was my problem. For a while I was certain this was an issue with FireFox and used Chrome for over a month without any issues, but the problem came back eventually and nearly everyday the driver would crash and I'd get a black screen. I'm not sure how hot Navi cards get when using a web browser but for me my Vega56 doesn't go above 33C when using Firefox. BTW the hotspot temp on my Vega56 hits 76C when playing a game, which is fanatically better than the 104C it got before. Thermal Grizzly Kryonaunt is really good thermal paste, which is what I'd recommend Navi users to try if they do reach insane junction temps. I don't think it'll solve the issue but I wouldn't want to use a card that reaches 110C.

Think your right. AMD seems to be much more worried about spinning fans up when needed the last couple generations. And ya no way 100c+ is good for anything. Highest I have noticed mine get was 95/6 or so... but at that point the fans where at 2800 rpm which is more then a bit loud. :) Since then though I have tuned my rear out take fan and I swapped the 2 120mm top fans I had with a couple 140s I had around. Have noticed the junction sits more around mid-high 80s now which makes me feel a lot better. Made things a bit quieter as well... the difference between 2200 rpm and 2800 rpm on those GPU fans is pretty huge.

I'm not on my gaming machine right now... but if I'm browsing on it, or doing pretty much anything non game related the fans shut off... so I think it sits around 45 degrees or something. I'm sure I could get it down by disabling fan shutoff and spinning the fans at 1000 rpm or something, but its nice when I'm not gaming that the system is very quite. No hiding when I stop doing things I should be doing to fire up a game though. :) I played a round of dota the other day... and I don't think my 5700 fans spun over 1000 rpm the entire match.
 
Haha, by this logic you ride Nvidia's dick. I use both companies , so fanboy or loyal costumer is not my label ;)

How does any of that apply to me? I call out Nvidia's shitty decisions all the time. And I also recommend the Navi cards, even with their driver issues. However, I make sure anyone I recommend one of the cards to knows about the complaints so that they're informed going in. You are falling into the typical trap of thinking "because I don't have problems it must mean it's perfect and no one should ever have problems". Its a gross misunderstanding of the nature of computer hardware and the interaction of hardware and software in people's systems.
 
The spec hasn't changed.... perhaps most MFGs surpass the standard. But I wouldn't count on it... every major PSU MFG I'm aware of recommend GPUs with 2 power connectors, get their own plug. Sure it may work not doing that but... why the hell would you do that, if you buy a 200+ watt GPU get a proper PSU and plug it in as recommended.

MSI 8-pin x 1+ 6-pin x 1 750w recommended.
Asus 2 x 8-pin 600w recommended
Power Colour Two 8-pin PCI Express Power connectors 700w recommended.
Gigabyte 8 Pin*1, 6Pin*1 600w recommended

I agree 750w probably isn't required to run a 5700 xt... I know from AMDs driver that my card pulls around 220 watts... and I'm sure even counting for a bit of inefficacy 350 or so watts for everything else is fine. Having said that... I have no doubt there are people trying to run these things on 450 watt power supplies. Cause AMD has hit a sweet spot on price where a lot of people that are probably lower end GPU buyers might splurge a bit to get a 5700 or a XT. Especially when all the review sites come out and say its a killer value. Knowing its going to pull 250 watts at least on most peoples bronze and less then rated PSUs... I would say 700w is really the min I would consider for a 5700 class navi.

As for being out of date.... not really. The spec hasn't been updated since 2013... the last revision was 7 years ago. I really don't think many MFGs are going much further above and beyond the current spec... and the voluntary industry stuff like the bronze/gold/platinum efficiency badges don't exactly mean all that much.
https://web.archive.org/web/2017071...ocuments/guides/power-supply-design-guide.pdf

The badges mean the PSU's meet efficiency standards, nothing more and nothing less.

Most PSUs these days put all of their PCI-E on a single rail, a lot only have a single rail period. It's been that way for a long time.

None of your links tell people to use two separate cables. I have NEVER seen a video card that tells users to use two separate cables.
 
Is the hardware actually good or really just broken that they can't figure it out with just software?

well going on their BIOS fiasco with the 5600, their cards are just as back as the drivers.
 
I'm still using Radeon drivers from September 2018 on one system because Devinci Resolve doesn't recognize my card for OpenCL processing with anything newer.

The solution is not to blame AMD but for you to switch to different software altogether.
 
The badges mean the PSU's meet efficiency standards, nothing more and nothing less.

Most PSUs these days put all of their PCI-E on a single rail, a lot only have a single rail period. It's been that way for a long time.

None of your links tell people to use two separate cables. I have NEVER seen a video card that tells users to use two separate cables.

https://seasonic.com/pub/media/pdf/consumer/user-manual/multilingual-user-manual.pdf" High powered graphics cards are usually outfitted with two (or three) PCIe
power connector slots. Seasonic prefers the use of two (or three) separate PCIe
cables to connect these cards to your power supply. "

https://www.bequiet.com/en/powersupply/883Be! Quite... also specifies 2 cables. Nothing to quote cause its in diagram form on page 41 of their dark power pro manual... they specify 2 8pin cards use 2 cables.

https://www.coolermaster.com/faq/Entry - "How to connect High Power Graphic Cards to Power Supply "
" High power graphic cards will usually have either the 8 pin +8 pin or the 8 pin+ 6 pin connector type.
If the VGA cards will be under heavy loading and for extended period of continuous use,it is highly recommended that 2 separate PCI-E cables are used when connecting to the power supply, for better stability in the performance. "

Don't think I need to say much else.... 3 min of hitting 2 higher end and 1 lower end PSU companies pages. Yep all 3 say the same thing... we recommend you use 2 cables. There is no good reason at all I think of to use one cable.... and yes the PSU manufacturers all agree. No doubt the GPU companies don't say anything... its not really their place. Its the PSU MFGs that say use 2 cables.
 
All my "cable" problems begin and end with overheated crimped connectors.
Pins tarnish, connecter turns brittle to charcoal, insulation migrates or smokes.

Very unusual for cable problems to start mid an undisturbed span of wire.
Except when dealing with absurdly thin or intentionally fusible wire...

ATX connector specs 9A per pin max, yet identically sized GPU pins allow 13A?
https://www.molex.com/webdocs/datasheets/pdf/en-us/0026013116_PCB_HEADERS.pdf
If we only choose to believe lesser specs, 12V*9A=108W per pair of power and ground.

Hopefully we get 48V GPUs sooner rather than later. Not so many amps...
 
Last edited:
I have a 5700XT rig in my guest room. The screen blacks out all the time and I have to unplug/replug the DP cable all the time. I should have just paid extra and got a 2070. I was a fan of AMDs for years, been using their cards solely from the 4870 until the 290x with a quad fire setup. The drivers were starting to see issues around that time, but I switch to a 1080 Ti and noticed a night and day difference. After the 5700 XT issue I’ve been having, I can’t recommend going back to AMD for awhile.
 
The cards are fine. The BIOS fiasco was a huge mess, but that's really only a launch/launch window concern. It's not going to be a long term issue. Plus, that has nothing to do with hardware quality.

It does when you cant fix it with drivers. Your average consumer has no idea how to bios flash a card and you run a chance of bricking your card.
 
Reproducing below a detailed comment, with list of issues from below the TechSpot article


https://www.techspot.com/news/84005...hics-cards-over-driver-issues.html#comment_26

Dsirius
Hi Steve,
I am one of the users who bought a new Sapphire RX 5700XT Nitro+ in november 2019.
I can confirm that I encountered a lot of drivers problems.
My comp specs are:
Ryzen 3900X, 64GB GSkill Trident Z Neo F4-3600C16Q-64GTZNC, Asus Crosshair VIII, Samsung C32HG70 FreeSync 2 HDR Monitor and up-to-date a Dell S3220DGF monitor.
I tested a lot of AMD configurations, I also had Ryzen 2700X, Ryzen 3800X, Sapphire RX 580, Vega 64 and Radeon VII.
When I update Radeon drivers I restart in safe mode and use DDU uninstaler every time. After that I reinstall also drivers for both monitors.
I can summarize these kinds of issues to help you replicate them if you want for Sapphire RX 5700XT Nitro+.
1. With Windows 10 fresh install - and drivers prior to 10 Dec 2019 not so many problems. With new AMD Radeon Software Adrenalin 2020 Edition driver the issues just flourished.
2. With Windows 10 with all drivers and more programs installed, especially more than 3 or 4 games from different platform - like steam, origin, blizzard, epic the issues also begin to manifest often.
So I think that drm games can cause a lot of problems but is not entirely from them.
3. Freesync and HDR can multiply the issues. If you have an HDR monitor - most of the black screen issues. If you disable HDR from Windows 10 and monitor settings the issues are present but not so often.
4. If you have dual monitor setup in extended configuration - the issues occur so often (almost every day with adrenalin 2020 drivers) especially with Freesync 2 and HDR enabled.
In extended configuration if you switch (or glide the mouse) from the game monitor to second monitor where you have an YouTube or twitch window - color issues, game crushes and more problems may appear.
5. With new AMD Radeon Software Adrenalin 2020 Edition driver - all of them until the last from February 2020 - I had the most issues. So I reverted to Adrenalin-2019-Edition-19.12.1-Dec2 drivers. Even with Dec2 2019 drivers I still have some issue but less than with Adrenalin 2020 drivers.
6. With AMD Radeon Software Adrenalin 2020 Edition driver those new Overlay and Relive options just screwed all the experience. A workaround solution is to delete (disable) all shortcuts from Relive. The issues are less but still present.
7. Games with DirectX 12 enabled - more issues than playing with DirectX 11. I tested with Deus Ex Mankind Divided.
8. Most game crushes I had was with Overwatch and Deus Ex Mankind Divided - with Overwatch it's like 10 times more.
In Overwatch, when you leave a game, even it's not your fault you are penalized. So I stopped playing Overwatch, because the game crushed so often.
9. I noticed (and also one of my friend noticed) like a low flicker on Dell S3220DGF HDR monitor even this monitor is flicker-free by its specs. My friend noticed in PUBG. Prior to, he had a GTX 1080TI and Acer XB27HU monitor. Now he has Dell S3220DGF and Sapphire RX 580 Nitro+.
9. Playing games from GOG platform and Guild Wars 2 from Arena Net - no issue with them or too few to bother. All games from GOG are DRM free.
10. An interesting fact - Radeon VII was the most stable video card, I had the least driver issues from all cards - like 1 issue to 20 issues from RX 5700XT.
This article and also your YouTube video "Can We Still Recommend Radeon GPUs? AMD Driver Issues Discussed" came right on time because I am at the point where I am very annoyed by these issues and was asking my self what is going on.
Hope to be helpful and if you have time, I'll be glad to hear your response or your thoughts and maybe, together from this forum, to urge AMD to FIX the DRIVERS!
Best regards.
 
4. If you have dual monitor setup in extended configuration - the issues occur so often (almost every day with adrenalin 2020 drivers) especially with Freesync 2 and HDR enabled.

Another person confirmed problem with multi monitor setups

https://www.techspot.com/news/84005...hics-cards-over-driver-issues.html#comment_40
I’ve now seen issues occurring on fresh installs of windows on more than one occasion. In fact on my very own system I even tried a different, older AMD graphics card and still encountered almost identical issues. I thought it was my monitor at first, however an old Nvidia 970 I have knocking about disproved that.

My only sort of fix was to underclock the graphics card quite significantly and you can mitigate most of the flickers. However the performance hit is rather enormous and not acceptable. I’ve also seemed to notice a trend in that the only people affected that I have seen are those with multi monitor setups, particularly using hdmi.
 
Another reason that has been flagged is many Nvidia users are switching to AMD (5700 xt). So need to do a clean install

https://www.techspot.com/news/84005...hics-cards-over-driver-issues.html#comment_51
Initially I just uninstalled the nvidia drivers and put in the 5700 xt and tossed on the amd drivers. As expected, I immediately had issues. DDU in safe mode fixed everything and I have only ever had any minor glitches since.

I suspect this is a great deal of the problem, considering this is one of the first amd cards that was enticing for multiple gpu generations, and almost everyone would have upgraded from nvidia and not Amd
 
Or maybe the aib didn't validate the board properly :(

https://www.techspot.com/news/84005...hics-cards-over-driver-issues.html#comment_68

Today 11:37 PMGothmog3vz
I wonder if there's a breakdown among card variants experiencing the issues? Brands and specific models.

I upgraded in December from my rock-solid 960 Strix to a 5700XT Tuf (because prior Asus experiences were great, and the price was competitive). Card had the same black-screen issues described, stuttering frame rates, system locking when playing Destiny... Turns out, ASUS basically didn't give a **** when they did their card variants, after a bit of research, and these Tuf 5700XT's suffer from inadequate cooling + some funky factory VBIOS configuration.

Returned the Tuf, bought a Sapphire Nitro+ 5700XT, and now the computer is just as stable as it was with my old 960 (but with much better graphics, naturally)
 
Had a 5700 XT and my experience with it was rather poor. About 10 minutes during gaing it will produce a solid green screen and the entire system will hang, prompting me to do a reboot. It was annoying.

Apparently the 5700 XT works fine on some PC's, some PC's like mine will play up.

I like AMD and I'm currently using an AMD CPU which is great, but I can't say the same with their GPU's. It is a shame because I really wanted to enjoy my 5700 XT but with it crashing on me, it was unacceptable. Sold it and got a 2080 Ti and what do you know, no more crashing or any issues.
 
It seems the return rates for Navi cards are higher than equivalent RTX cards ...

https://www.techspot.com/news/84005...hics-cards-over-driver-issues.html#comment_94

Good catch about Mindfactory, since I don't want to go over every model I pick the one with the highest sale number
PowerColor 5700XT Red Devil - 3560 unit sold - RMA rate 4%
KFA2 2070 Super EX - 4920 units sold - RMA rate 1%
(Cheapest 2070 Super vs highest reviewed premium 5700 XT model nonetheless)

Well 4x more return is pretty close to what Steve reported (5x) .
 
Last I checked... ATX spec still says 150 watts per rail. Granted the spec is old... and I have no doubt most PSUs these days go above.

Still 5700 xt is a 220 watt card... I would not use one 8 bin cable with a splinter.

Perhaps most modern power supplies will be fine... but if your running something a couple years old. Or your just barely meeting the 750 watt min most MFGs state as a requirement I would for sure use 2 rails unless I had zero choice.

75+150 = 225

Also: I had Dell workstations rated for 300W GPU's with two six pins split from one eight pin from the factory. 625W Gold rated pull out server style PSU's.
 
I'm still using Radeon drivers from September 2018 on one system because Devinci Resolve doesn't recognize my card for OpenCL processing with anything newer.

Under what system? The one with Windows 8?
 
Joel Hruska of Extremetech encountered issues with both 5700 & 5600 series cards, while doing 5600 review

https://www.extremetech.com/extreme...ng-black-screen-driver-issues-on-radeon-cards
After the review ran, I worked with AMD to troubleshoot the issue. I had missed a UEFI update from Asus released between November and January, and AMD recommended setting the motherboard to specifically use PCIe 3.0 on its x16 slots rather than “Auto” for link speed detection. The combination of the new UEFI and the PCIe 3.0 setting fixed my problem, though I didn’t test which of the two changes were responsible.

According to AMD, the black screen issue I had encountered was distinct from the problem affecting games. That’s a really important point in this. The reason I’m sharing information on what was deemed a separate issue is because there may be more than one contributing factor to the issue.

...

The thread for the public release driver for the 5600 XT also notes: “Some Radeon RX 5700 series graphics users may intermittently experience a black screen while gaming or on desktop. A potential temporary workaround is disabling hardware acceleration in applications running in the background such as web browsers or Discord.”
 
Status
Not open for further replies.
Back
Top