I’m Staying on AMD!

[Valaska]

41 minutes ago, aminit said:

So talking truthfully about their experience with the cards, highlighting all the driver issue they encountered and barely recommending them (like you said because "Oh it would suck for there to be no competition with Nvidia..." which is a valid argument price wise) even though AMD is one of their sponsor is irresponsible?

Theys till, at the end of the day, have these issues... unfixed, without fully understanding them or looking into if others were having them. ALL THREE of them had these issues, but they still recommend these cards? When there is a MASIVE chance you have a card which can't run all games, randomly will driver timeout on some, and can easily sneak by without you noticing it past the return window.

 

At that point you are stuck with a $1500 GPU which you cannot use fully. There is no fix. AMD is not acknowledging the issue, despite it being known and reported to them.

 

Still recommending a faulty product just because "I don't like Nvidia as a company!" Is horrible. That's anti-consumer, that is giving people a false belief that the AMD product does not have issues that should ward you away from it. nvidia and intel DO NOT have these issues right now, Linus and team are walking people into buying faulty products.

 

While yes, they acknowledge they exist... recommending them after acknowledging that simply because of competition? I dunno, I couldn't do that personally and I had to speak out about it. There are hundreds of people that I know of, dozens I talk to personally, and potentially thousands to tens of thousands who don't understand they have this issue or might be blaming it on other hardware (ie, PSU, RAM, etc). We need solidarity here to get AMD to knuckle down and investigate these problems... or do a recall.

 

People who are saying they don't have this issue are now after seeing this and my other posts I am spreading around, finding out they do have it. There is one here, even... this is not rare, this is not isolated. This is widespread to a point it might be affecting the entire range. But it is so random and sporadic many won't know until they get into an application/game their GPU will randomly timeout on for no explainable reason.

 

AMD needs to answer for this, they need to address it, and giving them a recommendation lets them ignore it. AMD needs to fix this GPU, because the 7XXX series would be an amazing value performance proposition... but right now it is... heck, I've never seen in my entire life since the Sony VAIO disaster such an awful issue with hardware and a lack of response.

[Valaska]

3 hours ago, ToboRobot said:

But that is for your usecase and workload. 

Not everyone is running the same software and having the same problems.

Just because it is a problem for you and people running the same software, doesn't mean it's a problem for everyone else, and maybe they find value in the products.

But you made an overly broad comment that isn't really accurate.  And are sticking to it, because I guess your workload is the most important thing in the world. 

https://docs.google.com/forms/d/e/1FAIpQLScL7m-EGEsA3x4-0Td7dD-HXXUngZvL52y-oRoqwvRXfK7j1Q/viewanalytics

 

Here is 183 people, with different hardware configs... different software being affected, all the same drive timeout issue. This is not JUST my use case, and this is not isolated. People on this forum are starting to find out they have the issue, people in other places I am spreading are starting to test out more applications/games and finding that... yeah, despite the main games they play, they too have drive timeout issues in others.

 

This is overly broad, because IT IS A BROAD ISSUE.

 

Cyberpunk 2077 works 100% fine for me, others? It times out. AC6 times out for me, for others? It doesn't. Unreal Editor, same story. These timeout issues are deep in the drive suite and it is not just my set up or not just the applications I use.

 

And even if it was, why would you ever recommend something with such massive glaring issues? Unfixable ones, so far. How is that a recommendable product? Can you explain to me why that would be acceptable?

 

Can you explain to me, how even in this video, all three of them had different hardware configs than me... they all had this issue. And continued to have the issue even after they updated. Explain to me, how that gets a recommend?

[Ultraforce]

4 hours ago, Valaska said:

The video and discussion is based around the GPUs, specifically. I still recommend their CPUs just fine, even AM5. AM5 gets overly maligned, but is solid so far with MINOR issues. Their GPUs do not have minor issues... and even on the forums you can see some people with, yes, RDNA 2 and 6XXX are actually meeting these same issues. Again, this issue is sporadic and you might not even know you have it, UNTIL you do. That's the scary thing.

Saying do not buy though is too much of a blanket statement.

For certain stuff the opposite is also true

Do not buy Nvidia GPUs. (If you want a good experience playing video games on Linux with Wayland and without having to install specific drivers.)

At the end of the day sometimes the problems people have issues with for GPUs is too random for it to be responsible to advocate for someone to use one specific brand of GPU. Read reviews, complaints and so on. My 6700XT is working fine on both Windows and Linux and was a bit simpler to set up then the 1070 was.

[Valaska]

47 minutes ago, Ultraforce said:

Saying do not buy though is too much of a blanket statement.

For certain stuff the opposite is also true

Do not buy Nvidia GPUs. (If you want a good experience playing video games on Linux with Wayland and without having to install specific drivers.)

At the end of the day sometimes the problems people have issues with for GPUs is too random for it to be responsible to advocate for someone to use one specific brand of GPU. Read reviews, complaints and so on. My 6700XT is working fine on both Windows and Linux and was a bit simpler to set up then the 1070 was.

 

Yeah that statement for Linux would be true, there are some serious issus with the RTX 3-4XXX series and I would say it would be fully apt to say "do not get this as a linux user." For 97% of PC users though? Nvidia is the vastly safer buy, as even 6XXX are having the drive timeout issue, and the 7XXX's get them a lot.

 

 

 

For 97% of PC users... imagine this.

 

A game series you love hasn't had a new game released for 10 years, ten years! And all the sudden an amazing story trailer drops... then another, then gameplay! It looks incredible (even if they soulsified it a bit too much with mid-mission loadout changing... and auto-tracking).  Some of the former players you played tournaments against are saying "You know what, even though it has that, it is amazing."

 

You preorder it, collectors edition, $250! Your 7900 XTX crushes its recommended specs... lol an RX 590? Give me a break! It finally arrives at your house a little past launch because the code is physical, and you finall download it off steam and get in!

 

It looks beautiful, the tutorial etc runs butter smooth. The raytracing in just the hangar looks jaw-dropping, and you wish it was in the game itself.  Two hours in, you are feeling the world, you are attached to the AC etc an- DRIVER TIMEOUT THE ENTIRE SYSTEM LOCKS FOR 5 MINUTES GAME CLOSES NOTHING SAVED.

 

This is what Linus and team decided to recommend still. I have no idea if I can buy a game... not because my GPU isn't powerful enough, but because the game COULD randomly and unpredictably timeout. Some old games I love do it. Some of the newest of the new do it. FSR on, FSR off, 1 monitor or 2, new cables, upgrade the RAM, MPO, registry edit timeout delay.

 

Linus and Company had the same issues, and they signed off on it anways. And you know what... I bet they're running Nvidia at home too, probably even because of this issue I am willing to bet. They are misleading literally hundreds of thousands of viewers to a situation where there is no solution currently. It's not a hardware compatibility thing, it's not a game developer thing it has NOTHING to do with developers it is NOT their fault. It is completely on the brand, AMD, and it doesn't matter if Nvidia and Intel are the only competitors int eh market... AMD put out an unreliable product that could ruin your enjoyment of the ENTIRE hobby. It has mine. It has to hundreds of others reporting these issues. We are home buyers we aren't multi millionaires... we can't just swap to a 4090 at a whim. Some people have spent thousands getting new components trying to fix this, when in actuality it wasn't their MOBO, it wasn't their RAM, it was not their GPU... people swapping out entire systems except the GPU because other people/forums led them astray.

 

I won't be doing that. I want to warn everyone here, and anyone who might see it. DO NOT BUY AMD GPUS. For now... will AMD take this issue seriously enough to fix it?

 

Not with people like the Linus team recommending them and burying any PR snafu... people blindly listen to Linus and even if it is crashing, like it is confirmed across all 3 presenters here, they'll listen to that recommendation. And if your game is affected, it is a 100% guarantee it will crash, when? Who knows. 10 minutes, 20, 10 hours... it will crash on you. Sometimes 20+ times in one day, sometimes 1 or 2 times a day. This is... insanity. And I don't think Linus and company would accept this risk in any other product. If this was Nvidia or Intel they probably would have made a huge piece admonishing them, and rightfully so.

 

Linus himself said it, if he didn't get it for free it would have been a much bigger issue... so why, why on earth, does he still recommend it? 3/3 people had the issues, the issue STILL persists with the most recent drivers... 3/3 was their sample and test size. And there are, more, many more of us.

 

Do not buy AMD GPUs.

[aminit]

1 hour ago, Valaska said:

Theys till, at the end of the day, have these issues... unfixed, without fully understanding them or looking into if others were having them. ALL THREE of them had these issues, but they still recommend these cards? When there is a MASIVE chance you have a card which can't run all games, randomly will driver timeout on some, and can easily sneak by without you noticing it past the return window.

 

At that point you are stuck with a $1500 GPU which you cannot use fully. There is no fix. AMD is not acknowledging the issue, despite it being known and reported to them.

 

Still recommending a faulty product just because "I don't like Nvidia as a company!" Is horrible. That's anti-consumer, that is giving people a false belief that the AMD product does not have issues that should ward you away from it. nvidia and intel DO NOT have these issues right now, Linus and team are walking people into buying faulty products.

 

While yes, they acknowledge they exist... recommending them after acknowledging that simply because of competition? I dunno, I couldn't do that personally and I had to speak out about it. There are hundreds of people that I know of, dozens I talk to personally, and potentially thousands to tens of thousands who don't understand they have this issue or might be blaming it on other hardware (ie, PSU, RAM, etc). We need solidarity here to get AMD to knuckle down and investigate these problems... or do a recall.

 

People who are saying they don't have this issue are now after seeing this and my other posts I am spreading around, finding out they do have it. There is one here, even... this is not rare, this is not isolated. This is widespread to a point it might be affecting the entire range. But it is so random and sporadic many won't know until they get into an application/game their GPU will randomly timeout on for no explainable reason.

 

AMD needs to answer for this, they need to address it, and giving them a recommendation lets them ignore it. AMD needs to fix this GPU, because the 7XXX series would be an amazing value performance proposition... but right now it is... heck, I've never seen in my entire life since the Sony VAIO disaster such an awful issue with hardware and a lack of response.

Seems like you dont care about a productive discussion and just want to rant and vilify LTT, I cant help you with that

[Valaska]

11 minutes ago, aminit said:

Seems like you dont care about a productive discussion and just want to rant and vilify LTT, I cant help you with that

I will have a productive conversation, but they got it wrong on this. Massively wrong... and I have seen zero arguments that are convincing the other way. 3/3 people had this issue, they had the issue after.

 

Can you imagine any product other than an AMD GPU they'd still sign off on that being... good? A recommended buy? Linus' userbase have proven they will ignore other people's testing and even the community and just buy on his recommend alone, you can see it in the comments of that video alone.

 

That's why I am here saying that it was absolutely irresponsible for them to recommend it at the end despite a 100% issue rate on their team, without diving more into it too, and just blaming it on... a riser? Riser cables work perfectly fine, and it has nothing to do with this issue. They didn't... even look into whether or not others were having the issue en-mass to this day, which we are. There are still journals and articles coming out by smaller publications that AMD can ignore.

 

Linus being blindly trusted by so many giving this recommendation regardless of a 100% issue rate means AMD doesn't have to address this. Ever. because the issue is so random, and sporadic, that people could go months past return periods before they see issue.

 

So while I am glad to have the conversation, I can't sit by and abide by the recommendation, and I feel the need to warn people away from buying this... especially in my situation. I am an indie game developer, I need a strong GPU for work, especially if we are to transition to Unreal and especially when checking over renders etc... but I can't afford to just pick up a 4090, or a 4080 even. And if I personally knew these issues existed and it was such a huge risk that 3/3 on a testing team had them, I would have just saved the extra $600 for a 4090 or even just $200 for the 4080... because I need this to be reliable. And you know what... you should too, as a gamer.

 

 

Edited September 29, 2023 by Valaska
BEAR IN MIND! I give Linus and his team HUGE PROPS for tackling and addressing this issue in SOME capacity.... many others are ignoring it or brushing it under the rug. So kudos for at least bringing it up. But the recommend, that hurts...

[Ultraforce]

51 minutes ago, Valaska said:

I will have a productive conversation, but they got it wrong on this. Massively wrong... and I have seen zero arguments that are convincing the other way. 3/3 people had this issue, they had the issue after.

 

Can you imagine any product other than an AMD GPU they'd still sign off on that being... good? A recommended buy? Linus' userbase have proven they will ignore other people's testing and even the community and just buy on his recommend alone, you can see it in the comments of that video alone.

 

That's why I am here saying that it was absolutely irresponsible for them to recommend it at the end despite a 100% issue rate on their team, without diving more into it too, and just blaming it on... a riser? Riser cables work perfectly fine, and it has nothing to do with this issue. They didn't... even look into whether or not others were having the issue en-mass to this day, which we are. There are still journals and articles coming out by smaller publications that AMD can ignore.

 

Linus being blindly trusted by so many giving this recommendation regardless of a 100% issue rate means AMD doesn't have to address this. Ever. because the issue is so random, and sporadic, that people could go months past return periods before they see issue.

 

So while I am glad to have the conversation, I can't sit by and abide by the recommendation, and I feel the need to warn people away from buying this... especially in my situation. I am an indie game developer, I need a strong GPU for work, especially if we are to transition to Unreal and especially when checking over renders etc... but I can't afford to just pick up a 4090, or a 4080 even. And if I personally knew these issues existed and it was such a huge risk that 3/3 on a testing team had them, I would have just saved the extra $600 for a 4090 or even just $200 for the 4080... because I need this to be reliable. And you know what... you should too, as a gamer.

 

 

They have litterally recommended for specific workloads or for tinkerers Intel A770 and A750 at a time because for the price the hardware is fine it's just drivers aren't great especially for older games.

[atxcyclist]

AMD Radeon driver issues have been what’s kept me from daily-driving one in my home machine, and to an extent maybe they are still somewhat problematic. That said, I really want something on the level of a 4070 Ti, but aside from limited raytracing capabilities, something like a 7800 XT would be much cheaper and compelling, if the drivers are solid.

[Si3Rra_7]

Yeah, getting an AMD GPU has felt exactly as I predicted. GPU was cheaper, great price performance, but since it's not even close to nvidia market share, so devs just don't care to put in work to fix AMD related bugs in their games. And some issues are AMD's fault too.

 

For example: CS2 launched after more than a year of various levels of beta access, and even after launch AMD GPUs still suffer from insane 1-2 second freezes EVERY SINGLE TIME you load into a map due to shader compilation. 

 

Current workaround is starting the game, and loading into every single map before joining a competitive game. Freaking cringe. 

Second workaround is kneecapping all your other DX11 games by disabling DXNAVI by doing some sketchy registry hacks. Freaking yikes. 

 

I'm not regretting getting the 7800xt, yet. But I'm on my way there. 

 

Edit: for those wondering yes, i did just try it on the driver that just came out that had CS2 support.

[BlueChinchillaEatingDorito]

On 9/28/2023 at 10:27 AM, Cave_Johnson82 said:

Does anyone know why Linus is using speakers from like 2011 (seems like he's using corsair sp 2500). I mean they're not bad, but I thought Linus would use something ultra high-tech stuff, you know? (You can see them at 2:45 in the video)

If it ain't broke, don't fix it. 

 

Honestly, it's a good thing. Tech YouTubers preach about the right to repair and the sheer amount of e-waste every day. But if you're upgrading practically everything each year, that argument falls apart. You're just creating more stuff that'll eventually become e-waste. Audio equipment is one of the only things that if they're good, they stay good, plus they rarely go bad. I'm still using a 2.1 set of speakers from the mid-90s because they still work and sound good to this day. 

[themrsbusta]

On 9/29/2023 at 1:36 PM, Valaska said:

The riser is not the issue. This is an issue, inherent, to the 7XXX series. I have no riser, I have the most modern PSU on the market by release date ready for PCIE5 and originates a 16 HPWR to 2x8 pin even. MSI MPG A1000G, this thing is among the best of the best PSUs you can get and guess what... same issues. Hundreds of us, have this issue. We have a hardware survey and barely any of us are using Risers but we to this day all have this issue still.

Wait, have you joined Thursday just for say this? Curious, really curious...

[themrsbusta]

On 9/29/2023 at 5:55 PM, Valaska said:

https://docs.google.com/forms/d/e/1FAIpQLScL7m-EGEsA3x4-0Td7dD-HXXUngZvL52y-oRoqwvRXfK7j1Q/viewanalytics

 

Here is 183 people, with different hardware configs... different software being affected, all the same drive timeout issue. This is not JUST my use case, and this is not isolated. People on this forum are starting to find out they have the issue, people in other places I am spreading are starting to test out more applications/games and finding that... yeah, despite the main games they play, they too have drive timeout issues in others.

 

This is overly broad, because IT IS A BROAD ISSUE.

 

Cyberpunk 2077 works 100% fine for me, others? It times out. AC6 times out for me, for others? It doesn't. Unreal Editor, same story. These timeout issues are deep in the drive suite and it is not just my set up or not just the applications I use.

 

And even if it was, why would you ever recommend something with such massive glaring issues? Unfixable ones, so far. How is that a recommendable product? Can you explain to me why that would be acceptable?

 

Can you explain to me, how even in this video, all three of them had different hardware configs than me... they all had this issue. And continued to have the issue even after they updated. Explain to me, how that gets a recommend?

And who are those people?

[MarkPol88]

On 9/30/2023 at 12:34 AM, Valaska said:

Do not buy AMD GPUs.

Thread on AVS with nVidia problems regarding HDMI signal on some TV's is 118 pages long.

Problems started after certain nVidia driver update.

 

Do not buy nVidia GPUs, right?

[Evelyn Harthbrooke]

On 9/29/2023 at 4:44 PM, Valaska said:

I will have a productive conversation, but they got it wrong on this. Massively wrong... and I have seen zero arguments that are convincing the other way. 3/3 people had this issue, they had the issue after.

 

Can you imagine any product other than an AMD GPU they'd still sign off on that being... good? A recommended buy? Linus' userbase have proven they will ignore other people's testing and even the community and just buy on his recommend alone, you can see it in the comments of that video alone.

 

That's why I am here saying that it was absolutely irresponsible for them to recommend it at the end despite a 100% issue rate on their team, without diving more into it too, and just blaming it on... a riser? Riser cables work perfectly fine, and it has nothing to do with this issue. They didn't... even look into whether or not others were having the issue en-mass to this day, which we are. There are still journals and articles coming out by smaller publications that AMD can ignore.

 

Linus being blindly trusted by so many giving this recommendation regardless of a 100% issue rate means AMD doesn't have to address this. Ever. because the issue is so random, and sporadic, that people could go months past return periods before they see issue.

 

So while I am glad to have the conversation, I can't sit by and abide by the recommendation, and I feel the need to warn people away from buying this... especially in my situation. I am an indie game developer, I need a strong GPU for work, especially if we are to transition to Unreal and especially when checking over renders etc... but I can't afford to just pick up a 4090, or a 4080 even. And if I personally knew these issues existed and it was such a huge risk that 3/3 on a testing team had them, I would have just saved the extra $600 for a 4090 or even just $200 for the 4080... because I need this to be reliable. And you know what... you should too, as a gamer.

 

 

Man I feel as if you're only in this thread to hate on AMD. Issues happen. They happen with NVIDIA too; they aren't somehow immune to having driver issues. I've seen loads of people have a ton of driver issues on NVIDIA cards as well. And as someone who has the 7900 XTX as well, I've noticed zero issues on my end personally. Instead of criticizing AMD, perhaps RMA your unit. RMAs exist for a reason, and if your card isn't reliable, it might be a hardware fault. Therefore I suggest an RMA. Additionally, your replies in this thread are unnecessarily hostile.

[Dingz]

I didn't get any problem with my AMD drivers since 10y+. And now you posted this video, my computer stutters like hell on hardware accelerated content (Youtube on Firefox, VLC...) with the latest drivers 23.9.3 and my 6800xt... had to revert back to 23.9.1 to get it normal... thanks :p

[Kra]

The issue Luke encountered is probably because he is using a pcie

3.0 riser cable considering his last card is a pcie3.0 Titan RTX. GPUs on idle mode will slow down to pcie1.1 which you can find it in GPUz and rise to 4.0 when it has loads on. That is why the system become unstable even if the work load is lightweight.

[Jaw709]

On 9/28/2023 at 12:59 PM, Other James said:

Back in March Luke, Jake, and myself all switched over to AMD to test our resolve after renouncing the 40 series. So are we sticking with AMD, or did we go crawling back to Nvidia?

 

Buy a PowerColor Red Devil AMD Radeon RX 7900 XTX Graphics Card: https://geni.us/mDl4bq8

Buy a Seasonic Prime TX-1000 80+ Titanium PSU: https://geni.us/w5pR4

Buy an XFX Speedster MERC310 AMD Radeon RX 7900XT Graphics Card: https://geni.us/muzK78

Buy a Seasonic Prime TX-850 80+ Titanium PSU: https://geni.us/B5CJz1

Buy a PowerColor Red Devil AMD Radeon RX 6950 XT Graphics Card: https://geni.us/gC3cd0

Buy a Seasonic Focus GX-850 80+ Gold PSU: https://geni.us/JUGwV9

Buy an XFX Speedster SWFT309 AMD Radeon RX 6700 XT Graphics Card: https://geni.us/yiueR

Buy a Seasonic Focus GX-750 80+ Gold PSU: https://geni.us/O6Taa

Purchases made through some store links may provide some compensation to Linus Media Group.

 

 

Was having similar random full PC crashing issues and awake-from-sleep lag as @Luke . After months of anguish and troubleshooting, it ended up being power settings. Even on "High / performance," the PCIE "Link State Power Management" was on moderate. I think it throttles power to the slot; after switching to "off" all the stuttering and crashing resolved.