Why no one has the right to be angry at AMD with regards to AM4

[TheLostSwede]

Your topic is a good reminder and wake up call. But at the same time, lots of problems AMD has around releases are not ever about money or being an underdog at all.

It is always, eternally, every single god damn time about COMMUNICATION.

They suck massively at it, plans change last minute or things get axed with no advance notice. And then, to make matters worse, they pile on with some driver and BIOS confusion to solidify the idea things are not perfect at all. That is fine, one time. A second time can be forgiven. But with AMD, its company culture.

I'm not fixed to any camp, blue green red, and if yellow comes around I'll take a long look too. But I don't reward companies that perform below par or drop the ball with my hard earned cash. I did already build several Zen rigs. I would now, since the latest release also recommend these CPUs for general gaming over Intel. Things do change, but they do it on MY pace based on MY view of a company. Not because the market somehow demands it for 'balance' or 'to do my part'. In that sense, I'm not touching AMD GPUs with a ten foot pole yet, completely unreliable. If they can produce a few good generations with no crappy events and drivers, I'll reconsider. Hopefully they get the memo with RDNA2, because the first iteration was again pretty shaky.



Its always about business and anyone thinking otherwise is deluded. Intel is no different in that regard.

One thing I've hopefully learnt from this is, don't be an early adopter of AMD hardware. They do seems to fix most of the stuff, but with my current rig, it took the better part of three months. Next time, I won't be so quick to jump in, that's for sure, but at the same time, it hasn't put me off AMD hardware. Admittedly I haven't used one of their graphics cards in something like the past 10 years, for various reasons and it seems like they still need to spend a lot more time and resources on their drivers, from what I can tell.

 

[Bill_Bright]

This would actually be an interesting question for the poll on TPU main page - how often do you upgrade your CPU and/or motherboard?

I don't think it would be useful. Businesses typically establish 5 year plans, for example, and might replace hardware on a timetable schedule. But I personally don't know any users (including enthusiasts) who do.

For myself and others I know, we upgrade when our current setup no longer meets our needs (perhaps forced by some failure), or when some new technology comes out we can't live without. Not because X number of months have passed.

AMD is like, well, we need more money.
Intel is like, mwaha we have cookies.

Nah! Not true at all. All companies (even non-profits) have the same goal - make money. And all companies have the goal to grow. If a company cannot grow and keep up with the times and technologies, they will fail. And all companies have the goal of not failing.

 

[xrobwx71]

I don't think it would be useful. Businesses typically establish 5 year plans, for example, and might replace hardware on a timetable schedule. But I personally don't know any users (including enthusiasts) who do.

For myself and others I know, we upgrade when our current setup no longer meets our needs (perhaps forced by some failure), or when some new technology comes out we can't live without. Not because X number of months have passed.

Nah! Not true at all. All companies (even non-profits) have the same goal - make money. And all companies have the goal to grow. If a company cannot grow and keep up with the times and technologies, they will fail. And all companies have the goal of not failing.

Spot on Bill.

 

[delshay]

Question:

Are any of the AM4 motherboards BIOS in a socket, or are all soldered to motherboard. Just need one or two users to post answer.

 

[TheLostSwede]

Question:

Are any of the AM4 motherboards BIOS in a socket, or are all soldered to motherboard. Just need one or two users to post answer.

Some have one of the chips in a socket. This is a bit off topic though, no?

 

[delshay]

Some have one of the chips in a socket. This is a bit off topic though, no?

[

Not really off topic because if it's in a socket & the rom is too small all the end user has to do is purchase a pre-programmed firmware from said company.

You may want to read this, so hold on users.

https://www.reddit.com/r/Amd/comments/gkgmwf

 

[TheLostSwede]

[

Not really off topic because if it's in a socket & the rom is too small all the end user has to do is purchase a pre-programmed firmware from said company.

You may want to read this, so hold on users.

https://www.reddit.com/r/Amd/comments/gkgmwf

No, sorry, but it's not that simple. AMD won't add support in the AGESA for the combination of 4000-series CPUs on any chipset prior to X570.
The board makers have been told it's a no go.
Sure, someone might be able to hack something together, but it's not likely to work reliably.

 

[delshay]

No, sorry, but it's not that simple. AMD won't add support in the AGESA for the combination of 4000-series CPUs on any chipset prior to X570.
The board makers have been told it's a no go.
Sure, someone might be able to hack something together, but it's not likely to work reliably.

Maybe so, but they could do something along the lines of trade in your old motherboard & buy the newer one at a discounted price. Let's see if Lisa SU can change anything to help end users.

 

[TheLostSwede]

Maybe so, but they could do something along the lines of trade in your old motherboard & buy the newer one at a discounted price. Let's see if Lisa SU can change anything to help end users.

Sure, but that's something entirely different. Doubt it'll happen though.

 

[Bronan]

Guess i am not a consumator i actually use since 1 month a 8700k the old still fantastic 6700k on a z170 board still works pretty darn well.
I have stopped wanting the newest and fastest because this only means a minor improvement for a incredible price tag which i have no actual need for.
The only reason i bought the 8700k for is that it can speedup certain tasks related to backup and storage.
I actually would love to see a much faster storage system which really is faster and by that not in idiot synthetic benchmarks but real world faster.
Ofcourse i am a bit spoiled because i have been working on systems with a massive raid storage devices which was build around multiple raid controllers with insane speeds but for home usage that is far above my budget. The price of one of these controllers alone cost more than i got for my whole pc budget

 

[newtekie1]

I know I'm late to the party, and frankly I'm not going to read 4 pages of posts, I don't have the time. So, if what I say has been discussed, I'm sorry.

I think the uproar is not that they can't support all the processors at once, it's that they outright refuse to support the 4000 series on the 400 chipset boards. There seems to be no technical reason they can't, they are just choosing not to in an effort to force people to buy new motherboards. Yes, there are limitation with BIOS size, but we've already hit that and addressed work-arounds. Using x470 as an example, it technically supports all the processors from the 1000 series up to the 3000 series, but most motherboards can't do it all at once. In fact, using my x470 Taichi Ultimate as an example, it has different BIOSes for different processors and warns you not to use certain BIOSes with certain CPUs. And there is always a cross-over with the BIOSes. So if I say had a 1700 and wanted to upgrade to a 3700, I could but it wouldn't be as easy as just swapping out CPUs. I'd have to upgrade the BIOS to one that supports the 1000 and 2000 CPUs. Then swap in a 2000 cpu, then flash the BIOS to one that supports 2000 and 3000 CPUs, then put in the 3700. It's annoying, but works, and is better than outright not supporting newer processors.

Now, do I really have a right to be mad at AMD for not supporting the 4000 CPUs on my x470 chipset? No, not really. The 400 chipsets came out along side the 2000 CPUs, and then I got support for 3000 CPUs. So I got 2 CPU generations out of one chipset, which Intel has made pretty par for the course. But AMD kind of hyped themselves up here, so they are deserving of some anger. Part of their marketing has been better platform longevity. Now the 400 series platform is only seeing 2 CPU generations, so it turns out it's not any longer lived than an Intel platform at this point. If there was a technical reasoning behind it, like the socket is actually changing, then OK. But this comes down solely to a software lockout, which is pretty scummy on AMD's side given their previous marketing. And it is a very Intel thing to do, which has a lot of people disappointed.* IMO, if they are going to market longer platform life compared to Intel, each platform should see support for at least 3 CPU generations after it is released. So the 400 chipsets should work with 2000, 3000, and 4000 CPUs.

*Not me though, because if you actually look at history, AMD is just as bad as Intel when they can be.

 

[TheLostSwede]

I know I'm late to the party, and frankly I'm not going to read 4 pages of posts, I don't have the time. So, if what I say has been discussed, I'm sorry.

I think the uproar is not that they can't support all the processors at once, it's that they outright refuse to support the 4000 series on the 400 chipset boards. There seems to be no technical reason they can't, they are just choosing not to in an effort to force people to buy new motherboards. Yes, there are limitation with BIOS size, but we've already hit that and addressed work-arounds. Using x470 as an example, it technically supports all the processors from the 1000 series up to the 3000 series, but most motherboards can't do it all at once. In fact, using my x470 Taichi Ultimate as an example, it has different BIOSes for different processors and warns you not to use certain BIOSes with certain CPUs.

Now, do I really have a right to be mad at AMD for not supporting the 4000 CPUs on my x470 chipset? No, not really. The 400 chipsets came out along side the 2000 CPUs, and then I got support for 3000 CPUs. So I got 2 CPU generations out of one chipset, which Intel has made pretty par for the course. But AMD kind of hyped themselves up here, so they are deserving of some anger. Part of their marketing has been better platform longevity. Now the 400 series platform is only seeing 2 CPU generations, so it turns out it's not any longer lived than an Intel platform at this point. If there was a technical reasoning behind it, like the socket is actually changing, then OK. But this comes down solely to a software lockout, which is pretty scummy on AMD's side given their previous marketing. And it is a very Intel thing to do, which has a lot of people disappointed.* IMO, if they are going to market longer platform life compared to Intel, each platform should see support for at least 3 CPU generations after it is released. So the 400 chipsets should work with 2000, 3000, and 4000 CPUs.

*Not me though, because if you actually look at history, AMD is just as bad as Intel when they can be.

Fair comment, although I'm not sure they deserve anger, but a bit of a scolding, sure.
From what they've said publicly and from what I've been told off the record, yes, this is an AGESA/software lockout.
However, there might be more things that we don't know at this point.
It also seems like AMD wants people to use their PCIe 4.0 CPUs/APUs with PCIe 4.0 compliant motherboards. Sure, it's a bit of a lame excuse but hey...
They did over hype the longevity, but we also don't know what's cooking at AMD. Maybe next year 16 cores will be mainstream and some older boards won't be able to cope, but this is just a wild guess on my side.
Time will tell if there is something more to this than what we know right now.

 

[newtekie1]

Fair comment, although I'm not sure they deserve anger, but a bit of a scolding, sure.
From what they've said publicly and from what I've been told off the record, yes, this is an AGESA/software lockout.
However, there might be more things that we don't know at this point.
It also seems like AMD wants people to use their PCIe 4.0 CPUs/APUs with PCIe 4.0 compliant motherboards. Sure, it's a bit of a lame excuse but hey...

I think a small amount of anger is due, I'm not saying rake them over the coals.

I think they just want to sell more chipsets and boards to make more profit while also not wanting to deal with the tech support of people that can't figure out what BIOS they should be using or doing that whole "loan a CPU" think again.

Honestly, if I was AMD, I'd put out the full update that supports all CPUs and let the motherboard manufacturers flounder for using low capacity BIOS ROMs. Let ASUS and AsRock and MSI and Gigabyte etc, explain why their boards can't handle the new CPUs and why their decision to save 25¢ on a part for a $200 motherboard means their customers can't upgrade to the latest generation of CPUs.

They did over hype the longevity, but we also don't know what's cooking at AMD. Maybe next year 16 cores will be mainstream and some older boards won't be able to cope, but this is just a wild guess on my side.

Even in that case, it doesn't really matter, that is on the motherboard manufacturers to test what their boards can handle and post CPU support lists. I've certainly got plenty of boards that won't support CPUs that are out right now that technically work, but the board just can't handle the power draw. It isn't AMD's decision in that case to limit an entire generation of chipsets just because a few weaker boards might not handle the power requirements.

 

[Vario]

First,
If you really want the new CPU, then sell the old board and old CPU. Buy a new board and CPU. Not as big a deal as its made out to be. You can often sell the old board on eBay for the basis price, there is that much demand for working mint old motherboards. If you keep the motherboard in mint condition and the box, manual, disks, antistatic bag, IO panel, the motherboards do not lose value. If the new motherboards do not support the old CPUs, then it will ensure the old motherboards retain value.
Second,
CPUs gain like 5 to 10% performance at the most between generations. The idea of upgrading a CPU each and every generation (all while staying on the same motherboard) is an insane waste of money. Go out a few years before upgrading CPUs, and then at that future time, it makes sense to get a new motherboard, newest generation of ram, newest generation of SSD storage, and so forth.
Third,
There is a bit of irony that one of the more annoying to read AMD fanboy talking points was the supposed upgrade-ability of AM4 motherboards for every Ryzen generation going forward, and yet we find it had roughly the same longevity as Intel.

 

[Sybaris_Caesar]

Nobody has the right to tell consumers what and what not to be angry at. Frankly speaking people defending AMD even now are buncha corporate simps.

 

[Bill_Bright]

Frankly speaking people defending AMD even now are buncha corporate simps.

Gee whiz. That's just stupid.

 

[TheLostSwede]

Nobody has the right to tell consumers what and what not to be angry at. Frankly speaking people defending AMD even now are buncha corporate simps.

So what do you gain from going around being angry at something you can do absolutely nothing about?
I mean, do you think it's acceptable to send threatening emails to AMD staff because of this? As some people clearly do.
Yes, it's fine to be pissed off about it in a forum thread or two, but how long is ok to go around being furious over something like this?
I fully understand why some people are miffed about this, but there's a limit to how far you're allowed to take it.

 

[Vayra86]

I know I'm late to the party, and frankly I'm not going to read 4 pages of posts, I don't have the time. So, if what I say has been discussed, I'm sorry.

I think the uproar is not that they can't support all the processors at once, it's that they outright refuse to support the 4000 series on the 400 chipset boards. There seems to be no technical reason they can't, they are just choosing not to in an effort to force people to buy new motherboards. Yes, there are limitation with BIOS size, but we've already hit that and addressed work-arounds. Using x470 as an example, it technically supports all the processors from the 1000 series up to the 3000 series, but most motherboards can't do it all at once. In fact, using my x470 Taichi Ultimate as an example, it has different BIOSes for different processors and warns you not to use certain BIOSes with certain CPUs. And there is always a cross-over with the BIOSes. So if I say had a 1700 and wanted to upgrade to a 3700, I could but it wouldn't be as easy as just swapping out CPUs. I'd have to upgrade the BIOS to one that supports the 1000 and 2000 CPUs. Then swap in a 2000 cpu, then flash the BIOS to one that supports 2000 and 3000 CPUs, then put in the 3700. It's annoying, but works, and is better than outright not supporting newer processors.

Now, do I really have a right to be mad at AMD for not supporting the 4000 CPUs on my x470 chipset? No, not really. The 400 chipsets came out along side the 2000 CPUs, and then I got support for 3000 CPUs. So I got 2 CPU generations out of one chipset, which Intel has made pretty par for the course. But AMD kind of hyped themselves up here, so they are deserving of some anger. Part of their marketing has been better platform longevity. Now the 400 series platform is only seeing 2 CPU generations, so it turns out it's not any longer lived than an Intel platform at this point. If there was a technical reasoning behind it, like the socket is actually changing, then OK. But this comes down solely to a software lockout, which is pretty scummy on AMD's side given their previous marketing. And it is a very Intel thing to do, which has a lot of people disappointed.* IMO, if they are going to market longer platform life compared to Intel, each platform should see support for at least 3 CPU generations after it is released. So the 400 chipsets should work with 2000, 3000, and 4000 CPUs.

*Not me though, because if you actually look at history, AMD is just as bad as Intel when they can be.

I reckon the core of the issue is not so much the fact it is or is not supported, but that the suggestion existed it probably would be, and AMD did nothing, until the very last moment (launch!) to change that idea.

They knew damn well what was going to happen long before we did. And that is the key here. Communication. Managing expectations and your PR properly, and not ripping your potential customers a new hole with empty promises. However unrealistic they might have been, or vague. It was up in the air and AMD profited off that, but now gets the backlash.

Its only fair really. I also can completely imagine how the pessimist would view this: the immediate kneejerk response is 'Omg, they're just like Intel after all, now that they have sales'.... Its really unfortunate but pessimists tend to be right. AMD can easily prove them wrong by communicating more clearly on their socket approach going forward. But... that is also shooting themselves in the foot, because it reveals strategy. Bottom line... maybe even they are reconsidering their approach right now. After all, if it does not benefit the company, why do it.

One thing I've hopefully learnt from this is, don't be an early adopter of AMD hardware. They do seems to fix most of the stuff, but with my current rig, it took the better part of three months. Next time, I won't be so quick to jump in, that's for sure, but at the same time, it hasn't put me off AMD hardware. Admittedly I haven't used one of their graphics cards in something like the past 10 years, for various reasons and it seems like they still need to spend a lot more time and resources on their drivers, from what I can tell.

Oh yes... early adopting is a recipe for frequent headaches. I've learned a while ago that trailing the new stuff by one or a half gen is the best way to go. Stuff is readily available, patched up and good to go, and the price is not quite as inflated. And on top of that you get an idea for widespread adoption which in tech is another huge thing. Its good to be part of the herd, more users is more bug fixing and maintenance, and therefore quality.

Nobody has the right to tell consumers what and what not to be angry at. Frankly speaking people defending AMD even now are buncha corporate simps.

The reasons do or do not apply to you. We each have our own truths, its good to hear others from time to time, to get inspired

 

[newtekie1]

I reckon the core of the issue is not so much the fact it is or is not supported, but that the suggestion existed it probably would be, and AMD did nothing, until the very last moment (launch!) to change that idea.

They knew damn well what was going to happen long before we did. And that is the key here. Communication. Managing expectations and your PR properly, and not ripping your potential customers a new hole with empty promises. However unrealistic they might have been, or vague. It was up in the air and AMD profited off that, but now gets the backlash.

Its only fair really. I also can completely imagine how the pessimist would view this: the immediate kneejerk response is 'Omg, they're just like Intel after all, now that they have sales'.... Its really unfortunate but pessimists tend to be right. AMD can easily prove them wrong by communicating more clearly on their socket approach going forward. But... that is also shooting themselves in the foot, because it reveals strategy. Bottom line... maybe even they are reconsidering their approach right now. After all, if it does not benefit the company, why do it.

I believe a big part of it too is that people view AMD as "the good guys" that don't do evil things like Intel. I know better, I remember history and know that it repeats itself. Now that they are ahead, they are proving once again that they'll do exactly what their fans have bashed Intel for for years. At least with Intel there was socket improvements over the years that explained why new boards were necessary. But AMD is doing an outright software lockout, which is even worse than what Intel does.

 

[TheLostSwede]

Oh yes... early adopting is a recipe for frequent headaches. I've learned a while ago that trailing the new stuff by one or a half gen is the best way to go. Stuff is readily available, patched up and good to go, and the price is not quite as inflated. And on top of that you get an idea for widespread adoption which in tech is another huge thing. Its good to be part of the herd, more users is more bug fixing and maintenance, and therefore quality.

I wasn't going to be that cautious, only three to six months after launch

I believe a big part of it too is that people view AMD as "the good guys" that don't do evil things like Intel. I know better, I remember history and know that it repeats itself. Now that they are ahead, they are proving once again that they'll do exactly what their fans have bashed Intel for for years. At least with Intel there was socket improvements over the years that explained why new boards were necessary. But AMD is doing an outright software lockout, which is even worse than what Intel does.

Please list out what those socket improvements were from LGA-1155 to now.
Intel didn't need a new socket to go from PCIe 2.0 to PCIe 3.0.
Intel didn't need a new socket to add more SATA 6Gbps ports.
Intel didn't need a new socket to add more cores.
Intel didn't need a new socket to add USB 3.x.
Intel didn't need a new socket to add NVMe support.
Intel didn't need a new socket to add faster memory support.

New chipsets and other changes to the motherboard, yes.
But not a new socket, as there's no significant pin changes, so getting more interfaces for the CPU is not part of the problem here.
Yes, the new sockets were "needed" due to Intel changing the power domain inside the CPUs, but beyond this, all of the new features are related to new chipsets, as most of what I listed above was introduced on second generation CPUs/chipsets on the same socket.

 

[Vayra86]

I wasn't going to be that cautious, only three to six months after launch


Please list out what those socket improvements were from LGA-1155 to now.
Intel didn't need a new socket to go from PCIe 2.0 to PCIe 3.0.
Intel didn't need a new socket to add more SATA 6Gbps ports.
Intel didn't need a new socket to add more cores.
Intel didn't need a new socket to add USB 3.x.
Intel didn't need a new socket to add NVMe support.
Intel didn't need a new socket to add faster memory support.

New chipsets and other changes to the motherboard, yes.
But not a new socket, as there's no significant pin changes, so getting more interfaces for the CPU is not part of the problem here.
Yes, the new sockets were "needed" due to Intel changing the power domain inside the CPUs, but beyond this, all of the new features are related to new chipsets, as most of what I listed above was introduced on second generation CPUs/chipsets on the same socket.

Small caveat. If you want to push those new features, you will require new chipsets in the wild and the socket change is a perfect way to force that to happen.

It enables business cases for, for example, new storage media.

 

[TheLostSwede]

Small caveat. If you want to push those new features, you will require new chipsets in the wild and the socket change is a perfect way to force that to happen.

It enables business cases for, for example, new storage media.

Now that's an entirely different kettle of fish though. But this is exactly what AMD is now getting shit for. They want to push the platform forward, much like what Intel has done, but some of AMD's customers, seemingly disagree and are calling AMD out over it...

Yes, AMD failed from a marketing perspective, but we shouldn't pretend as if a new socket is the only reason why Intel has progressed its platforms forward, as it wasn't a true requirement, but rather the way Intel implemented it.

 

[Vayra86]

Now that's an entirely different kettle of fish though. But this is exactly what AMD is now getting shit for. They want to push the platform forward, much like what Intel has done, but some of AMD's customers, seemingly disagree and are calling AMD out over it...

Yes, AMD failed from a marketing perspective, but we shouldn't pretend as if a new socket is the only reason why Intel has progressed its platforms forward, as it wasn't a true requirement, but rather the way Intel implemented it.

Yep, well I've always been an advocate of the idea that you buy a board with a CPU because the two do tend to last equally long, even if its not your main rig anymore, why separate them to find a new buddy... Also the idea that the board will age less quickly than a CPU is... a bit strange. A board is much more prone to failure than a CPU in every single way, it will also go obsolete faster as new standards arrive.

This affects the way people purchase stuff though. If your base notion is that a board will be having multiple CPUs, you start off with the goal not of finding the optimal CPU you can ever find, but one 'that will do for now' and gets resold down the line. Its not my cup of tea, as it seems rather wasteful.

 

[newtekie1]

Please list out what those socket improvements were from LGA-1155 to now.
Intel didn't need a new socket to go from PCIe 2.0 to PCIe 3.0.
Intel didn't need a new socket to add more SATA 6Gbps ports.
Intel didn't need a new socket to add more cores.
Intel didn't need a new socket to add USB 3.x.
Intel didn't need a new socket to add NVMe support.
Intel didn't need a new socket to add faster memory support.

New chipsets and other changes to the motherboard, yes.
But not a new socket.

In short:

Going from 1155 to 1150 the socket was redesigned to simplify trace routing.
Going from 1150 to 1151 unused pins were repurposed to add Displayport output from the iGPU.
And going from 1151 to 1151(300 Chipsets) more power delivery pins were added to support higher core count CPUs.

 

[TheLostSwede]

In short:

Going from 1155 to 1150 the socket was redesigned to simplify trace routing.
Going from 1150 to 1151 unused pins were repurposed to add Displayport output from the iGPU.
And going from 1151 to 1151(300 Chipsets) more power delivery pins were added to support higher core count CPUs.

Those are still fairly minor things. Sure, simplifying trace routing makers for easier board designs, so it's a good thing, but did it require a socket change? Most likely not.
Unused pins being repurposed happens all the time, so no big deal.
The last one is questionable, as some people have modded older boards to work with newer CPUs, but sure. I'm not convinced it required Intel locking out older parts, as AMD didn't have to do it for AM4 where we've gone from eight to 16 cores.