The Best CPUs: Gaming, Productivity, and the Best Bang for Your Buck

[Steve]

The CPU market looks very different from six months ago. Intel's back in the conversation, AMD's X3D chips still rule gaming, and there's genuine value at every price point.

Read the full article here.

The Best CPUs: Gaming, Productivity, and the Best Bang for Your Buck

 

[poshflamingos]

Sounds solid, it's my currect rule of thumb too. If you only care about gaming, only 2 options are relevant, Ryzen 7500F (or 7600 if not available) and Ryzen 7800X3D (or 9800X3D if not available). If you care about productivity, only 1 option matters, the 270K Plus.

The problem is the RAM situation means all of this is moot for most people.

 

[cldmstrsn]

All I know is that for 4K gaming my 7700X is holding up remarkably well. I'll hold on to it and see what AMD's next set of chips will bring.

 

[Rapob]

Sounds solid, it's my currect rule of thumb too. If you only care about gaming, only 2 options are relevant, Ryzen 7500F (or 7600 if not available) and Ryzen 7800X3D (or 9800X3D if not available). If you care about productivity, only 1 option matters, the 270K Plus.

The problem is the RAM situation means all of this is moot for most people.

That entirely depends on which productivity tasks you want to do. Some don't scale well with threads, others like a big fast L2 cache. And if you truly need good productivity performance, you don't settle with mainstream and buys into HEDT space: Threadripper obliterates Intel here. You pay more, sure but you can't get anything that even comes close with Intels mainstream offerings. Intel shutdown their HEDT segment in 2019. AMD forced them.

270K Plus is decent value, if you want to accept a dead platform with no upgrade path. With AM5 you get an actual upgrade path, with a +50% core increase next time with Zen 6 and rumours about Zen 7 support as well. You can be 100% sure that Zen 6 with up to 24C/48T will make 270K Plus look dated in terms productivity, while bringing better gaming performance as well.

With 270K Plus, you buy into a dead platform that will get nothing in the future. Even 290K Plus was scrapped. Intel abandons this socket in half a year as their entire mainstream focus shifts to socket 1954, which hopefully gets 4 full generations instead of socket jumping every other generation to sell people a new motherboard/chipset.

 

[poshflamingos]

That entirely depends on which productivity tasks you want to do. Some don't scale well with threads, others like a big fast L2 cache. And if you truly need good productivity performance, you don't settle with mainstream and buys into HEDT space: Threadripper obliterates Intel here. You pay more, sure but you can't get anything that even comes close with Intels mainstream offerings. Intel shutdown their HEDT segment in 2019. AMD forced them.

270K Plus is decent value, if you want to accept a dead platform with no upgrade path. With AM5 you get an actual upgrade path, with a +50% core increase next time with Zen 6 and rumours about Zen 7 support as well. You can be 100% sure that Zen 6 with up to 24C/48T will make 270K Plus look dated in terms productivity, while bringing better gaming performance as well.

With 270K Plus, you buy into a dead platform that will get nothing in the future. Even 290K Plus was scrapped. Intel abandons this socket in half a year as their entire mainstream focus shifts to socket 1954, which hopefully gets 4 full generations instead of socket jumping every other generation to sell people a new motherboard/chipset.

I was talking about mainstream platforms. If you can afford HEDT, then go with HEDT, and Intel is dead in that segment.

That being said, for mainstream platforms, while long-term platform support is a huge benefit, it becomes kinda moot when the cost difference is that large. The 270K is $320 and you can drop it into a decent ~$150 B860 board and have 9950X peformance. That means Intel's CPU + platform is still cheaper than AMD's CPU alone. AMD is being a bit delusional with Zen 5 prices in the current environment, if they drop the price of the 9950X to $400-ish then I agree they win due to platform longevity.

 

[Theinsanegamer]

That entirely depends on which productivity tasks you want to do. Some don't scale well with threads, others like a big fast L2 cache. And if you truly need good productivity performance, you don't settle with mainstream and buys into HEDT space: Threadripper obliterates Intel here. You pay more, sure but you can't get anything that even comes close with Intels mainstream offerings. Intel shutdown their HEDT segment in 2019. AMD forced them.

270K Plus is decent value, if you want to accept a dead platform with no upgrade path. With AM5 you get an actual upgrade path, with a +50% core increase next time with Zen 6 and rumours about Zen 7 support as well. You can be 100% sure that Zen 6 with up to 24C/48T will make 270K Plus look dated in terms productivity, while bringing better gaming performance as well.

With 270K Plus, you buy into a dead platform that will get nothing in the future. Even 290K Plus was scrapped. Intel abandons this socket in half a year as their entire mainstream focus shifts to socket 1954, which hopefully gets 4 full generations instead of socket jumping every other generation to sell people a new motherboard/chipset.

Dead platform doesn't matter for 99% of people. They run the CPU till it can't do the job, at which point the socket has been replaced and new motherboards with new interfaces and better memory support are available.

Most people who bought into ryzen 5000s did so on 500 series boards, not 400 or 300 series.

 

[poshflamingos]

Dead platform doesn't matter for 99% of people. They run the CPU till it can't do the job, at which point the socket has been replaced and new motherboards with new interfaces and better memory support are available.

Most people who bought into ryzen 5000s did so on 500 series boards, not 400 or 300 series.

Now this is just unsourced nonsense again. You're erring by going to the opposite extreme.

Platform longevity is a huge benefit and many people who have access to it (I.e. not Intel customers) do take advantage of it. Someone who builds an AM5 workstation today can just drop a Zen 6 chip in as soon as it releases, which is a huge advantage over Intel. The only reason Intel users don't do the same is because they can't, but even Intel themselves are moving to longer term support for their next socket starting with Nova Lake. Once Intel also has long term platform support, Intel fanboys like yourself will start singing praise to it just as much.

To clarify, the only reason the 270K Plus is an exception to this is due to how cheap it is, which offsets the cost of the dead platform. If it weren't undercutting the AMD competitor by that much, it would be a bad deal precisely because of the (bad) platform support.

Also, again, your claim that "most people who bought into ryzen 5000s did so on 500 series boards" has no basis in reality. You made this claim based on (alledged) personal observations you made on some random, undisclosed "forum". That's not any kind of valid data.

 

[Loadedaxe]

Now this is just unsourced nonsense again. You're erring by going to the opposite extreme.

Platform longevity is a huge benefit and many people who have access to it (I.e. not Intel customers) do take advantage of it. Someone who builds an AM5 workstation today can just drop a Zen 6 chip in as soon as it releases, which is a huge advantage over Intel. The only reason Intel users don't do the same is because they can't, but even Intel themselves are moving to longer term support for their next socket starting with Nova Lake. Once Intel also has long term platform support, Intel fanboys like yourself will start singing praise to it just as much.

To clarify, the only reason the 270K Plus is an exception to this is due to how cheap it is, which offsets the cost of the dead platform. If it weren't undercutting the AMD competitor by that much, it would be a bad deal precisely because of the (bad) platform support.

Also, again, your claim that "most people who bought into ryzen 5000s did so on 500 series boards" has no basis in reality. You made this claim based on (alledged) personal observations you made on some random, undisclosed "forum". That's not any kind of valid data.

You’re mixing valid points with a lot of assumptions and calling it “fact.”

Platform longevity is a benefit...no one’s arguing that. AMD AM5 is clearly positioned for longer support than most of Intel’s recent sockets, and that’s a real selling point. But acting like most people actually use that longevity is where this falls apart.

Most users don’t buy a CPU and then drop in a new one every generation. They do full platform upgrades every 3 - 5+ years. By the time they upgrade, they’re often moving to a new board anyway, for features, I/O, power delivery, or just because it makes more sense as a package deal. That’s been true for both AMD and Intel buyers, whether you agree or not.

And we’ve already seen the reality check with AM4. Yes, it had great longevity, yes, but how many people actually went from something like a 1600 - 3600 - 5800X3D on the same board? A small percentage. Most people either upgraded once at most, or just rebuilt entirely.

That doesn’t make longevity useless, but it’s not the universal advantage you’re making it out to be either.

On Intel, same thing. Intel Core Ultra 7 270K Plus being on a shorter lived platform matters on paper, but in practice its lower entry cost can absolutely offset that for a lot of buyers. Value isn’t just lifespan, it’s what you pay today vs what you actually use.

Also, the “fanboy” angle doesn’t really land well either. Recognizing that AMD has better platform longevity right now isn’t controversial. The pushback is against pretending that longevity automatically translates into real world value for most users...it often doesn’t.

And on the Ryzen 5000 board claim, fair enough, anecdotal data isn’t gospel. But neither is assuming widespread multi generation upgrading without actual data either. Both sides need evidence, not just “that’s how people use it" , that isn't any proof at all.

 

[Theinsanegamer]

Now this is just unsourced nonsense again. You're erring by going to the opposite extreme.

Platform longevity is a huge benefit and many people who have access to it (I.e. not Intel customers) do take advantage of it. Someone who builds an AM5 workstation today can just drop a Zen 6 chip in as soon as it releases, which is a huge advantage over Intel. The only reason Intel users don't do the same is because they can't, but even Intel themselves are moving to longer term support for their next socket starting with Nova Lake. Once Intel also has long term platform support, Intel fanboys like yourself will start singing praise to it just as much.

To clarify, the only reason the 270K Plus is an exception to this is due to how cheap it is, which offsets the cost of the dead platform. If it weren't undercutting the AMD competitor by that much, it would be a bad deal precisely because of the (bad) platform support.

Also, again, your claim that "most people who bought into ryzen 5000s did so on 500 series boards" has no basis in reality. You made this claim based on (alledged) personal observations you made on some random, undisclosed "forum". That's not any kind of valid data.

Imagine writing a diatribe about "unsourced claims", calling someone a fanboy, then not providing a single source for any of your counterclaims.

Now me personally, I wouldn't want to post something like that, making an absolute fool out of myself for the world to see, but good on you for embracing your lack of inhibition around poor choices.

You’re mixing valid points with a lot of assumptions and calling it “fact.”

Platform longevity is a benefit...no one’s arguing that. AMD AM5 is clearly positioned for longer support than most of Intel’s recent sockets, and that’s a real selling point. But acting like most people actually use that longevity is where this falls apart.

Most users don’t buy a CPU and then drop in a new one every generation. They do full platform upgrades every 3 - 5+ years. By the time they upgrade, they’re often moving to a new board anyway, for features, I/O, power delivery, or just because it makes more sense as a package deal. That’s been true for both AMD and Intel buyers, whether you agree or not.

And we’ve already seen the reality check with AM4. Yes, it had great longevity, yes, but how many people actually went from something like a 1600 - 3600 - 5800X3D on the same board? A small percentage. Most people either upgraded once at most, or just rebuilt entirely.

That doesn’t make longevity useless, but it’s not the universal advantage you’re making it out to be either.

On Intel, same thing. Intel Core Ultra 7 270K Plus being on a shorter lived platform matters on paper, but in practice its lower entry cost can absolutely offset that for a lot of buyers. Value isn’t just lifespan, it’s what you pay today vs what you actually use.

Also, the “fanboy” angle doesn’t really land well either. Recognizing that AMD has better platform longevity right now isn’t controversial. The pushback is against pretending that longevity automatically translates into real world value for most users...it often doesn’t.

And on the Ryzen 5000 board claim, fair enough, anecdotal data isn’t gospel. But neither is assuming widespread multi generation upgrading without actual data either. Both sides need evidence, not just “that’s how people use it" , that isn't any proof at all.

Plus the 300 series STILL relies on beta AGESA for 5000s, which can play havoc with power tables. And their memory support sucked. My cross hair x470 could not go over 2400 without major issues, so when I got a 5800x3d and wanted 3200 mhz memory, I had to get a new motherboard too.

Instead of chasing the upgrade dragon, had I just built a coffee lake PC and held on until the 7800x3d, I would have spent less money and had a better experience.

 

[SRB]

My question is "How long can Intel afford to dump their top processor at $300?" It's a great strategy to move units, but the "free" MB chipsets and subsidized processor can't be profitable for Intel. A recent analysis from a brokerage pointed out that while Intel's revenue is great, they're still NOT a profitable company.

 

[Hecate91]

I assume Intel can keep subsidizing their high end cpu's for cheap, because Intel has shown they don't have to be profitable, as long as they're getting help from the US govt and Nvidia.

However the fanboying with claiming no one needs socket longevity is getting very tiring, most people will never upgrade their PC, but we aren't talking about CPU's most people will use in the first place. LGA1851 is basically a one and done socket, because the refresh is what Arrow Lake should've been, not the mess it launched as being slower than 14th gen. The only redeeming feature Intel CPU's have are being cheap to offset the cost of RAM and an SSD, which is only a temporary offset for any enthusiast because you'll just be spending more later on another CPU and motherboard, instead of being able upgrade the CPU.

 

[poshflamingos]

Most users don’t buy a CPU and then drop in a new one every generation. They do full platform upgrades every 3 - 5+ years. By the time they upgrade, they’re often moving to a new board anyway, for features, I/O, power delivery, or just because it makes more sense as a package deal. That’s been true for both AMD and Intel buyers, whether you agree or not.

And we’ve already seen the reality check with AM4. Yes, it had great longevity, yes, but how many people actually went from something like a 1600 - 3600 - 5800X3D on the same board? A small percentage. Most people either upgraded once at most, or just rebuilt entirely.

And you're making those statements based on what?
Let's see the evidence backing up your claims that "most users don't buy a new CPU for the same motherboard", "most users only do full platform upgrades", and "only a small percentage of AM4 users upgraded their CPUs on the same board". If you can't provide the data that backs up those claims, then you're just like that other user I was responding to, this is just unsourced nonsense.

And on the Ryzen 5000 board claim, fair enough, anecdotal data isn’t gospel. But neither is assuming widespread multi generation upgrading without actual data either. Both sides need evidence, not just “that’s how people use it" , that isn't any proof at all.

1) I don't need "evidence" to say long platform support is positive for consumers. That's a blatantly obvious fact. It's you and the other commenter who are trying to claim that long platform support is not positive because nobody actually makes use of those upgrade options. That claim absolutely does need evidence.
2) You think both sides need evidence? So where's yours? How curious you make a whole comment like this completely devoid of any evidence. Let's see it.

 

[poshflamingos]

Imagine writing a diatribe about "unsourced claims", calling someone a fanboy, then not providing a single source for any of your counterclaims.

Now me personally, I wouldn't want to post something like that, making an absolute fool out of myself for the world to see, but good on you for embracing your lack of inhibition around poor choices.

You're the one who needs to provide source, not me. Do you comprehend how burden of proof works? You're the one who brought the claim that platform longevity is useless because nobody makes use of it, the burden is on you to provide the evidence that backs up your claim, not on me to provide the evidence that disproves it.

Go ahead, provide the evidence. It's very interesting how you've been given the opportunity to provide this evidence multiple times, and refused it repeatedly, isn't it?

Also, you repeatedly making nonsense claims without evidence and then trying to say someone else is "making an absolute fool out of myself for the world to see" is absolutely hilarious. Zero self-awareness.

 

[Loadedaxe]

You're the one who needs to provide source, not me. Do you comprehend how burden of proof works? You're the one who brought the claim that platform longevity is useless because nobody makes use of it, the burden is on you to provide the evidence that backs up your claim, not on me to provide the evidence that disproves it.

Go ahead, provide the evidence. It's very interesting how you've been given the opportunity to provide this evidence multiple times, and refused it repeatedly, isn't it?

Also, you repeatedly making nonsense claims without evidence and then trying to say someone else is "making an absolute fool out of myself for the world to see" is absolutely hilarious. Zero self-awareness.

You’re trying to turn this into a “gotcha” over burden of proof instead of addressing how people actually use their systems.

I never said platform longevity is “useless.” I said it’s overstated relative to real world behavior. There's a big difference.

There isn’t some clean, universal dataset that tracks “how many users swapped CPUs on the same board across generations.” That data simply doesn’t exist publicly in any meaningful, comprehensive way. So pretending this is a courtroom where one side can just drop a perfect citation is unrealistic...

What we do have are strong indicators, OEM and prebuilt sales dominate the market, those users don’t upgrade CPUs at all. DIY builders tend to upgrade infrequently, often alongside platform changes like DDR generations, PCIe, I/O, etc.

Even within AM4, the best example of longevity...most people didn’t do 2 - 3 CPU swaps. At most, they did one, if any.

That’s not “unsourced nonsense,” that’s a synthesis of how the market behaves.

Also, burden of proof cuts both ways. You’re claiming “many people take advantage of it” as if it’s established fact. Where’s your data quantifying that? Because anecdotes and assumptions go both directions.

So no...this isn’t about refusing evidence. It’s about acknowledging that, hard data on this specific behavior is limited and real world usage patterns don’t line up with the idealized “drop-in upgrade every generation” narrative.

You’re arguing theory. I’m talking about how people actually build and upgrade PCs. We are enthusiasts, most of us don't do half baked upgrades...maybe you do?

And here’s the part you’re avoiding....Even if platform longevity is a benefit (which it is), that doesn’t mean it’s a deciding factor for most buyers...especially when pricing, features, and total platform cost are involved.

But, if you have some data I am wrong, like you so eloquently stated, drop it for us to read...I'll wait.

 

[Rynoman17]

All I know is that for 4K gaming my 7700X is holding up remarkably well. I'll hold on to it and see what AMD's next set of chips will bring.

I Have the same chip and just wanted to say that I agree completely. It has been doing a fantastic job for me. I do have one question that I hope you can help me with. I was using an NZXT Kraken Elite 360, until recently when it completely failed. It never did a great job at keeping temps low. What cooler do you have for yours and do you have good results? Thanks!

 

[Megalomaniac]

I'm happy with my 285K.

 

[poshflamingos]

You’re trying to turn this into a “gotcha” over burden of proof instead of addressing how people actually use their systems.

You cannot make claims about "how people actually use their systems" without sources/data. You don't know how people actually use their systems. You either point to data that shows it, or you're just making an uninformed guess.

I never said platform longevity is “useless.”

The other commenter did. I was responding to him before you inserted yourself in the conversation.

I said it’s overstated relative to real world behavior. There's a big difference.

Except you do not know what "real world behavior" is, because you do not have data or sources to back it up.

There isn’t some clean, universal dataset that tracks “how many users swapped CPUs on the same board across generations.” That data simply doesn’t exist publicly in any meaningful, comprehensive way.

Thank you for admitting you have no clue what you're talking about.

OEM and prebuilt sales dominate the market, those users don’t upgrade CPUs at all.

Citation needed. Where is your source showing that home users that buy a prebuilt "never upgrade it at all"?

DIY builders tend to upgrade infrequently, often alongside platform changes like DDR generations, PCIe, I/O, etc.~

Citation needed. Where's the source that shows DIY builders "upgrade infrequently/only with platform changes"? Once again, you're making up nonsense with no basis in reality.

Even within AM4, the best example of longevity...most people didn’t do 2 - 3 CPU swaps. At most, they did one, if any.

Citation needed. While I would tend to believe "most people didn't do 2 or more upgrades, just 1", you still can't make that claim without sources.
Also, "most people did only 1 upgrade" already refutes the other guy who is claiming platform longevity is useless.

That’s not “unsourced nonsense,” that’s a synthesis of how the market behaves.

No, it is not lmao
You can't just pull a "synthesis of how the market behaves" out of your buttocks. To make a synthesis of how the market behaves you need data/sources showing how the market behaves, and you already admitted you don't have any.

Also, burden of proof cuts both ways.

LMAO
This is the funniest thing you could have written.
No, it does not, genius. The concept of burden of proof is very explicitly about how it DOESN'T cut both ways. It explicitly means that burden of proof is NOT on both people in an argument, but rather only on the person who made the initial argument. The other commenter is the one who made the initial argument that platform longevity is useless and that nobody makes use of it. The burden of proof is on him (and by extension you, since you're here to support him), not on me. He (and by extension you) is the one who has to produce evidence to support it.

You’re claiming “many people take advantage of it” as if it’s established fact.

Yes, see how I intentionally left it vague? I don't have the data either, so I made a claim that an undefined amount of people do take advantage of it. I didn't try to claim "most" (I.e. over 50%) people or "nobody" (I.e. close to 0%) takes advantage of it, because I don't have to. The burden of proof isn't on me. All I need to do is expose that he doesn't have a source and his claims are unfounded.

Where’s your data quantifying that? Because anecdotes and assumptions go both directions.

Burden of proof isn't on me. It wasn't me who made the initial claim.

So no...this isn’t about refusing evidence. It’s about acknowledging that, hard data on this specific behavior is limited and real world usage patterns don’t line up with the idealized “drop-in upgrade every generation” narrative.

LMAO again
If you don't have the data, what makes you think you can make any claims about "real world usage"? You don't have the data, you have no clue what the "real world usage" is. You cannot make any claims about what it aligns or doesn't align with because you don't have the data and you do not know.

You’re arguing theory. I’m talking about how people actually build and upgrade PCs.

You don't have data about how people actually build and upgrade PCs, so you don't know. You can only make ignorant guesses.

We are enthusiasts, most of us don't do half baked upgrades...maybe you do?

Citation needed. Where's the data showing that "enthusiasts" "don't do half baked upgrades"? How do you struggle so much with comprehending you can't make those claims without evidence?

And here’s the part you’re avoiding....Even if platform longevity is a benefit (which it is)

What you're saying is at odds with what the commenter you're trying to defend says.

that doesn’t mean it’s a deciding factor for most buyers...especially when pricing, features, and total platform cost are involved.

That's not what the argument is about. The other commenter claimed it has no value because "nobody uses it/99% of people don't use it". I'm saying that's unsourced nonsense, which it is.
This isn't about whether it's a deciding factor. Maybe it is, maybe it isn't, I don't know. If you have data about how much of a deciding factor it is, provide it. If you don't, accept you don't know either and you can't make any claims about it.

But, if you have some data I am wrong, like you so eloquently stated, drop it for us to read...I'll wait.

Burden of proof isn't on me. You're the one who has to provide data to support your (and the other commenter's you chose to defend) claims. You wrote an 11 paragraph reply and still failed to present any source whatsoever. Looks like my assertion that what you two wrote is unsourced nonsense is correct.

 

[Loadedaxe]

You cannot make claims about "how people actually use their systems" without sources/data. You don't know how people actually use their systems. You either point to data that shows it, or you're just making an uninformed guess.


The other commenter did. I was responding to him before you inserted yourself in the conversation.


Except you do not know what "real world behavior" is, because you do not have data or sources to back it up.


Thank you for admitting you have no clue what you're talking about.


Citation needed. Where is your source showing that home users that buy a prebuilt "never upgrade it at all"?


Citation needed. Where's the source that shows DIY builders "upgrade infrequently/only with platform changes"? Once again, you're making up nonsense with no basis in reality.


Citation needed. While I would tend to believe "most people didn't do 2 or more upgrades, just 1", you still can't make that claim without sources.
Also, "most people did only 1 upgrade" already refutes the other guy who is claiming platform longevity is useless.


No, it is not lmao
You can't just pull a "synthesis of how the market behaves" out of your buttocks. To make a synthesis of how the market behaves you need data/sources showing how the market behaves, and you already admitted you don't have any.


LMAO
This is the funniest thing you could have written.
No, it does not, genius. The concept of burden of proof is very explicitly about how it DOESN'T cut both ways. It explicitly means that burden of proof is NOT on both people in an argument, but rather only on the person who made the initial argument. The other commenter is the one who made the initial argument that platform longevity is useless and that nobody makes use of it. The burden of proof is on him (and by extension you, since you're here to support him), not on me. He (and by extension you) is the one who has to produce evidence to support it.


Yes, see how I intentionally left it vague? I don't have the data either, so I made a claim that an undefined amount of people do take advantage of it. I didn't try to claim "most" (I.e. over 50%) people or "nobody" (I.e. close to 0%) takes advantage of it, because I don't have to. The burden of proof isn't on me. All I need to do is expose that he doesn't have a source and his claims are unfounded.


Burden of proof isn't on me. It wasn't me who made the initial claim.


LMAO again
If you don't have the data, what makes you think you can make any claims about "real world usage"? You don't have the data, you have no clue what the "real world usage" is. You cannot make any claims about what it aligns or doesn't align with because you don't have the data and you do not know.


You don't have data about how people actually build and upgrade PCs, so you don't know. You can only make ignorant guesses.


Citation needed. Where's the data showing that "enthusiasts" "don't do half baked upgrades"? How do you struggle so much with comprehending you can't make those claims without evidence?


What you're saying is at odds with what the commenter you're trying to defend says.


That's not what the argument is about. The other commenter claimed it has no value because "nobody uses it/99% of people don't use it". I'm saying that's unsourced nonsense, which it is.
This isn't about whether it's a deciding factor. Maybe it is, maybe it isn't, I don't know. If you have data about how much of a deciding factor it is, provide it. If you don't, accept you don't know either and you can't make any claims about it.


Burden of proof isn't on me. You're the one who has to provide data to support your (and the other commenter's you chose to defend) claims. You wrote an 11 paragraph reply and still failed to present any source whatsoever. Looks like my assertion that what you two wrote is unsourced nonsense is correct.

You’re so focused on “burden of proof” that you’re missing the actual discussion.

You keep demanding hard data for any statement about user behavior, while simultaneously making your own claims like “many people take advantage of it” and hiding behind semantics to avoid backing it up. That’s not how this works either.

There is no public dataset that cleanly tracks multi generation CPU upgrades on the same motherboard across the entire market. You know that, I know that. So pretending this debate can be settled with a single citation is just deflection.
You’re trying to force this into “prove exact percentages or you can’t say anything,” which is just a way to shut down discussion, not engage with it.

If your position is simply “longevity has value,” then we agree. If your position is “it’s widely utilized and a major factor for most buyers,” that also needs evidence...whether you want to acknowledge that or not.

You’re not proving anything either, you’re just demanding sources while making assumptions of your own. You’re hiding behind “burden of proof” to avoid the actual point.

 

[poshflamingos]

You keep demanding hard data for any statement about user behavior

Because any statement about user behavior requires hard data. Without hard data, any claims you make about user behavior are unfounded and useless.

There is no public dataset that cleanly tracks multi generation CPU upgrades on the same motherboard across the entire market. You know that, I know that.

Oh, you know that, do you? Then what makes you think you can make claims about "real world behavior", "OEM and prebuilt users don’t upgrade CPUs at all", "DIY builders tend to upgrade infrequently", "a synthesis of how the market behaves", "how people actually build and upgrade PCs", and "enthusiasts don't do half upgrades" like you did on your last comment? You're admitting you don't have sources or evidence for any of these claims you made. So where are they coming from? Why are you dodging answering this?

You’re trying to force this into “prove exact percentages or you can’t say anything,” which is just a way to shut down discussion, not engage with it.

There's no discussion to shut down. You don't have any valid information for us to discuss. You're making up unbased nonsense and expecting me to accept it as true. That's not a discussion, and the only thing it deserves is to be shut down.
Provide the sources for your claims or GTFO.

If your position is simply “longevity has value,” then we agree.

WTF do you mean "if"? Did you not read my comments before responding to me? Why do you not even know what my point is?
I'm responding to someone who claimed platform longevity is useless because nobody takes advantage of it, by saying that platform longevity does have value, and that his claims are invalid because he doesn't have any source or evidence. And here you hit me with an "if" several comments in. You don't even know what you're arguing against.

If your position is “it’s widely utilized and a major factor for most buyers,”

Nowhere have I said this. I'm here disproving the other commenter, not making sweeping statements of my own.
Again, why are you replying if you didn't even bother reading what I wrote? You're just wasting everyone's time.

You’re not proving anything either, you’re just demanding sources

YES! That's exactly what I'm doing. I'm responding to the commenter who wrote unsourced nonsense and demanding his sources (that I know he doesn't have) to show that what he wrote is in fact unsourced nonsense. That's the entirety of my argument. That's how burden of proof works.
Then you jumped in and piled up even more unsourced nonsense on top of it. So, again, either show the sources, or accept that what you (and the other commenter) wrote is unsourced nonsense.

 

[Rapob]

Dead platform doesn't matter for 99% of people. They run the CPU till it can't do the job, at which point the socket has been replaced and new motherboards with new interfaces and better memory support are available.

Most people who bought into ryzen 5000s did so on 500 series boards, not 400 or 300 series.

Many people use 300 and 400 series chipset with Ryzen 5000. I know for a fact. I bet more people uses 300/400 than 500 series as few needs what 500 offer.. Ryzen 5000 is an easy slot-in upgrade on any 300/400 board. Did it at least 10 times.

And the only reason you think 99% don't care about platform support, is because you never had the option on Intel and you have zero AMD experience. It it always pointless to upgrade on same socket with Intel, unless you go from low-end to high-end maybe. Intel made sure of that. This will change with Socket 1954 tho.

I had Ryzen 3700X, upgraded to 5800X3D. Massive upgrade in gaming and it was no compromise compared to Intel with the 3700X + OC to begin with. The 5800X3D slapped Intel in gaming with absolute ease.

The 3700X replaced the 1700X in my server, which is also AM4, socket longevity matters for several reasons.

The 5800X3D was sold for 75% of the initial price, when I went AM5. 5800X3D has huge demand in second hand market.

I had 7800X3D, upgraded to 9800X3D and will probably upgrade to 10800X3D again, or Zen 7, we will see. Sold my 7800X3D for 300 bucks, paid 375 or so. X3D sells like hotcakes in the used market so the upgrade is cheap. Why do they sell easily? Because of good platform support. Tons of people can use them. Tons of people want them.

Think.

Even Intel knows platform support matters and officially said socket 1954 will get 4 generations, just like AMD... Funny how the tables turn when there is competition.

New interfaces, better memory support, a nope-bucket. Brings nothing really. Buy a decent board to begin with - AM5 with PCIe 5.0 x16, multiple M.2 x4, what more do I need for the next 5 years? Nothing. Even RTX 5090 performs at 100% on a PCIe 4.0 x16 or even 4.0 x8 / 3.0 x16 with almost 100% performance still.

X3D means memory speed won't matter much, cache is used instead of slower RAM. Intel requires fast and expensive RAM for good gaming performance, AMD don't. Hence why you can re-use the memory in pretty much all cases. Used the same 6000/30 RAM for my 7800X3D and 9800X3D. Overclocked them to 6400/28 on the 9800X3D, top tier dies.

No GPU will saturate my PCIe 5.0 x16 board till I replace the entire platform with AM6 or next-next Gen Intel (Whats coming after socket 1954). My RTX 4090 is not even close to maxing out a PCIe 5.0 x16 port... Even 4.0 x16 gives me 100% perf with OC/UV meaning 5-10% higher than stock 4090 perf.

You have alot of points, that just makes no sense.

Good platform support is the reason AMD dominate the CPU space. Many people upgraded CPU on the same board, instead of considering Intel.

Bad platform support on Intel, also forced many to try AMD because they needed to replace everything anyway.

So yeah, think.

 

[Rapob]

I was talking about mainstream platforms. If you can afford HEDT, then go with HEDT, and Intel is dead in that segment.

That being said, for mainstream platforms, while long-term platform support is a huge benefit, it becomes kinda moot when the cost difference is that large. The 270K is $320 and you can drop it into a decent ~$150 B860 board and have 9950X peformance. That means Intel's CPU + platform is still cheaper than AMD's CPU alone. AMD is being a bit delusional with Zen 5 prices in the current environment, if they drop the price of the 9950X to $400-ish then I agree they win due to platform longevity.

Intel is cheaper, because they need to be cheaper to actually sell well enough. Do you think Intel lowered their prices to be nice? AMD forced them to. Just like AMD forced Intel to drop their HEDT lineup many years ago.

AMD dominate the DIY space and did for years now. Threadripper and EPYC sells very well in HEDT space and datacenter segment too. Intel is struggling hard here, and often needs to undercut AMD to make people go Intel. Just like desktop, and laptop space.

Intel mostly sells in the OEM space, due to giving huge discounts on their chips, mostly 14th generation now, which they make on Intel fabs, hence cheaper to make (than Arrow Lake using TSMC - Higher margins on 14th gen). I do B2B for a living and know this for a fact. More and more companies buy AMD, both server and laptop.

They generally don't want to pay extra for Intel, like in the old days.
AMD lost the bad rep. This is nothing new really.

AMD is bigger than Intel these days. Luckily Intel is doing better today than a few years ago.

I don't actually want Intel to die. I want competition. And I bought Intel stock 1 year ago at the rock bottom.

The worst days are hopefully over for Intel now. They were almost dead and buried at some point. Removing the coffee for employees ... then you know business is bad.

 

[OortCloud]

The world of getting a new PC and graphics card every few years is at an end. At first I was sad about that because as an enthusiast I enjoy buying and building new hardware but now I'm ok with it as is my wallet. On the AMD side if you have AM4 then just stick with it for now, the gains for an AM5 board + DDR5 aren't worth the cost. If you have a low-end CPU look for a cheap or second-hand 57000/5800X3D etc. Same is true for graphics cards. Get something cheap/second-hand for now. 3080's for instance are very reasonable - probably because people are scared off by the 10GB VRAM - but it's enough for anything to run ok for now especially if you aren't targeting 4k. Even the worst culprit - Indiana Jones - runs fine on a 3080 especially if you lower texture quality a notch.

 

[Rapob]

The world of getting a new PC and graphics card every few years is at an end. At first I was sad about that because as an enthusiast I enjoy buying and building new hardware but now I'm ok with it as is my wallet. On the AMD side if you have AM4 then just stick with it for now, the gains for an AM5 board + DDR5 aren't worth the cost. If you have a low-end CPU look for a cheap or second-hand 57000/5800X3D etc. Same is true for graphics cards. Get something cheap/second-hand for now. 3080's for instance are very reasonable - probably because people are scared off by the 10GB VRAM - but it's enough for anything to run ok for now especially if you aren't targeting 4k. Even the worst culprit - Indiana Jones - runs fine on a 3080 especially if you lower texture quality a notch.

I went AM4 to AM5 and it was very much worth it. From 5800X3D to 9800X3D was a massive upgrade. Night vs day literally. From good to much better. Made me actually able to use my high refresh rate monitor to the fullest with a truly massive increase in 1% lows.

Let me guess, you are a GPU bound gamer thinking 60 fps is good enough?

What you claim entirely depends on requirements. It is subjective. I'd never settle with AM4 or a 6 year old GPU. No enthusiast would, unless money is the issue.

I see what you mean tho, but then you are not really an enthusiast anymore.

Some would also say 8700K and RTX 2070 is perfectly fine still, do you agree? It might be, for their 1080p 60 Hz IPS monitor. For me, dumpster worthy. It is all subjective.

 

[Ashford]

Checking in for the first time in ages. I stopped coming here because of the staff bias toward the most toxic people who frequent these boards.

Seems nothing has changed.

 

[Loadedaxe]

Because any statement about user behavior requires hard data. Without hard data, any claims you make about user behavior are unfounded and useless.


Oh, you know that, do you? Then what makes you think you can make claims about "real world behavior", "OEM and prebuilt users don’t upgrade CPUs at all", "DIY builders tend to upgrade infrequently", "a synthesis of how the market behaves", "how people actually build and upgrade PCs", and "enthusiasts don't do half upgrades" like you did on your last comment? You're admitting you don't have sources or evidence for any of these claims you made. So where are they coming from? Why are you dodging answering this?


There's no discussion to shut down. You don't have any valid information for us to discuss. You're making up unbased nonsense and expecting me to accept it as true. That's not a discussion, and the only thing it deserves is to be shut down.
Provide the sources for your claims or GTFO.


WTF do you mean "if"? Did you not read my comments before responding to me? Why do you not even know what my point is?
I'm responding to someone who claimed platform longevity is useless because nobody takes advantage of it, by saying that platform longevity does have value, and that his claims are invalid because he doesn't have any source or evidence. And here you hit me with an "if" several comments in. You don't even know what you're arguing against.


Nowhere have I said this. I'm here disproving the other commenter, not making sweeping statements of my own.
Again, why are you replying if you didn't even bother reading what I wrote? You're just wasting everyone's time.


YES! That's exactly what I'm doing. I'm responding to the commenter who wrote unsourced nonsense and demanding his sources (that I know he doesn't have) to show that what he wrote is in fact unsourced nonsense. That's the entirety of my argument. That's how burden of proof works.
Then you jumped in and piled up even more unsourced nonsense on top of it. So, again, either show the sources, or accept that what you (and the other commenter) wrote is unsourced nonsense.

You’re still arguing a different point than the one I’m making.

I already acknowledged that hard public data on exact upgrade behavior is limited. That was never controversial. The difference is I’m willing to admit uncertainty, while you’re treating the absence of perfect data as proof nobody can discuss market behavior at all.

That’s not how real world discussion works. You just want to argue.

People make observations about markets, trends, and buyer behavior all the time using a mix of public sales trends, platform adoption, OEM dominance, enthusiast habits, pricing behavior, and upgrade cycles. None of that becomes “invalid” simply because there isn’t a peer reviewed spreadsheet tracking every AM4 owner’s upgrade path.

And for the third time now, I never defended the “platform longevity is useless” argument. I specifically said longevity is a real benefit. My point was that its practical value varies depending on how the buyer upgrades.

You keep trying to drag me back to defending a claim I didn’t make because it’s easier than addressing the nuance of the actual point.

At this stage, your entire argument boils down to “If there’s no perfect dataset, nobody can make any generalized observations whatsoever.”

That’s an impossible standard that you yourself aren’t applying consistently either, because you’re still making assumptions about how widely platform longevity is utilized while refusing to quantify it.

So no, I’m not claiming certainty. I’m saying reality is more nuanced than “platform longevity automatically matters equally to everyone,” which is a perfectly reasonable position whether you like it or not.

At this point, arguing with someone who absolutely has to have the last word is like arguing with ..... you can fill in the blank.