Question Is the i7 10700K still a good option for a gaming PC?

[vultusprime]

Hi
I will be buying a new 1080p gaming PC in the near future.
AMD 5600X seemed like a good option, but considering that I would keep the new PC for at least 5 years, an 8 cores CPU looked more appealing.
Intel latest releases were a little disappointing, so I was thinking about the last generation i7 10700K.
What do you think?

 

[Rigg]

And before we return to our regularly scheduled programming; They do the same thing with GPUs. The one that really grinds my gears is the frames per dollar charts. Why the Fudge do I care what that value is, if I am getting 27 fps average at 1080p low? Am I suppose to pat myself on the back because I paid $1.50 per frame and not $4.50 per frame like that "poor sucker" playing at ultra on high refresh with buttery smooth performance? Or care that I could have been getting another 50 cents worth of frames by buying the competition's card? It is all smoke and mirrors. And like it or not, they condition us to the point where we repeat their silly rhetoric and marketing driven info to each other.

Back to our show: Buy what you want, but keep things in context. Because millions of dollars are being spent to ensure you don't.

This an absurd example. The frames per dollar charts are completely relevant (especially under normal market circumstances) when comparing GPU's in the same or similar performance tier. Yeah....comparing an RX 570 4gb to a 2080ti on this basis is irrelevant. Comparing a 6800xt/3080 to a 6900xt/3090 in this context is completely relevant. This should be completely obvious to anyone with an IQ high enough to install a graphics card. Do you assume the people looking at this data are so stupid they can't differentiate in what context the data is relevant? Especially when It's likely to be included in a review that has overall performance data for all of the same products in the FPS/Dollar chart? What is a better metric to asses value?

The whole "shill reviewers are everywhere" paranoia is getting really tired. Maybe I'm inferring something you're not implying, but it sure sounds like you are taking a veiled shot at hardware unboxed. It's pretty simple. Look at the data, figure out how it's relevant to you, and make your purchase decision.

 

[DAPUNISHER]

This an absurd example. The frames per dollar charts are completely relevant (especially under normal market circumstances) when comparing GPU's in the same or similar performance tier. Yeah....comparing an RX 570 4gb to a 2080ti on this basis is irrelevant. Comparing a 6800xt/3080 to a 6900xt/3090 in this context is completely relevant. This should be completely obvious to anyone with an IQ high enough to install a graphics card. Do you assume the people looking at this data are so stupid they can't differentiate in what context the data is relevant? Especially when It's likely to be included in a review that has overall performance data for all of the same products in the FPS/Dollar chart? What is a better metric to asses value?

The whole "shill reviewers are everywhere" paranoia is getting really tired. Maybe I'm inferring something you're not implying, but it sure sounds like you are taking a veiled shot at hardware unboxed. It's pretty simple. Look at the data, figure out how it's relevant to you, and make your purchase decision.

Your entire response was nothing but an insult from start to finish. You should lay off the passive aggressive ad hom nonsense, it detracts from any valid point you make. But you do you. Also, if I decide to throw shade at any particular reviewer, I won't be shy about it.

 

[Rigg]

Your entire response was nothing but an insult from start to finish. You should lay off the passive aggressive ad hom nonsense, it detracts from any valid point you make. But you do you. Also, if I decide to throw shade at any particular reviewer, I won't be shy about it.

That came out a bit snarkier than necessary. Apologies for that but It wasn't intended to be an insult. On the flip side, the examples you used in your post were pretty hyperbolic which detracts from any valid point that you made. It also displayed a pretty low intelligence expectation for the viewers/readers of the charts being discussed. This rubbed me the wrong way and I layed on the snark too heavy as a result. My bad.

I guess we can agree to disagree but I think charts like these are very useful for assessing the value proposition of a potential GPU purchase. At least under normal market conditions.

 

[Ranulf]

The 5800x is not too expensive. We are just being beaten over the head by reviewers, shills, viral marketing, and manipulative math, that says it is. The rest of the builds being roughly equal in cost, if you pay $150 more and keep it 5 years, that's a $30 a year difference. That is an inconsequential amount of money, in North America anyways.

Ah, yes... that manipulative math for pointing out the obvious truth. That $150 is upfront cost that means you have less to spend elsewhere if on a budget. The 5800X is $50 more expensive than the 3800X was, MSRP to MSRP. Nevermind that the the 3800X was not well reviewed, much like the 3600X simply based on their prices. The 3800X was on sale during the 2019 holidays, 5 months after its release in July. Down $60 or so to $340, with a free game if I remember.

If I'm stuck paying $450 for an 8 core cpu right now, I'd rather spend $100 more and get a 12 core. $10 less per core for a 5900X. I mean, its only $100 more and I keep it for 5 years, thats a $20 per year difference. That is an inconsequential amount of money, in NA anyways.

At least AMD is mostly winning the performance crown now that they want to charge Intel prices (already overpriced for years and even worse with this 11th gen). But I'm gonna criticize them the same I do Intel on pricing idiocy. AMD can get away with it simply because of high demand and low supply. If the silicon supply was normal, they'd either have the non x chips out by now or dropped the price on the 5800X. That or artificially kept its supply low because why not make more money selling 12 cores for $100 more.

 

[DAPUNISHER]

@Rigg

No worries, it is water off a duck's back with me. I take no personal offense. And thanks for the more measured reply.

You are not wrong about the hyperbole, but I was hoping the Peter Griffin what really grinds my gears schtick would tip off what was coming. And I put a comma between reviewers and shills, so the "shill reviewers are everywhere" comment was such a obvious strawman, that I thought you were baiting me, since I don't know your posting style yet.

And I will effort to explain myself better. Scientific testing methodology is a good thing, and it has value. But I cannot get on the bandwagon for value assessments like you posted above. It is based on canned benchmarks and short in game runs. Depends on their interpretation of how many DX11 and 12 games to test, and which ones. And is further dependent on numbers I won't be able to reproduce at home, since I am not using the same hardware.

Don't get me wrong I am a fan of benchmark Steve. And he stirred the pot with the Nvidia driver overhead investigation. Which brings us back to those frame per dollar charts. For those affected by the issue, the frame per dollar chart goes right out the window. And there are other problems I have with anyone deciding the value of a product that way. But I will leave it at that.

 

[Rigg]

Ah, yes... that manipulative math for pointing out the obvious truth. That $150 is upfront cost that means you have less to spend elsewhere if on a budget. The 5800X is $50 more expensive than the 3800X was, MSRP to MSRP. Nevermind that the the 3800X was not well reviewed, much like the 3600X simply based on their prices. The 3800X was on sale during the 2019 holidays, 5 months after its release in July. Down $60 or so to $340, with a free game if I remember.

If I'm stuck paying $450 for an 8 core cpu right now, I'd rather spend $100 more and get a 12 core. $10 less per core for a 5900X. I mean, its only $100 more and I keep it for 5 years, thats a $20 per year difference. That is an inconsequential amount of money, in NA anyways.

At least AMD is mostly winning the performance crown now that they want to charge Intel prices (already overpriced for years and even worse with this 11th gen). But I'm gonna criticize them the same I do Intel on pricing idiocy. AMD can get away with it simply because of high demand and low supply. If the silicon supply was normal, they'd either have the non x chips out by now or dropped the price on the 5800X. That or artificially kept its supply low because why not make more money selling 12 cores for $100 more.

I agree. Upon seeing the MSRP's for Zen 3 it was pretty obvious to me the 5800x was the odd man out in the lineup. I didn't need a reviewer to tell me that. In fact I would be pissed if reviewers weren't pointing out the obvious. Especially after testing them. The same logic applies to 6800/6900xt GPU's vs. 6800xt.

The 5800x would be nearly irrelevant in any normal market IMO.

 

[DAPUNISHER]

Ah, yes... that manipulative math for pointing out the obvious truth. That $150 is upfront cost that means you have less to spend elsewhere if on a budget.

I don't disagree with you on most of your post, and you callout anything you want, your prerogative. On the quote above, every amount is a budget, but I know what you meant. Next, build the system, and tell me what the percentage difference is in the total build cost. Hint it is rather insignificant. And I KNEW when I typed that the way I did someone was going to pull the slippery slope card. We are spending this much, why not just go hog wild? Whatev.

 

[DigDog]

i was gonna say something but moonbogg already did.

 

[VirtualLarry]

I think that @DAPUNISHER 's point, correct me if I'm wrong, is that if you base your GPU purchase primarily off of the "cost per FPS" chart, you're doing it wrong, because you'll pretty-much always end up in GTX 1050 ti 4GB territory or something for a GPU choice. And have an unsatisfactory gaming experience because of that.

 

[Rigg]

@Rigg

No worries, it is water off a duck's back with me. I take no personal offense. And thanks for the more measured reply.

You are not wrong about the hyperbole, but I was hoping the Peter Griffin what really grinds my gears schtick would tip off what was coming. And I put a comma between reviewers and shills, so the "shill reviewers are everywhere" comment was such a obvious strawman, that I thought you were baiting me, since I don't know your posting style yet.

And I will effort to explain myself better. Scientific testing methodology is a good thing, and it has value. But I cannot get on the bandwagon for value assessments like you posted above. It is based on canned benchmarks and short in game runs. Depends on their interpretation of how many DX11 and 12 games to test, and which ones. And is further dependent on numbers I won't be able to reproduce at home, since I am not using the same hardware.

Don't get me wrong I am a fan of benchmark Steve. And he stirred the pot with the Nvidia driver overhead investigation. Which brings us back to those frame per dollar charts. For those affected by the issue, the frame per dollar chart goes right out the window. And there are other problems I have with anyone deciding the value of a product that way. But I will leave it at that.

Those are Steve's charts (if you aren't aware) and he doesn't do canned benches. The short runs argument is fair but the API/Driver overhead argument would seemingly apply to any benchmark chart.

I guess I just don't get the argument. Buying a computer component is nearly always a value judgment. While some people just buy god tier kit priced be damned, (which is fine) most people are looking to hit a sweet spot in price/performance. It seems to me most people are making something similar to this calculation in their head when looking at regular bench charts anyway.

 

[DAPUNISHER]

I think that @DAPUNISHER 's point, correct me if I'm wrong, is that if you base your GPU purchase primarily off of the "cost per FPS" chart, you're doing it wrong, because you'll pretty-much always end up in GTX 1050 ti 4GB territory or something for a GPU choice. And have an unsatisfactory gaming experience because of that.

No, my opinion is way more unpopular than that. I don't think they should be doing it at all.

 

[VirtualLarry]

No, my opinion is way more unpopular than that. I don't think it they should be doing it at all.

I guess that we'll have to agree to disagree, then, because I see no difference between cost-per-CPU-core comparisons, and "cost per FPS" comparisons, arranged in an upward pattern of performance, LogicalIncrements-style, allowing a user to determine how much more $$$ to spend, to get how many more FPS or CPU cores or whatnot, as part of the build budget calculus.

I just don't see that the average "cost per FPS" should be the primary basis for that calculus, without taking absolute performance in the game(s) that you want to play, or the CPU core(s) in the applications that you want to run, into account.

 

[DAPUNISHER]

Those are Steve's charts (if you aren't aware) and he doesn't do canned benches. The short runs argument is fair but the API/Driver overhead argument would seemingly apply to any benchmark chart.

Unfortunately it doesn't, because they test with flagship CPUs.

 

[DAPUNISHER]

I guess that we'll have to agree to disagree, then, because I see no difference between cost-per-CPU-core comparisons, and "cost per FPS" comparisons, arranged in an upward pattern of performance, LogicalIncrements-style, allowing a user to determine how much more $$$ to spend, to get how many more FPS or CPU cores or whatnot, as part of the build budget calculus.

I don't like the cost per core either. Could they try to spoon feed us any harder? Maybe I am just being anti authority lately. Down with the man.

EDIT - I told you it was unpopular.

 

[Rigg]

Unfortunately it doesn't, because they test with flagship CPUs.

I don't follow.

I guess that we'll have to agree to disagree, then, because I see no difference between cost-per-CPU-core comparisons, and "cost per FPS" comparisons, arranged in an upward pattern of performance, LogicalIncrements-style, allowing a user to determine how much more $$$ to spend, to get how many more FPS or CPU cores or whatnot, as part of the build budget calculus.

I just don't see that the average "cost per FPS" should be the primary basis for that calculus, without taking absolute performance in the game(s) that you want to play, or the CPU core(s) in the applications that you want to run, into account.

Cost per frame charts would definitely be more useful on a per game basis. That's really a better argument against the usefulness of overall FPS averages though.

 

[Mopetar]

The 5800x is not too expensive. We are just being beaten over the head by reviewers, shills, viral marketing, and manipulative math, that says it is.

I think that's going a bit far, because here's the prices:

5600X: $300
5800X: $450
5900X: $550

To go from a 5600X to a 5800X you pay 50% more for 33% more cores. There's a slightly clock speed bump, but it's relatively negligible.
To go from a 5800X to a 5900X you pay 22% more for 50% more cores. Again you get a slightly higher boost clock, but largely negligible.

The 5800X is stuck in this weird position where you're almost always better off going with either of the two other CPUs. You either get a considerable cost saving dropping down to the 5600X without sacrificing a lot in gaming or other non-professional workloads or you get considerably more cores that will make a big difference in professional workloads for not too much additional money.

I think the market has spoken pretty clearly on this as well because the 5800X is almost always in stock whereas every other Zen 3 CPU is sold out. I suppose that "too expensive" is subjective, but I don't see how it's unfair to say that it's badly priced relative to other offerings. At $400 it wouldn't have that problem and I don't think you'd hear anyone grumbling about it.

 

[DAPUNISHER]

I don't follow.

The big OOF! is with older and slower CPUs like a 4790K or Ryzen 1400. It is not pronounced with a 10900K or 5900x.

I suppose that "too expensive" is subjective

It is subjective, that is my entire point. And $50 being a deal breaker, in THIS market, and ending up with an inferior platform or CPU, well it is just silly IMO. And I will assume providing the math is for the reader's edification and not mine, because otherwise Imma come at you bro!

 

[ondma]

AMD is somewhat justified for charging high prices now, because they have both the gaming and productivity lead. In isolation, one could argue that the 5800x is not "too expensive", but in the context of both the 5600x and 5900x, it still seems to lack value.

 

[Mopetar]

It is subjective, that is my entire point. And $50 being a deal breaker, in THIS market, and ending up with an inferior platform or CPU, well it is just silly IMO. And I will assume providing the math is for the reader's edification and not mine, because otherwise Imma come at you bro!

If your argument is between a 5800X and some other non-Zen 3 CPU that's an entirely different argument and one that I make no claims regarding. If, however, you've decided on a Zen 3 CPU then there's no good reason to get a a 5800X other than everything else is sold out so the real choice is buying a 5800X or waiting some indeterminant amount of time to get a better Zen 3 CPU. Either you can save $150 for negligible differences in gaming or you can spend another $100 for a substantial performance gain in productivity applications.

I only specifically mention $50 because it at least puts the 5800X more inline with the CPUs on either side of it in terms of what you get for what you pay. I'd argue that even at $400, you're still probably better off buying a 5600X or a 5900X depending on what you need. The 5800X just occupies this uncomfortable middle position where it's not so much better at the things a 5600X is good as to justify the extra cost, but similarly quite a bit worse at the things the 5900X is good at where it's hard not to justify the upgrade.

The only reason to consider it at all is that it will probably age a bit more gracefully than the 5600X, but right now you can just pocket that $150 and put it towards the next upgrade so it's not nearly as cut and dry.

 

[moonbogg]

The 5800X is stuck in this weird position where you're almost always better off going with either of the two other CPUs. You either get a considerable cost saving dropping down to the 5600X without sacrificing a lot in gaming or other non-professional workloads or you get considerably more cores that will make a big difference in professional workloads for not too much additional money.

This appears to be a technique designed to frustrate the consumer into buying the more expensive chip. Ignoring the 10700K screaming value in people's face, people looking for bang for buck will go with the 5600X. Most people looking for a fully capable gaming CPU these days are targeting 8 cores, and of course AMD knows that. So they charge a high price for 8 cores that's close in price to the 5900X, so the consumer is frustrated because they have to choose between getting a bad deal on 8 cores or paying even more for a better deal on 12 cores. I believe it's designed to frustrate.
This reminds me a lot of how Intel was so frustrating for a very long time, albeit for somewhat different reasons. You had to choose between 4 faster cores or 6 slower ones for a higher price. Which do you choose? Most value buyers always go with the cheaper ones anyway, but those looking for a more forward-looking, fully capable gaming CPU had a much tougher choice between the 2 platforms. It was so frustrating that it almost seemed by design to irritate people into skipping the cheaper option because 4 cores were getting dated even though they were faster at the time for most games. Meanwhile, Intel often refused to offer the most modern architecture for their HEDT platform, making you choose between the two when neither of them were more capable than the other in all areas.
At least Intel had the excuse of having two different platforms: one for work and one for play. AMD doesn't have that excuse. They choose to frustrate you into becoming a more profitable wallet with legs solely for the sake of doing so.

 

[Mopetar]

I don't know if that's the case because you can't buy a 5900X for the most part. That particular sales tactic only works if you can actually get the upgrade option and I don't even know if it's fair to call the 5900X that because if you look at how it's priced it offers more cores/dollar than the 5800X so it's just the better option from that perspective. Also, normally this type of tactic is used to upsell customers to the "middle" option like Apple tends to do where the entry-level model has some pathetic amount of memory, but for just $100 more, you can get 256 GB and not have to worry about it. Here, the entry option is great and it's the middle option that actively drives customers either high or low. Maybe the price only makes sense when AMD has the whole lineup shipped, but they don't have that so the price is just bad.

 

[moonbogg]

I don't know if that's the case because you can't buy a 5900X for the most part. That particular sales tactic only works if you can actually get the upgrade option and I don't even know if it's fair to call the 5900X that because if you look at how it's priced it offers more cores/dollar than the 5800X so it's just the better option from that perspective. Also, normally this type of tactic is used to upsell customers to the "middle" option like Apple tends to do where the entry-level model has some pathetic amount of memory, but for just $100 more, you can get 256 GB and not have to worry about it. Here, the entry option is great and it's the middle option that actively drives customers either high or low. Maybe the price only makes sense when AMD has the whole lineup shipped, but they don't have that so the price is just bad.

They were supposed to have the 5900X in stock. Their plans got SPIKED into the dirt ACE2 style when everyone got stuck at home and decided to buy a console instead.

 

[Mopetar]

They were supposed to have the 5900X in stock. Their plans got SPIKED into the dirt ACE2 style when everyone got stuck at home and decided to buy a console instead.

They use chiplets so it's not terribly difficult for them to adjust the product mix. They could probably take 100,000 chiplets that would make 100,000 5800X CPUs and turn them into at least 40,000 5900X CPUs with the remaining 20,000 still going towards making a 5800X.

I'm not sure what people deciding to purchase a console has anything to do with AMD's allocation of chiplets towards various products. The orders from Sony and Microsoft were always going to be large for the initial launch and require a sizeable number of wafers.

 

[blckgrffn]

I dont think that is totally fair to reviewers. Personally, 150.00 is not a lot of money to me, so I would not let the price deter me if the 5800x was the chip I wanted.

However, there *are* 2 problems with its price. First, there is only 1 sku available. Normally there is an initial model and a higher priced, slightly tweaked model comes later. The 5800x is more like the more expensive tweaked model, but there is no lower priced sku available. Secondly, the 5800x is bracketed by a cheaper 5600x with very good performance, and the 5900x which gives a lot more cores for not much more money. It is kind of academic now, though, since the 5900x has very poor availability.

It is my impression, given a background in retail, that AMD really wanted us to work our way up to the 5900x as a high volume - and I am guessing high margin/asp SKU.

If I could buy it with a board discount at MC it would be my current bogey. Given how many more cores you get for that $100, I really think AMD expected us to tip in at 5600x for the most sales, the 5800x as a stair stepper to the 5900x but for fully enabled ccx that binned out, and 5950x as a lowest volume halo.

My guess is AMD is frustrated with the volume of 5900x/5950x available as well.

 

[Rigg]

They use chiplets so it's not terribly difficult for them to adjust the product mix. They could probably take 100,000 chiplets that would make 100,000 5800X CPUs and turn them into at least 40,000 5900X CPUs with the remaining 20,000 still going towards making a 5800X.

I'm not sure what people deciding to purchase a console has anything to do with AMD's allocation of chiplets towards various products. The orders from Sony and Microsoft were always going to be large for the initial launch and require a sizeable number of wafers.

It is my impression, given a background in retail, that AMD really wanted us to work our way up to the 5900x as a high volume - and I am guessing high margin/asp SKU.

If I could buy it with a board discount at MC it would be my current bogey. Given how many more cores you get for that $100, I really think AMD expected us to tip in at 5600x for the most sales, the 5800x as a stair stepper to the 5900x but for fully enabled ccx that binned out, and 5950x as a lowest volume halo.

My guess is AMD is frustrated with the volume of 5900x/5950x available as well.

I wish they'd offer up the 5800 & 5900 OEM only cpu's to the DIY market. They'd probably have to drop the 5800x/5600x price a bit for it to make sense but they'd sure fill out the lineup nicely. It would be a really nice option for cheapskates, SFF builders, and overclockers. You'd have to assume that the quality of bins required for these aren't as good as what goes into the X counterparts and they'd have an easier time keeping them stocked.