News Intel's 1500W TDP for Falcon Shores AI processor confirmed — next-gen AI chip consumes more power than Nvidia's B200

[bit_user]

Don't get me started on clothes. Can't buy a pure cotton sports socks if your life depended on it.

This.

On certain occasions, I need to avoid wearing synthetic fabrics (other than Nomex) and it's getting harder and harder to find certain clothing items in 100% cotton. Athletic socks are one of them.

If I were a conspiracy theorist I'd say that clothing manufacturers have been bought by the plastic industry as they seem to be dumping ever-increasing amounts of plastic waste into our clothes AND WE ARE PAYING FOR THE PRIVILEGE OF WEARING PLASTIC TRASH ON OUR SKIN.

Hey, the microplastics in the oceans and air didn't all put themselves there! We need to do our part!
; )

 

[bit_user]

"However, circular use of plastics and textiles could lead to the accumulation of a variety of contaminants in the recycled product. "

From what I've read, there's a steam-based cracking process for recovering the monomers from mixed plastic waste, which can then be used to form new plastics of nearly new quality. I have no idea how far away such processes are from being scaled up, but it should be doable and has been demonstrated in labs.

 

[TheHerald]

I said that the power and cooling requirements for CPUs and GPUs over the last couple of years seem to be seeping from HPC / datacenter into the desktop market because chip manufacturers sell basically the same HPC / datacenter chips just a bit gimped for desktop use

That's AMD, that's why I don't like their desktop chips. Consume too much power even for simple tasks. Their laptops and their APUs though are good, but they don't really have a high end option, so for desktop, buy an Intel if you care about your energy usage.

 

[bit_user]

That's AMD, that's why I don't like their desktop chips. Consume too much power even for simple tasks. Their laptops and their APUs though are good, but they don't really have a high end option, so for desktop, buy an Intel if you care about your energy usage.

Heh, nice try.

It depends on what you're doing. If you're just playing video games and browsing the web, then sure. However, where people tend to care most about energy efficiency is on heavy compute jobs. That's where Intel earns its reputation of being inefficient.

Now, we've been through this enough times that I know you understand this. I really expected better of you.

I'd point out that AMD does have power-efficient options available, for those with more modest compute requirements (although it still consistently outperforms a i9-12900). Here's an example of what's possible with their APUs:

Beelink GTR7 Changes the Mini PC Game with AMD Ryzen 7 7840HS

The Beelink GTR7 takes the best of the company's last three mini PC generations along with a new AMD Ryzen APU to make something great

www.servethehome.com

Idles at just 9.6 to 11 W.

 

[TheHerald]

Heh, nice try.

It depends on what you're doing. If you're just playing video games and browsing the web, then sure. However, where people tend to care most about energy efficiency is on heavy compute jobs. That's where Intel earns its reputation of being inefficient.

Now, we've been through this enough times that I know you understand this. I really expected better of you

​

Most people at home don't actually run blender 24/7. If they do then

A) They shouldn't buy an intel K part (there are T and non k options available)
B) If they did buy an intel k part they shouldn't run it power unlimited

If they follow A or B, Intel is fine / better than AMD even in MT performance. It's an easily fixable solution, idle and light load power draw isn't, it's inherent to multi die cpus. If Intel's new CPUs are similar and they hit up to 60w just browsing the web - trust me - I will be flaming the crap out of them as well.

Now, I'd add that AMD does have power-efficient options available, for those with more modest compute requirements. Here's an example of what's possible with their APUs:

​

Beelink GTR7 Changes the Mini PC Game with AMD Ryzen 7 7840HS

The Beelink GTR7 takes the best of the company's last three mini PC generations along with a new AMD Ryzen APU to make something great

www.servethehome.com

​

Idles at just 9.6 to 11 W.

But this is an APU. I already mentioned that, yes, the monolithic AMD chips are actually pretty good on the efficiency department. Their normal desktop parts though are not. They stink.

 

[bit_user]

It's an easily fixable solution, idle and light load power draw isn't, it's inherent to multi die cpus.

That's not what TweakTown found.

Source: https://www.tweaktown.com/reviews/1...zen-7600-7700-and-7900-cpus-tested/index.html
​

If Intel's new CPUs are similar and they hit up to 60w just browsing the web - trust me - I will be flaming the crap out of them as well.

If you consider web browsing to be mostly single-threaded, then AMD's 65W chiplet-based options look quite promising:

Source: https://www.techpowerup.com/review/amd-ryzen-7-7700-non-x/22.html
​

Let's see how that looks in terms of 1T efficiency:

​

Oh ...but gaming!?

​

I'm sure I don't even need to show MT efficiency, but for anyone curious:

Spoiler: More #winning

🧹 🧹 🧹

Right at the top of the list of Pros, TechPowerUp credits the R7 7700 with "Fantastic energy-efficiency". They were so pleased with the Ryzen 7 7700 they gave it their Editors Choice award.

​

The TweakTown review concludes:

"I've been pleasantly surprised by the range of AMD 65W CPUs launched today. Looking at gaming first, we saw minimal performance loss at 1080p when comparing these new SKUs to the original launch stack and, in some cases, a small uptick in performance. At 1440p and 4K, performance was still quite good, though down a few FPS compared to higher TSP parts.
...
Power numbers across the entire range were mind-blowing.
..."
​

They awarded it their Must Have recommendation.

​

the monolithic AMD chips are actually pretty good on the efficiency department. Their normal desktop parts though are not. They stink.

You're overgeneralizing. It's actually just the X-series that push Zen 4 outside its efficiency window.

 

[TheHerald]

That's not what TweakTown found.

​

Source: https://www.tweaktown.com/reviews/1...zen-7600-7700-and-7900-cpus-tested/index.html​

​

If you consider web browsing to be mostly single-threaded, then AMD's 65W chiplet-based options look quite promising:

​

Source: https://www.techpowerup.com/review/amd-ryzen-7-7700-non-x/22.html​

​

Let's see how that looks in terms of 1T efficiency:

​

​

Oh ...but gaming!?

​

​

I'm sure I don't even need to show MT efficiency, but for anyone curious:

​

Spoiler: More #winning

🧹 🧹 🧹

Right at the top of the list of Pros, TechPowerUp credits the R7 7700 with "Fantastic energy-efficiency". They were so pleased with the Ryzen 7 7700 they gave it their Editors Choice award.

​

​

The TweakTown review concludes:

​

"I've been pleasantly surprised by the range of AMD 65W CPUs launched today. Looking at gaming first, we saw minimal performance loss at 1080p when comparing these new SKUs to the original launch stack and, in some cases, a small uptick in performance. At 1440p and 4K, performance was still quite good, though down a few FPS compared to higher TSP parts.​

...​

Power numbers across the entire range were mind-blowing.​

..."​

​

They awarded it their Must Have recommendation.

​

​

You're overgeneralizing. It's actually just the X-series that push Zen 4 outside its efficiency window.

We've been through this a billion times yet you keep on going with your pro AMD propaganda BS. Facts are

1) AMD is horrible at idle and light workloads. Everybody who's used both know this, AMD cpus draw up to 5-10 times MORE power than the Intel equivalent. I've had them both side by side, 7w for the intel part, 40w for the amd part. Just drop this.

2) Intel cpus are more efficient in MT workloads and it's not even close. The only AMD cpu that is actually good in MT efficiency is the 7950x. Everything else kinda stinks.

Please, let's just stop it. I know it's popular to hate Intel and all that but let's focus on the facts. Sure, let's see some efficiency on a simulated 14900t and a 14900 (non k)

Spoiler

 

[bit_user]

We've been through this a billion times yet you keep on going with your pro AMD propaganda BS. Facts are

Your claims have been repeatedly shown to be based on questionable logic, shaky data, or both. I don't care what you claim. You need to provide quality data.

1) AMD is horrible at idle and light workloads.

This is directly contradicted by the data I cited from both TweakTown and TechPowerUp.

I've had them both side by side, 7w for the intel part, 40w for the amd part.

I can't debug your anecdote. I'm sure there was a reason for what you observed. Instead of just throwing up your hands and assuming the worst, maybe you should actually look into the matter, next time.

2) Intel cpus are more efficient in MT workloads and it's not even close.

Again, that's not supported by the data. It's right in the spoiler tag of my previous post. The only way you can make this claim stick is with wacky product match-ups that numerous people have challenged you on.

Please, let's just stop it.

The only way you're going to stop people challenging your wild claims is if you can support them with quality data, which has been sadly deficient in your posts.

The other way you can end these debates is to stop making claims we don't accept and that you can't support.

I know it's popular to hate Intel

I just bought an Intel 1 month ago and I have another on order. In contrast, the last AMD I bought was a Zen 3 upgrade for my fileserver, a couple years ago. I'm not an Intel hater. I just want clear and accurate information so people can make well-informed decisions.

P.S. Yes, the T-series CPUs are good for people who really care about efficiency and are willing to accept the tradeoffs. However, you can't transfer the efficiency wins of a i9-14900T over to other CPUs. What efficiency it achieves belongs to it, alone. Also, I'd point out that you can't buy T-series CPUs retail, which limits their applicability to the DIY market.

 

[TheHerald]

Your claims have been repeatedly shown to be based on questionable logic, shaky data, or both. I don't care what you claim. You need to provide quality data.

I have. Multiple times.

This is directly contradicted by the data I cited from both TweakTown and TechPowerUp.

No it didn't. You are just cherrypicking the data.

As I've said, we've been through here before, there is no way you are going to admi it, but let's go once more. Let's check the techpowerup so called "data" you provided.

Spoiler

Clear picture, cannot possibly get any clearer, but you will keep freaking arguing. Intel is lightyears ahead in light load efficiency like ST. It's a freaking slaughterhouse, even if you completely ignore the X parts. The non X parts are still getting hammered by Intel. Not only are they losing in efficiency, they are losing while being SLOWER!

But yeah, 2 days later you are going to pretend this never happened and post a graph again with half the CPUs missing to make your point, lol 😆😆

Again, that's not supported by the data. It's right in the spoiler tag of my previous post. The only way you can make this claim stick is with wacky product match-ups that numerous people have challenged you on.

Of course it's supported by data, the simulated 14900t and 14900 completely obliterate anything AMD has in MT efficiency. Nothing even gets close.

The only way you're going to stop people challenging your wild claims is if you can support them with quality data, which has been sadly deficient in your posts.

No, I have provided top quality data but 2 days later you pretend it never happened and start all over again. The techpowerup graph up there, we talked about it for 5 pages, and what did you just do? You ignored it to make your nonsensical point, that amd is efficient in light loads when it's freaking OBVIOUS that isn't. All modern high end AMD chips are at the freaking bottom of the efficiency chart. While being SLOWER as well.

 

[TheHerald]

P.S. Yes, the T-series CPUs are good for people who really care about efficiency and are willing to accept the tradeoffs. However, you can't transfer the efficiency wins of a i9-14900T over to other CPUs. What efficiency it achieves belongs to it, alone. Also, I'd point out that you can't buy T-series CPUs retail, which limits their applicability to the DIY market.

????

Of course you can

https://www.alternate.de/listing.xhtml?q=14900t

Sends to the whole EU.

But you don't have to. The point isn't that you should buy a 14900t, the point is it's completely dishonest and is obviously done on purpose to run blender on a 14900k without any power limits and then pretend it's not efficient. That's the point. If you bought a 14900k - you care about efficiency - and you want to run blender, just power limit the damn thing or buy a non k. You are being dishonest REPEATEDLY on purpose.

And I find it funny - cause if amd cpus were indeed as efficient as people are claiming, you wouldn't freaking need to show me graphs with power limits completely removed running blender to make your point, lol. Yet that's what people are doing all the time, as if people that care about efficiency buy 14900k cpus without power limits and run blender on a loop. Who are we trying to fool?

 

[bit_user]

No it didn't. You are just cherrypicking the data.

That's exactly what you're doing. You pick leaderboards where Intel sits at the top, and then pretend it applies to Intel's entire range of CPUs. That's why valid product matchups are so important.

Clear picture, cannot possibly get any clearer, but you will keep freaking arguing.

Clear as mud. Here's something you need to learn: Apples to Apples Comparisons.

Intel is lightyears ahead in light load efficiency like ST. ... The non X parts are still getting hammered by Intel. Not only are they losing in efficiency, they are losing while being SLOWER!

Nope. You can't pull efficiency data from one product and performance from another. If you're going to use the efficiency of the i5-13400F, then you also have to take its performance data. Because that's where we see the tradeoffs it makes for better efficiency.

​

Yes, it beat the R7 7700 on efficiency, but not performance!

Let's just recall how we got here. The only reason I went to the R7 7700 was to challenge your claim that AMD's chiplet-based CPUs are inherently inefficient at idle and ST tasks. I'm not saying that CPU's efficiency somehow rubs off onto the others in AMD's lineup.

2 days later you are going to pretend this never happened

Cool trick how you're asserting a "win" that never happened.

I have provided top quality data but 2 days later you pretend it never happened

No, I said you need sound logic and quality data. You can't make up for a lack of one with excess of the other. When making product comparisons, having a valid matchup is key.

The techpowerup graph up there, we talked about it for 5 pages,

What we talked about, for 5 pages, is why you need to match up like-for-like CPUs.

 

[bit_user]

????

Of course you can

https://www.alternate.de/listing.xhtml?q=14900t

You seem unaware of what the phrase "(Tray-Version)" means. That means it's not retail-boxed, but instead sold in bulk, as a tray of bare CPUs. When you buy one of these CPUs, you get just the CPU and typically any warranty must be through the seller.

If you had bought an i9-14900K via tray and started encountering errors with it, your warranty claim (at least in the USA) would be with the seller and you'd be at the mercy of whether they felt like honoring it.

the point is it's completely dishonest and is obviously done on purpose to run blender on a 14900k without any power limits and then pretend it's not efficient.

I never quoted any graph that didn't include the CPUs running at stock settings. If any graph included both stock + unlimited, I always intended you to look at the Stock settings, unless I explicitly stated otherwise.

If you bought a 14900k - you care about efficiency - and you want to run blender, just power limit the damn thing or buy a non k.

I'm fine with you citing data for a non-K CPU, as long as the efficiency and performance are taken together and they're compared with other CPUs of that class.

It's when you get into tweaks like undervolting that we run into trouble. We see now that Intel spec'd these CPUs like they did for reasons. Most people don't custom tweak their settings and motherboard BIOS cannot be trusted to do it safely. The only solution we're left with is comparing at Intel stock settings. Those are the settings guaranteed to work for every CPU and not susceptible to the "golden sample" problem and they characterize how most users will experience these products.

If you want to make broader points about the architecture, then microbenchmarks are really the best way to do that and we really need to look at perf/W curves - not single-point measurements.

You are being dishonest REPEATEDLY on purpose.

My rejection of your argument doesn't equate to dishonesty. I always cite a problem with the logic or the data, if I disagree with something. Furthermore, lack of agreement is no reason to start calling people dishonest. That's getting into ad hominem territory, which is not justified simply because you're frustrated.

 

[TheHerald]

That's exactly what you're doing. You pick leaderboards where Intel sits at the top, and then pretend it applies to Intel's entire range of CPUs. That's why valid product matchups are so important.


Clear as mud. Here's something you need to learn: Apples to Apples Comparisons.


Nope. You can't pull efficiency data from one product and performance from another. If you're going to use the efficiency of the i5-13400F, then you also have to take its performance data. Because that's where we see the tradeoffs it makes for better efficiency

​

No, that's exactly what YOU are doing. As I've explained to you AGAIN for 5 freaking pages (and you keep ignoring) im doing exactly that, comparing equally performing products to how efficient they are.

The i5 13400f has the same ST performance as the 7800x 3d, and it absolutely nails it in efficiency.
The i5 13600k has the same ST performance as the 7700x, and guess what....yeah. Not pretty.

And as we move further up the stack the difference start becoming HUGE. The 13700k is way faster than the 7900x while being way more efficient than both the 7900x and the7900. And let's not even talk about the 13900k / 14900k / 14900ks.

So what the actual heck are you talking about man? LOL.

to challenge your claim that AMD's chiplet-based CPUs are inherently inefficient at idle and ST tasks.

Well, you failed. The graph makes it clear, on average Intel cpus are faster in ST while being a lot more efficient. See this is a problem, the graph makes it obvious that Intel is way more efficient in ST tasks and you keep arguing.

I just took the average efficiency of all 12+13 gen cpus vs Zen 3 + zen 4. We got 92.5 pts / watt for Intel, 70.06 pts / watt for amd, yet here you are arguing it's not the case. Meanwhile Intel cpus are much faster on top of being more efficient.....😍

 

[TheHerald]

You seem unaware of what the phrase "(Tray-Version)" means. That means it's not retail-boxed, but instead sold in bulk, as a tray of bare CPUs. When you buy one of these CPUs, you get just the CPU and typically any warranty must be through the seller.

If you had bought an i9-14900K via tray and started encountering errors with it, your warranty claim (at least in the USA) would be with the seller and you'd be at the mercy of whether they felt like honoring it.

I know what a tray is, my 13900k is a tray CPU. In EU warranty is covered by the seller anyways. But doesn't matter, right below it there is a boxed 14900.

It's when you get into tweaks like undervolting that we run into trouble.

I agree, I never use undervolted data for comparisons unless both CPUs are undervolted, else they are meaningless.

Most people don't custom tweak their settings and motherboard BIOS cannot be trusted to do it safely. The only solution we're left with is comparing at Intel stock settings. Those are the settings guaranteed to work for every CPU and not susceptible to the "golden sample" problem and they characterize how most users will experience these products.

Then those "most people that don't go into the bios even to enable XMP" should not buy a K cpu to run blender on a loop.

My rejection of your argument doesn't equate to dishonesty. I always cite a problem with the logic or the data, if I disagree with something. Furthermore, lack of agreement is no reason to start calling people dishonest. That's getting into ad hominem territory, which is not justified simply because you're frustrated.

What's dishonest is posting power unlimited data to show that Intel CPUs are inefficient, when there are Intel CPUs that are power limited out of the box and top the efficiency charts. I've explained that millions of times. Posting data of a CPU that is MEANT to run inefficiently and then use that data to apply it to all intel cpus is very dishonest actually. If you don't like being characterized like that then stop posting those graphs.

 

[TerryLaze]

The 14900k is a 125W TDP CPU and if you run it at the official TDP it is more power efficient than the 7700x and basically matches the 7950x (128W avg) ,it is within 5% of the performance anyway, which is good enough for pretty much anybody if efficiency is a big factor for them.

73.7% vs 91.4% against the 7700x is a huge difference.
The 14900k limited to 65W matches the 7700x in performance while the 7700x uses 78W on average

Efficiency is not the same thing as "how dumb are users"
Link

 

[bit_user]

No, that's exactly what YOU are doing. As I've explained to you AGAIN for 5 freaking pages (and you keep ignoring) im doing exactly that, comparing equally performing products to how efficient they are.

The i5 13400f has the same ST performance as the 7800x 3d, and it absolutely nails it in efficiency.

No, establishing parity on a single benchmark doesn't make the products peers. I've said this a dozen times, at least: you need to consider performance, price, and power, when aligning products. And performance doesn't mean just on a single axis, but overall performance with special attention to how the CPU is marketed. The R7 7800X3D is marketing as a gaming CPU and indeed provides much better gaming performance than the i5-13400F (which is only 69.4% as fast). Most people intuitively understand this. I don't know why you're having so much trouble with it.

Spoiler: R7 7800X3D pwning the i5-13400F

The 13700k is way faster than the 7900x while being way more efficient than both the 7900x and the7900.

It's not "way" faster than the X, but it does take some wins on MT. Here are some cases where it doesn't:

Spoiler: R9 7900X MT wins

As for winning on efficiency vs. the 65W 7900, that's a negative.

​

What a weird thing to say. Now, let me point out that I only went there to refute your claim. I don't consider these products direct peers, because they differ so much in TDP and in other respects. As I've said before, I consider the R9 7900X to be the closest peer of the i7-13700K.

Well, you failed. The graph makes it clear, on average Intel cpus are faster in ST while being a lot more efficient. See this is a problem, the graph makes it obvious that Intel is way more efficient in ST tasks and you keep arguing.

I can talk to a brick wall, but I can't make it listen. The point of the graph was to show an example of a chiplet-based CPU with good ST efficiency.

See, if you make a sweeping claim and I show an example of where it doesn't hold, then your claim is invalidated. That's why you should avoid sweeping claims unless you actually have data that supports it, and clearly you didn't.

I just took the average efficiency of all 12+13 gen cpus vs Zen 3 + zen 4.

Please don't. That's just numerology. It's not the proper way to analyze microarchitectures and it doesn't actually say anything about specific product matchups. Stuff like that is just a waste of time. If you're interested in architectural comparisons, it would be a much better use of your time to read ChipsAndCheese.

 

[bit_user]

The 14900k is a 125W TDP CPU and if you run it at the official TDP it is more power efficient than the 7700x

Like I just told Harold, you can't simply take efficiency data from one CPU and put it up next to a CPU in a totally different class. That doesn't really tell us very much about these products and it certainly doesn't help anyone make purchasing decisions, because nobody is trying to decide between those two specific models!

 

[TheHerald]

No, establishing parity on a single benchmark doesn't make the products peers. I've said this a dozen times, at least: you need to consider performance, price, and power, when aligning products. And performance doesn't mean just on a single axis, but overall performance with special attention to how the CPU is marketed. The R7 7800X3D is marketing as a gaming CPU and indeed provides much better gaming performance than the i5-13400F (which is only 69.4% as fast). Most people intuitively understand this. I don't know why you're having so much trouble with it.

First of all, we were talking about ST efficiency (You literally brought it up). Now that the 3d is getting absolutely destroyed on it, you changed it to gaming. What in the world?

And why all of a sudden do you care about how a product is marketed? The R7 7700x is marketed as an i7 13700k competitor, hence the name (and the price), but you were adamant about me not comparing them against one another.......

It's not "way" faster than the X, but it does take some wins on MT. Here are some cases where it doesn't:

It is way faster in ST performance, which is EXACTLY what we were talking about. You are again changing the subject.

As for winning on efficiency vs. the 65W 7900, that's a negative.

​

Again, we were talking about ST efficiency. You are changing the subject for a 3rd time in a row, lol. How can you do that and then complain about me calling you dishonest? 😆

What a weird thing to say. Now, let me point out that I only went there to refute your claim.

Good job, posting MT data really refuted my point that Intel has a huge lead in ST efficiency. You win

I can talk to a brick wall, but I can't make it listen. The point of the graph was to show an example of a chiplet-based CPU with good ST efficiency.

But it doesn't have good ST efficiency, what are you talking about? The 400f still have a 37% lead man.....

See, if you make a sweeping claim and I show an example of where it doesn't hold, then your claim is invalidated. That's why you should avoid sweeping claims unless you actually have data that supports it, and clearly you didn't.

Of course it holds. The majority of Intel cpus have better ST efficiency while being faster compared to amd's desktop parts. That's a fact, and the graph shows it clearly. Yes, an avid AMD fan won't accept it, but there is nothing that can convince that kind of person so I'm out of luck on that one.

 

[TheHerald]

The 14900k is a 125W TDP CPU and if you run it at the official TDP it is more power efficient than the 7700x and basically matches the 7950x (128W avg) ,it is within 5% of the performance anyway, which is good enough for pretty much anybody if efficiency is a big factor for them.

73.7% vs 91.4% against the 7700x is a huge difference.
The 14900k limited to 65W matches the 7700x in performance while the 7700x uses 78W on average

Efficiency is not the same thing as "how dumb are users"
Link

the 14900k limited to 65w pulls an average of 45w. They have a graph with the total power pulled for the whole application suite

But I agree, when it comes to Intel, efficiency is measured by removing power limits, using LN2 for cooling and pressing the blender button. Intel is just doomed, no matter how much better their products are, they are going to get heavily criticized. They had top gaming performance (by a big margin btw) for so many years and people only cared about MT performance. Now that they have huge MT performance on all segments, people really care that the 7800x 3d is 5% faster than the equivalent intel cpu in games even though it's half as fast in MT. It would be funny if it wasn't true, but sadly it's true.

 

[bit_user]

First of all, we were talking about ST efficiency (You literally brought it up). Now that the 3d is getting absolutely destroyed on it, you changed it to gaming. What in the world?

I'm not the one who broadened the scope, but I've said all I have to say about that specific CPU matchup.

And why all of a sudden do you care about how a product is marketed?

That's relevant to how people use it. If you take a gaming-oriented CPU and show it's not great at rendering, that doesn't make it a bad CPU overall. It just means you shouldn't buy it if your main concern is rendering performance.

The R7 7700x is marketed as an i7 13700k competitor,

No, that's different (assuming it were even true). I was talking about the selling point of a specific model, and now you're trying to gin up a theory to legitimize you wacky product matchups.

hence the name (and the price), but you were adamant about me not comparing them against one another.......

Wow, so you remember at least some of what I said! That's progress, I suppose.

Okay, so the parts you left out were:

• AMD's numbering scheme is tied to the amount of cores they have, among other things. Ryzen 7000 is now AMD's 3rd generation of desktop products to use the same scheme, so don't pretend it's something new they just came up with.
• In general product numbering schemes are relevant for comparing products from the same manufacturer between different generations. You can't simultaneously have inter-generational naming consistency and use the names to try and align with competing products.
• It's also incredibly tricky to pre-align your products with your competitors, before either have launched, because you have no final performance data on either. That makes the whole notion they even attempted this that much more far-fetched.
• MSRP is similarly problematic as the point above, because you're merely taking a guess about how performance is going to match up. MSRP is set well in advance of product launches, so final specs & performance data isn't available.

I think that about covers it. So, basically the main factors that should be used to establish product matchups are: (current) price, performance, and power.

It is way faster in ST performance, which is EXACTLY what we were talking about. You are again changing the subject.

If I'm talking about ST performance & efficiency of one CPU, and then you decide to use it as a basis for establishing another product matchup, that's not me changing the subject.

Good job, posting MT data really refuted my point that Intel has a huge lead in ST efficiency.

You didn't say "ST efficiency". You just made another of your sweeping claims.

"And as we move further up the stack the difference start becoming HUGE. The 13700k is way faster than the 7900x while being way more efficient than both the 7900x and the7900."
​

If your wording is sloppy, that's your problem not mine. Maybe take more time to compose and proof-read your posts, instead of always being so quick to reply. A good reply is worth a lot more than a quick one.

But it doesn't have good ST efficiency, what are you talking about? The 400f still have a 37% lead man.....

You used efficiency data from one model and performance from another, in the same claim. That's a no no. The i5-13400F beats the R7 7700 on ST efficiency, but not ST performance. Both must be taken into account.

Of course it holds.

"... idle and light load power draw isn't (fixable, it's inherent to multi die cpus."
​

So, that's the point I was attempting (successfully, I might add) in refuting.

The majority of Intel cpus have better ST efficiency

But, if I can cite an example of a chiplet-based CPU with good efficiency, then perhaps it's not chiplets that are the main reason.

 

[bit_user]

the 14900k limited to 65w pulls an average of 45w. They have a graph with the total power pulled for the whole application suite

Heh, that's because those apps aren't all multithreaded. With either Intel or AMD, a well-threaded benchmark will pretty much max out the power limit. With ST tasks, they obviously don't. So, if you set PL2 at 65 W and average together MT and ST, then you're guaranteed to get a number less than 65 W. That's how math works.

 

[TheHerald]

That's relevant to how people use it. If you take a gaming-oriented CPU and show it's not great at rendering, that doesn't make it a bad CPU overall. It just means you shouldn't buy it if your main concern is rendering performance.

Then every product in the history of products is great. If you ignore all the things that a product doesn't do well , you can make the worst products that ever existed look good.

AMD markets the 7800x 3d as a gaming CPU because it is pretty horrible (for the price) on anything but games. If it was actually good on anything other than gaming it wouldn't be marketed as just a gaming CPU.

Would you ever use that argument for the 14900ks? Guys, it's a Single Threaded oriented CPU, so let's ignore every negative it has and focus on single threaded performance. That would be silly, so I refuse to do it for the 7800x 3d as well.

If I'm talking about ST performance & efficiency of one CPU, and then you decide to use it as a basis for establishing another product matchup, that's not me changing the subject.


You didn't say "ST efficiency". You just made another of your sweeping claims.

​

"And as we move further up the stack the difference start becoming HUGE. The 13700k is way faster than the 7900x while being way more efficient than both the 7900x and the7900."​

​

If your wording is sloppy, that's your problem not mine. Maybe take more time to compose and proof-read your posts, instead of always being so quick to reply. A good reply is worth a lot more than a quick one.

There is nothing sloppy about it, we were talking about ST for like 3 posts in a row. Obviously within that context I mentioned that the 700k is way faster.

You used efficiency data from one model and performance from another, in the same claim. That's a no no. The i5-13400F beats the R7 7700 on ST efficiency, but not ST performance. Both must be taken into account.

I compared the 400f to the 7800x 3d and the 600k to the 7700x. Both share the same ST performance. The 7700 doesn't have a 13th gen CPU matching it in ST performance so we can't perfectly match it with anything. Still you can clearly see that it's not topping any charts in ST efficiency.

"... idle and light load power draw isn't (fixable, it's inherent to multi die cpus."

​

So, that's the point I was attempting (successfully, I might add) in refuting.


But, if I can cite an example of a chiplet-based CPU with good efficiency, then perhaps it's not chiplets that are the main reason.

In order to succeed you need to compare it with a non chiplet based cpu. There are plenty of those (I'm sorry, im too bored to start throwing links around) where the 7700 is compared to eg. the 7500f. It loses horribly, exactly because it's chiplet based.

Regardless, the 7700 is the best example of what a chiplet based CPU can get in ST efficiency, and it's anything but impressive. Look at the top 5 places of the chart below, that's what impressive looks like

Spoiler

Heh, that's because those apps aren't all multithreaded. With either Intel or AMD, a well-threaded benchmark will pretty much max out the power limit. With ST tasks, they obviously don't. So, if you set PL2 at 65 W and average together MT and ST, then you're guaranteed to get a number less than 65 W. That's how math works.

Yes, but exactly because those apps aren't multithreaded the 14900k is only 3% faster than the 7700. If they were fully MT it would be way faster.

 

[slightnitpick]

From what I've read, there's a steam-based cracking process for recovering the monomers from mixed plastic waste, which can then be used to form new plastics of nearly new quality. I have no idea how far away such processes are from being scaled up, but it should be doable and has been demonstrated in labs.

That's probably my last link. Best part in that link is that the scientist behind it has also used some of this technology to clean up contaminants. So it might be possible to not only recycle some of these plastics but also clean up any contaminating chemicals at the same time.

 

[Silas Sanchez]

The 14900ks is a great example, it's - due to binning - the most efficient CPU intel has ever produced. It's now down to the user to utilize that efficiency by running sensible power limits or go kamehameha mode at 450 watts. But you can't be doing the latter and then complain that hardware isn't efficient nowadays.

This is what he keeps doing, making general blanket statements that are unsound. The problem is people like him can't see their flaws as they are full of ego and attitude. That is exactly what cognitive bias is, and why we have a discourse to open up to other experts so we have a system that weeds out our bias, he can't be part of that system because his full of ego and attitude. A rational system by nature weeds out people like him.

I'd question whether binning makes a CPU great, I think binning is more a marketing term used to prey on people all pumped up and is really just a word for sorting grades of quality. In semiconductors grades of quality fundamentally has to do with the mobility of electrons & holes, more mobility equals more faster switching possibly, which is affected by the quality of the doping impurities as bad doping affects the precision of the dynamic interplay between drift and diffusion. The level of impurities like defects and grain boundaries in the crystal lattice which shunt away current. The quality of the lithograph technique and technology primarily etc. Higher quality and luck give rise to variation and affect these things which can help reduce the heat produced of the cpu at higher frequency. But the real advancements are due to architecture and overall implementation.

Saying "the most efficient CPU intel has produced" is just an unsound statement. It's unsound because it makes assumptions of behalf of others, a very rudimentary mistake but one he is not aware of. He simply can't get it through his head that efficiency can mean many things and their are different sounding ways to speak of efficiency, but he is asserting his own definition in a blanket statement. In science we have precise definitions so we can all be in agreement and to avoid troublemakers who dont really understand the scientific method. but outside of these precise definitions there are other types of efficiencies, some being very abstract.

A solar cell is said to be very efficient as it represents years of engineering and research, yet from a physics pov they are morbidly inefficient as they leave over 80% of the suns energy on the table in the realworld. The new generation of thin films are very efficient in a way as they are very cheap to make even for their rather poor conversion efficiency and lifespan, yet from an environmental pov they are horribly inefficient due to their short lifespan and waste. For many people solar is inefficient as it takes many years for it to pay for itself when they could just use less electricity and spend that money elsewhere. For some who prioritize pleasure they can actually lead to unhappiness which means an inefficient way of living life.

 

[bit_user]

Then every product in the history of products is great. If you ignore all the things that a product doesn't do well , you can make the worst products that ever existed look good.

This is a simplistic argument. A better take is to say that not every product is right for every user. This is consistent with the idea that you cannot make a value judgement for other users. Just because one CPU has better MT performance than another doesn't mean it's a better fit for my needs. Only I know what are my priorities and can weigh its pros vs. its cons.

Seriously, if we could just assign one or two numbers that characterized the entirety of these products, don't you think people would do that? But, it's not that simple. That's why there are lots of different benchmarks and metrics. To make the optimal decision, each person needs to consider their constraints (cost, heat, noise, compatibility, etc.) and priorities and look at how the data on different products measures up.

AMD markets the 7800x 3d as a gaming CPU because it is pretty horrible (for the price) on anything but games.

"Horrible" is a strong word. Also, its efficiency is pretty good, especially considering other factors like its gaming performance.

And for someone whose top priority is gaming performance, maybe it's the best CPU for them. Only they can make that decision.

Would you ever use that argument for the 14900ks? Guys, it's a Single Threaded oriented CPU, so let's ignore every negative it has and focus on single threaded performance.

I can't do that and neither can you, because you don't know how important the different aspects are to someone else.

There is nothing sloppy about it, we were talking about ST for like 3 posts in a row. Obviously within that context I mentioned that the 700k is way faster.

Just take the time to say what you mean and then there won't be any confusion.

I compared the 400f to the 7800x 3d and the 600k to the 7700x. Both share the same ST performance. The 7700 doesn't have a 13th gen CPU matching it in ST performance so we can't perfectly match it with anything.

Matching in one axis of performance (and on one benchmark, no less) does not make two CPUs direct competitors. A corollary of this is that products don't have to be a perfect match in all respects, in order to be competitors.

In order to succeed you need to compare it with a non chiplet based cpu.

You just said:

"... idle and light load power draw isn't (fixable), it's inherent to multi die cpus."
​

I showed where a multi-die CPU can have decent power and efficiency. So, that's that.

If you are now trying to move the goalposts, then it's a different discussion.

Regardless, the 7700 is the best example of what a chiplet based CPU can get in ST efficiency, and it's anything but impressive. Look at the top 5 places of the chart below, that's what impressive looks like

Spoiler

It's an example. If we take the 5700G, that's interesting, but if someone wants better performance then it might not be an option for them. That's another example where it could be problematic to compare different classes of CPUs. It's like I said about the i5-13400F: you can't just look at its 1T efficiency in isolation. You also have to take into account its other attributes.