Discussion Intel current and future Lakes & Rapids thread

[dullard]

Haha, pricing on CPUs, grouching about $850 chips that are gorillas. You guys forget this type of pricing?

Competition is good. That's what I am saying.

Competition is good, but competition doesn't have that much to do with pricing. A duopoly (two companies competing) generally drops prices about 20% compared to a monopoly. It is something, but not a whole lot. That 20% drop vanishes if companies can have marketable differences.

I remember the Intel/AMD MHz wars. Basically the same chips, just kept bumping up the MHz every month or so. Great increases in speed, almost identical competitors. What happened to prices? They went up ~$100/month. What really controls prices? Demand. That MHz war happened at the Y2K transition when demand skyrocketted.

AMD:


Intel:

 

[blckgrffn]

Competition is good, but competition doesn't have that much to do with pricing. A duopoly (two companies competing) generally drops prices about 20% compared to a monopoly. It is something, but not a whole lot. That 20% drop vanishes if companies can have marketable differences.

I remember the Intel/AMD MHz wars. Basically the same chips, just kept bumping up the MHz every month or so. Great increases in speed, almost identical competitors. What happened to prices? They went up ~$100/month. What really controls prices? Demand. That MHz war happened at the Y2K transition when demand skyrocketted.

I think it worth pointing out that back then we were largely accustomed to paying well over $1000 for a usable PC.

We've been a bit spoiled in the meantime.

Also, it further reinforces that having the top "desktop" SKU at $800 is ¯\_(ツ)_/¯ because 20 years ago you could have parted with more than that many dollars, regardless of inflation for a top flight SKU. It seems like flirting with $1k for the top desktop SKU across the last 20 year has been commonplace and should seem less and less outrageous.

I'd expect if Intel has a faster than 5950x CPU it will cost more than the 5950x. Seems pretty straightforward to me. It might be $999 even

 

[dullard]

I think it worth pointing out that back then we were largely accustomed to paying well over $1000 for a usable PC.

We've been a bit spoiled in the meantime.

Also, it further reinforces that having the top "desktop" SKU at $800 is ¯\_(ツ)_/¯ because 20 years ago you could have parted with more than that many dollars, regardless of inflation for a top flight SKU. It seems like flirting with $1k for the top desktop SKU across the last 20 year has been commonplace and should seem less and less outrageous.

I'd expect if Intel has a faster than 5950x CPU it will cost more than the 5950x. Seems pretty straightforward to me. It might be $999 even

I agree. Price points and marketing terms aren't great ways to distinguish CPU types. AMD has a red griffin, but Intel has a purple griffin, they are completely different!!!! Or not, those can be just marketing fluff. What distinguishes CPUs to me is performance and capabilities. Do I need 4 TB of memory? Do I need massive amounts of PCI lanes? Do I need ECC memory, etc. It sucks to pay $999, but if the product is worth the money, I'll do so. If not, I'll wait ~12 months for something equivalent that is far cheaper.

 

[Mopetar]

Yeah, but the child can provide feedback to the OS: "Hey, i am carrying heavy bricks for what is 45 meters now, help" - and OS can decide to move the heavy load from child to adult, even if there are 5 meters left, cause on average it will benefit.
This way it can react to characteristics of workload dynamically.

Unless the code being executed has some built-in functionality specifically designed such that it can be monitored and used for reporting purposes, that's just not possible in a general sense. It also doesn't even get into complex problems where a particular task is waiting on something else (e.g. disk access) such that it won't benefit in any speed up from being transferred to a more powerful core.

You can get some simplistic approaches to handling work distribution, but nothing that approaches intelligently solving the problem.

 

[JoeRambo]

Unless the code being executed has some built-in functionality specifically designed such that it can be monitored and used for reporting purposes, that's just not possible in a general sense. It also doesn't even get into complex problems where a particular task is waiting on something else (e.g. disk access) such that it won't benefit in any speed up from being transferred to a more powerful core.

You can get some simplistic approaches to handling work distribution, but nothing that approaches intelligently solving the problem.

It is complex problem, and some parts like "5 meters" left are obviously unknown to CPU and OS.

But just like children and adults can talk, CPU cores also can provide feedback to OS, and that is what Intel's "Hardware-guided scheduling" probably does - it provides hints in the form of extra performance counters about the load. Now what those hints contain so far is a bit of mystery, i have looked at LInux changes in the spring and there were still none.
Speculation - CPU can provide quite some characteristics about the thread being executed, like is it saturating small core resources, how rich the code is in SSE/AVX instructions and so on. OS can then take a look and make decisions about moving the load around.


Still my plan for Alder Lake is to buy it and disable atom cores and probably HT and put it on static clocks. Just like i don't miss HT on my static desktop 5.1Ghz 10900K, i won't be missing extra threads, while enjoying 30MB of L3 and Intel's IPC advances.
No hybrid, no HT means no problems for schedulers

 

[lobz]

Wow! You just erased every doubt I have for your AMD bias with this remark. The 5950x is an HEDT chip by every measure. The fact that AMD wanted to artificially maintain multithreaded superiority when Intel started releasing 8 core chips into the mainstream segment is not lost on some of us. AND, I doubt this chip is going to exhibit clear single-threaded weakness. You sound like you gave up the ghost. Relax. It's early days yet. The day @lobz types a concession is the day we'll all know for sure.

Man, take off the blinders and don't read only those posts that fit your opinion about people. I'm 100% sure ADL will have significantly higher 1T IPC than Zen 3 does, some of which the new Vcache will mitigate, but not all and certainly not in all cases, and I certainly believe that AMD really missed the proper release cadence this time. What I hate is every moronic rumor and the hyperboles built around them. I hated adored's 5GHz 8C Zen 2 for $250 nonsense with a passion, and so I hated RKL's legendary performance tales, justified by pages long mathematical equations using GB5 scores.... Granted I'm just an enthusiast and no technical expert at all, I tend to step into the conversation when someone shows symptoms of losing their minds, to try and apply some common sense when possible. I also like it when something is funny, regardless of which side/product/company/ideal it's about. As I said, blinders. Off. Please.

 

[Thunder 57]

Man, take off the blinders and don't read only those posts that fit your opinion about people. I'm 100% sure ADL will have significantly higher 1T IPC than Zen 3 does, some of which the new Vcache will mitigate, but not all and certainly not in all cases, and I certainly believe that AMD really missed the proper release cadence this time. What I hate is every moronic rumor and the hyperboles built around them. I hated adored's 5GHz 8C Zen 2 for $250 nonsense with a passion, and so I hated RKL's legendary performance tales, justified by pages long mathematical equations using GB5 scores.... Granted I'm just an enthusiast and no technical expert at all, I tend to step into the conversation when someone shows symptoms of losing their minds, to try and apply some common sense when possible. I also like it when something is funny, regardless of which side/product/company/ideal it's about. As I said, blinders. Off. Please.

That Zen 2 nonsense, ugh. I, along with a few others like author Joel Hruska, got a lot of crap from a lot of people for saying that rumor was bunk. It was nice to get the last laugh there, even if it meant higher prices.

 

[Asterox]

The person who tested Alder Lake 8+8 in Cinebench R20 score, i can bet that the same person intentionally does not want to show power consumption numbers.It would be very interesting to see, stock R9 5950X vs stock Alder Lake 8+8 Cinebench R20 red details.

If Alder lake has very good power efficiency, there is no reason to hide that red dragon.

AMD Ryzen 9 5950X Zen 3 CPU Review - KitGuru

AMD's new Zen 3 processors are finally upon us and they bring to the market a brand-new architecture

www.kitguru.net

 

[lobz]

That would mean he doesn't wanna show ST scores because they're bad too? You're reaching there.

 

[eek2121]

A thought I had this morning: assuming Intel doesn’t massively increase pricing, midrange is about to get way more exciting.

The Core i5 12600k will have 6 big + 4 little for a total of 10 cores/16 threads. If Intel keeps pricing where it is, it will be a great value at $270 or less. AMD will have to drop prices…

Of course, knowing Intel, they will probably jack up prices.

 

[blckgrffn]

A thought I had this morning: assuming Intel doesn’t massively increase pricing, midrange is about to get way more exciting.

The Core i5 12600k will have 6 big + 4 little for a total of 10 cores/16 threads. If Intel keeps pricing where it is, it will be a great value at $270 or less. AMD will have to drop prices…

Of course, knowing Intel, they will probably jack up prices.

VP's gotta eat too, you know. And shareholders. Think of the shareholders

 

[Thunder 57]

That Zen 2 nonsense, ugh. I, along with a few others like author Joel Hruska, got a lot of crap from a lot of people for saying that rumor was bunk. It was nice to get the last laugh there, even if it meant higher prices.

How silly of me to not include a link to the original article. Here it is. I go by a different name there though. Shouldn't be too hard to pick me out

VP's gotta eat too, you know. And shareholders. Think of the shareholders

Careful now, lest we get another warning about P&N.

 

[Thunder 57]

The person who tested Alder Lake 8+8 in Cinebench R20 score, i can bet that the same person intentionally does not want to show power consumption numbers.It would be very interesting to see, stock R9 5950X vs stock Alder Lake 8+8 Cinebench R20 red details.

If Alder lake has very good power efficiency, there is no reason to hide that red dragon.

AMD Ryzen 9 5950X Zen 3 CPU Review - KitGuru

AMD's new Zen 3 processors are finally upon us and they bring to the market a brand-new architecture

www.kitguru.net

Can we use bold and alternate colors less often please?

 

[Mopetar]

Can we use bold and alternate colors less often please?

You'd like to think that'd be nice, but it's just going to result in posts WITH A LOT OF CAPS LOCKS SECTIONS.

I lived through that dark age of the internet and I don't care to relive it again.

 

[scannall]

You'd like to think that'd be nice, but it's just going to result in posts WITH A LOT OF CAPS LOCKS SECTIONS.

I lived through that dark age of the internet and I don't care to relive it again.

Hey, at least we don't have Flash banner ads everywhere now. ;-)

 

[Mopetar]

Hey, at least we don't have Flash banner ads everywhere now. ;-)

Ah the good old days where you could punch the monkey while waiting for the rest of the page to load. Load times have certainly improved, but how many websites have gone from a single Flash banner ad to a page full of Javascript bloat ads and trackers? It was a more innocent time for sure, but I'm sure there are some pining for a time before that eternal September.

 

[DrMrLordX]

Hey, at least we don't have Flash banner ads everywhere now. ;-)

I want animated gifs like on JEFFK!!!!!!! 's site.

Bonus points if you remember jeffk

 

[JasonLD]

You'd like to think that'd be nice, but it's just going to result in posts WITH A LOT OF CAPS LOCKS SECTIONS.

I lived through that dark age of the internet and I don't care to relive it again.

Well the problem is half of his posts are borderline unreadable on dark mode lol.

 

[Thunder 57]

I don't. What was jeffk for us pleabs?

 

[jpiniero]

I don't. What was jeffk for us pleabs?

JEFF K."S BRAND NEW HOEMPAGE !!! NO HAX0RING!!!

You were warned...

 

[Mopetar]

Well the problem is half of his posts are borderline unreadable on dark mode lol.

I can see that colored text might cause issues (though it seems red shouldn't be an issue in dark mode) but I don't see how bolding something would cause a problem.

I think the easiest fix would be a personal setting that just strips out non-text formatting from posts entirely. I don't know if the forum has that kind of support built-in, but I would imagine that there are browser plug-ins that would definitely do it for you.

 

[dullard]

I can see that colored text might cause issues (though it seems red shouldn't be an issue in dark mode) but I don't see how bolding something would cause a problem.

Bolding isn't the problem. The problem is going from red text to black text (instead of default color. I think you have to use the remove color button to get default, no one should ever use black as a color, especially in software like Office since that conflicts with default). Although, honestly, we aren't missing much. That poster is here solely to bring up AMD in every single Intel thread.

 

[Mopetar]

Bolding isn't the problem. The problem is going from red text to black text (instead of default color).
View attachment 47594

Yeah that's kind of silly, but this exciting piece of information means that I can put hidden messages in my posts that only people in dark mode can see!

I apologize for being easily amused.

 

[Hulk]

Man, take off the blinders and don't read only those posts that fit your opinion about people. I'm 100% sure ADL will have significantly higher 1T IPC than Zen 3 does, some of which the new Vcache will mitigate, but not all and certainly not in all cases, and I certainly believe that AMD really missed the proper release cadence this time. What I hate is every moronic rumor and the hyperboles built around them. I hated adored's 5GHz 8C Zen 2 for $250 nonsense with a passion, and so I hated RKL's legendary performance tales, justified by pages long mathematical equations using GB5 scores.... Granted I'm just an enthusiast and no technical expert at all, I tend to step into the conversation when someone shows symptoms of losing their minds, to try and apply some common sense when possible. I also like it when something is funny, regardless of which side/product/company/ideal it's about. As I said, blinders. Off. Please.

If we are to believe the 12900K leaked Cinebench ST and MT scores then this part looks to be faster than the 5950X, at least in this bench. While we don't know how it will perform in actual applications that do useful work, I have to say I'm surprised and cautiously optimistic.

I'd love to know what happens with this MT score when the Gracemont cores are disabled?

 

[CHADBOGA]

If we are to believe the 12900K leaked Cinebench ST and MT scores then this part looks to be faster than the 5950X, at least in this bench. While we don't know how it will perform in actual applications that do useful work, I have to say I'm surprised and cautiously optimistic.

I'd love to know what happens with this MT score when the Gracemont cores are disabled?

Didn't rocketlake show far more improvement in Cinebench, than in most other applications?

I will wait for more benchmarks on other applications before getting too excited about Alderlake.