Discussion Intel current and future Lakes & Rapids thread

[dmens]

Singlethread performance is a result of frequency and IPC, this should be more than obvious.

And yet somehow so many people ignored RKL's obvious frequency limitations within a reasonable power envelope while drinking the IPC gainz kool-aid.

Simply ask yourself why they even bother working on faster iGPUs when it's just a monitor driver. Saying worst in class is strange, you should understand that not all mobile devices are suited for a dedicated GPU, be it for cost reasons, cooling reasons, size constraints etc. There is a big market for non dGPU devices and not everyone is fine with an UHD 620 like performance forever. A decent iGPU also helps removing the crap dGPUs from Nvidia, MX130 and MX230 are no longer a thing since Tigerlake.

Then why did you bring up the supposed perf gains of UHD750?

Also the reason you work on faster iGPU is so you can shrink it down even more for parity performance, thus saving cost and power.

 

[dmens]

Right now AMD is losing mostly in frame rates, check out ultrabookreview or notebookcheck. Maybe they should work on better driver optimizations. But why even bother? According to some people it's just a monitor driver and there is no need for something faster.

Hah. You might want to do a area/power normalized comparison. Also, consider what AMD was able to achieve on CPU performance by saving area on graphics. As in, the point I just made in the last post.

 

[lyonwonder]

I wonder if Intel will eventually migrate its big and little cores strategy with Alder Lake to workstation and server class Xeons and HEDT chips, or will Intel just limit little cores to mobile and consumer desktop, though I'm unsure if Xeons with little cores would benefit servers? In any case Intel's current server chips already have most of its die space committed to dozens of big cores.

 

[Asterox]

Very similar iGPU performace, or for AMD old high clocked Vega 8 iGPU is still quite enough.As we see, in general as expected iGPU(Iris XE 96 EU) gaming Tiger Lake eats more power.

If we just replace old Vega 8CU iGPU with RDNA2 8CU iGPU...........................................

Intel and GPU-s, well it is sad I hope no one expects too much from Intel discrete GPU-s.

 

[tomatosummit]

I wonder if Intel will eventually migrate its big and little cores strategy with Alder Lake to workstation and server class Xeons and HEDT chips, or will Intel just limit little cores to mobile and consumer desktop, though I'm unsure if Xeons with little cores would benefit servers? In any case Intel's current server chips already have most of its die space committed to dozens of big cores.

I don't think we'll see them for quite a while. With chiplets, io dies and active interposers coming I do think having some spare silicon space for a 4core low power cluster would be easy to fit into these cpus but then it's competing with space for igpu, fpga, l4 cache sort of things and to go too far in the other direction using the efficient cores for the majority of the compute you end up with knights landing which has soured the idea and will be where XE devices come in.
I like the idea of a small cluster somewhere on a big server cpu doing low priority tasks while the big cores are doing the actual serving but I don't design cpus.
Xeons on lga1700 are where you will see alderlake-s and that kind of budget server is probably the best place for intel's hybrid designs to go, these machines running overnight/all day doing tertiary tasks with occasional bursts required.

 

[Shivansps]

Very similar iGPU performace, or for AMD old high clocked Vega 8 iGPU is still quite enough.As we see, in general as expected iGPU(Iris XE 96 EU) gaming Tiger Lake eats more power.

If we just replace old Vega 8CU iGPU with RDNA2 8CU iGPU...........................................

Intel and GPU-s, well it is sad I hope no one expects too much from Intel discrete GPU-s.

Well if Nvidia sells GT1030, AND GT1030 DDR4, and AMD sells RX550 and RX550 64 bit, i dont see why Intel cant do the same. The level of performance you see on a high clocked Vega 8 is around the RX550 128 bit, and that is not that far away from a RX560. a 8CU RDNA2 iGPU with DDR4 would already be RX560 performance class.

 

[jpiniero]

I wonder if Intel will eventually migrate its big and little cores strategy with Alder Lake to workstation and server class Xeons and HEDT chips, or will Intel just limit little cores to mobile and consumer desktop, though I'm unsure if Xeons with little cores would benefit servers? In any case Intel's current server chips already have most of its die space committed to dozens of big cores.

Unlikely. Even the Alder Lake Xeon E will probably won't have any small cores enabled. It's a good place to dump dies with busted small cores and IGP.

 

[Mopetar]

APU performance in some ways is always going to be limited by memory bandwidth so there's not a lot of incentive to chase performance. Rather, being able to achieve that performance level with less die space or better yet power draw is going to be better for AMD/Intel and the customers.

Also, we should be glad the memory bandwidth holds them back, or crazed Etherium miners would be buying up all of the APUs in order to mine with them. Sadly, I'm only half joking here.

 

[reb0rn]

Forget DDR5, with this new damn chia crypto everyone will be looking into having +250gb ram in order to use a ramdisk for plotting since that process kills a 1TB NVME in matter of days. The high-capacity of DDR5 will make it impossible to get.

You can not plot on RAM there is almost zero benefit from using it vs 2x NVME disk also when you done with ploting you do not need NVME disk!

 

[mikk]

Very similar iGPU performace, or for AMD old high clocked Vega 8 iGPU is still quite enough.As we see, in general as expected iGPU(Iris XE 96 EU) gaming Tiger Lake eats more power.

If we just replace old Vega 8CU iGPU with RDNA2 8CU iGPU...........................................

Intel and GPU-s, well it is sad I hope no one expects too much from Intel discrete GPU-s.

This is a showcase model, similar story to 4800U. Relevant models are 4500U and 4700U - they are losing against i5-1135G7 and i7-1165G7. What jpiniero told there is nonsense. Only a small percentage of AMD devices can keep up. Lucienne is slowly coming and basically nothing from Cezanne-U.

 

[uzzi38]

This is a showcase model, similar story to 4800U. Relevant models are 4500U and 4700U - they are losing against i5-1135G7 and i7-1165G7. What jpiniero told there is nonsense. Only a small percentage of AMD devices can keep up. Lucienne is slowly coming and basically nothing from Cezanne-U.

Except he bought a model at retail. TechEpiphany doesn't get samples, he buys the laptops like any normal person would.

At this point you're just coping.

 

[Asterox]

Mindfactory CPU sales numbers for April 2021.

https://www.reddit.com/r/Amd/comments/n39hxn

Huh, Rocket Lake is seling like a flat tire at Tour de France or Giro d'Italia.



As expected, best seling R5 3600 it is still sold in large quantities.

Approximately a month ago, Mindfactory started seling R5 4650G Renoir APU.It is currently sold out or not available, but still it managed to sell more units vs two Rocket Lake CPU models.

https://www.mindfactory.de/product_...-4-3GHz-So-AM4-11MB-Cache--Tray-_1370228.html

 

[eek2121]

Huh, Rocket Lake is seling like a flat tire at Tour de France or Giro d'Italia.

This is a thread for discussing current and future Intel architectures, not a thread for constant hating on Intel, and definitely not a thread for talking about APU performance except as it relates to the topic. Please take your Anti-Intel banter elsewhere.

We get it. AMD has a better (current) architecture. We know this. I have all AND machines in my house, however I have an interest in both new tech from both companies, and the constant off topic banter is an annoyance.

EDIT: I want to be clear. that I think your post has merit, but not in it’s thread. Such a topic would be better in a new thread.

That being said, apparently the 11980HX will be available May 11th. Hopefully this means we will get an Anandtech preview pretty soon. I am extremely interested in this chip, in 65W mode it should perform similar to the 11700K while using significantly less power.

 

[DrMrLordX]

@Asterox

The 10900k and 11900k aren't even in the top 20 there. Whew.

@eek2121

Even if you don't like the presentation of data, you have to admit that it's telling which Intel CPUs are still selling and which aren't. My main concern with the 11980HX and all other 10nm chips would be supply. How many of these can Intel actually push onto the open market? Will they be available to the DiY crowd?

 

[coercitiv]

Even if you don't like the presentation of data, you have to admit that it's telling which Intel CPUs are still selling and which aren't.

Remember the last time people used Mindfactory data in this thread and the derail that followed? Even if I openly support taking into account this type of vendor data as long as it's being used to interpret a DYI trend, I also see how toxic it is for the thread when used as bait.

 

[Hulk]

I think the MindFactory data is interesting and does belong in this thread.

I didn't take it as a dig against Intel/Rocket Lake. It's the sales ramifications, albeit from one source, of Intel's recent engineering/design decisions with the "Lakes."

 

[LightningZ71]

Since this thread has gone meta:

Do we think that the sales information on a "lakes" processor, either directly, or in relation to its market competition, is relevant to a discussion about those processors? If so, what are the allowed sources of such sales information, especially since the only freely accessible data we have is from a source that seems to be disallowed here?

I argue that, since the primary goal of making these processors is to sell them for commercial purposes, the sales data on them is quite important, in addition to the technical discussion of those products. This thread isn't titled "Only technical discussion on Intel Lakes processors", so it isn't excluded by the thread title.

 

[Hulk]

Since this thread has gone meta:

Do we think that the sales information on a "lakes" processor, either directly, or in relation to its market competition, is relevant to a discussion about those processors? If so, what are the allowed sources of such sales information, especially since the only freely accessible data we have is from a source that seems to be disallowed here?

I argue that, since the primary goal of making these processors is to sell them for commercial purposes, the sales data on them is quite important, in addition to the technical discussion of those products. This thread isn't titled "Only technical discussion on Intel Lakes processors", so it isn't excluded by the thread title.

I concede those are valid points. I think more information is a good thing in general. It's up to the reader to vet that data for relevance and accuracy. I am fully aware that Dell alone will purchase magnitudes of orders more parts than the DIY market. But the "possible" lack of DIY sales in relation to the competition is also interesting because it provides additional insight into the part as it functions as a whole in the environment of computing.

Furthermore, at 418 pages I wouldn't exactly call this thread "tight and on topic." It has rambled about and is more a blog at this point so again, in this thread, under these circumstances I think the post is beneficial and warranted as it does directly relate to Rocket Lake.

Finally, the thread title doesn't specifically mention architecture, it's a pretty broad title, which can encompass sales data.

"Intel current and future Lakes & Rapids thread"

 

[repoman27]

Well, if we're going full-meta, the thread title was originally "Intel Cannonlake, Ice Lake, Tiger Lake & Sapphire rapid thread". The fact that nearly 4 years and over 10,000 posts later, Intel still hasn't finished launching the full product stack of even the second "lake" on that list and has changed their roadmap so many times that the thread was renamed, speaks volumes.

 

[coercitiv]

Do we think that the sales information on a "lakes" processor, either directly, or in relation to its market competition, is relevant to a discussion about those processors? If so, what are the allowed sources of such sales information, especially since the only freely accessible data we have is from a source that seems to be disallowed here?

My take is using MF data to discuss DYI trends is open game, using it as bait should be deterred. Comparing Rocket Lake to a flat tire does sound more like bait than sales analysis to me.

 

[NTMBK]

I concede those are valid points. I think more information is a good thing in general. It's up to the reader to vet that data for relevance and accuracy. I am fully aware that Dell alone will purchase magnitudes of orders more parts than the DIY market. But the "possible" lack of DIY sales in relation to the competition is also interesting because it provides additional insight into the part as it functions as a whole in the environment of computing.

Furthermore, at 418 pages I wouldn't exactly call this thread "tight and on topic." It has rambled about and is more a blog at this point so again, in this thread, under these circumstances I think the post is beneficial and warranted as it does directly relate to Rocket Lake.

Finally, the thread title doesn't specifically mention architecture, it's a pretty broad title, which can encompass sales data.

"Intel current and future Lakes & Rapids thread"

So what you're saying is that this thread is ridiculously broad and unfocused?

 

[Gideon]

IMO it is absurd to have a single thread for everything Intel while AMD has a thread to every family of products, every generation.

It made more sense to have 1 thread for everything skylake based, but IMO now, when we multiple quite different products we should move on from this thread.

 

[NTMBK]

IMO it is absurd to have a single thread for everything Intel while AMD has a thread to every family of products, every generation.

It made more sense to have 1 thread for everything skylake based, but IMO now, when we multiple quite different products we should move on from this thread.

See also the Atom thread, which covers "every single small core Intel product since Silvermont"...

 

[blckgrffn]

IMO it is absurd to have a single thread for everything Intel while AMD has a thread to every family of products, every generation.

It made more sense to have 1 thread for everything skylake based, but IMO now, when we multiple quite different products we should move on from this thread.

So who's making the Alder Lake Hype Train thread? I mean, they are always fun and when we are invariably disappointed in some aspect of the product (Intel, AMD, nvidia, whoever) there's always another train leaving the station.

 

[Hulk]

So what you're saying is that this thread is ridiculously broad and unfocused?

Ha! Not exactly. I'm saying this thread is broad in relations to the "Lakes" and somewhat unfocused. It meanders from focus-to-focus. If you been reading it you know as news pops up we'll discuss for a few pages, then move to new news, but as it relates to the "Lakes."

This thread is unique and fun.

It somehow achieved sentience, it's alive, and can't quite be controlled and that adds to the fun.

But, yes we should try to keep it more focused. Still, I think the possibility of poor DIY Rocket Lakes sales could likely be a result of of underwhelming performance compared to the competition, and a problem with pricing as well. All of these issues, ... performance, sales, and pricing are "Lakes" issues.