311: The Fab Floor

00:00SPEAKEROK, apparently I need to use this cold open to defend my honor.

00:04SPEAKEROh, no, you're listening to the other podcast, aren't you?

00:07SPEAKERFrom the vicious slander that apparently has been aired

00:11SPEAKERon a competing Linux podcast.

00:13SPEAKERWait, what did Adam say?

00:18SPEAKERI think it was more you.

00:19SPEAKERWhat did I do?

00:20SPEAKERThe refined is not a bootloader conversation came up.

00:23SPEAKEROh, yeah, it doesn't matter.

00:24SPEAKERAnd it does matter.

00:26SPEAKERNope, doesn't matter.

00:27SPEAKERHere we are to talk about why it matters.

00:29SPEAKERYou brought this on yourself.

00:30SPEAKERHere we go. Technically correct.

00:32SPEAKEROh, it's not. It is. Listen, OK.

00:34SPEAKERAll right. This is this is the diary's material potentially.

00:37SPEAKERSo, OK, I would keep it short.

00:39SPEAKERI will just say, though, if you are getting to the point where you are manually

00:42SPEAKERconfiguring your boot process, which you are clearly flirting with now.

00:45SPEAKEROh, no, you haven't listened to this week's episode yet.

00:48SPEAKEROh, OK. Did you abandon the.

00:50SPEAKERNo, I did it. It's done.

00:52SPEAKEROK, well, then that's the only point I would make is that it's important

00:55SPEAKERto know what each piece is doing and what is happening at each stage

00:58SPEAKERof the boot process.

00:59SPEAKERIf you are going to manually configure things, you don't know what.

01:02SPEAKERThe reason I say refined is not a bootloader is because you still need a bootloader.

01:06SPEAKERIf you have refined, you still need a bootloader downstream of refined.

01:09SPEAKERYeah, it's a bootloader loader.

01:10SPEAKERYeah, OK. He calls it a boot manager.

01:12SPEAKERI read I spent a lot of time on Rod's books.

01:14SPEAKERI learned about it all.

01:15SPEAKERDude, you've been on Rod's books.

01:17SPEAKERNow we can speak a common language.

01:19SPEAKERI you're going to you're probably going to yell at your radio again this week

01:22SPEAKERbecause I explained to Adam how a UEFI boots into the OS

01:27SPEAKERto explain a problem he's been having.

01:29SPEAKERAnd I know and I prefaced it by saying this is the gist, not an expert.

01:36SPEAKERUh huh. I know you're going to be yelling at your radio.

01:39SPEAKERThat's why I spent a lot of time on Rod's books and I have a fully functional

01:42SPEAKERrefined dot com with a manually configured bootload.

01:46SPEAKEROK, now we can now we can speak the same language.

01:48SPEAKERI don't know about that, but I still say bootloader loader.

01:51SPEAKERThat's fine. I think as long as you're differentiating.

01:54SPEAKERBut I mean, again, though, the real point there is if you're actually going

01:57SPEAKERto manually write a refined dot com as you are now doing.

02:00SPEAKERYeah, you still have to point refined at a bootloader.

02:03SPEAKERThat's true. Refined can't run an operating system kernel.

02:06SPEAKERIt has to boot a bootloader that then launches an operating system kernel,

02:09SPEAKERunless you may have noticed.

02:12SPEAKERI don't know if you've gotten this far yet,

02:13SPEAKERbecause this is how I'm doing my butter if a sub volume distro set up.

02:17SPEAKERRefined can manually start the Linux kernel itself.

02:20SPEAKERThat's what I'm doing.

02:21SPEAKERAnd you might be asking, well, where is the bootloader in that situation?

02:25SPEAKERIs it refined? It's the Linux kernel.

02:27SPEAKEROh, right. Because it can boot itself if it speaks.

02:30SPEAKERYeah, I think that's fucking crazy.

02:31SPEAKERAnd rad is it's called.

02:33SPEAKERI think they call it like an EFI stub loader.

02:35SPEAKERAnd it has to be it has to be compiled into the kernel.

02:38SPEAKERLike it's possible that you have a kernel that does not support that

02:40SPEAKERif they didn't compile it that way.

02:41SPEAKERBut yes, like the kernel, the kernel can be its own bootloader.

02:45SPEAKERAnd that ties into this whole move toward.

02:47SPEAKERI think they refer to it as unified kernel images. Yeah.

02:50SPEAKERNow, which is another thing that's coming about, I feel like alongside

02:54SPEAKERlike immutable distros as a way of saying like, hey, here is a whole

02:57SPEAKERand complete Linux OS that can't be fucked up, that I can't not ever boot.

03:02SPEAKERLike if we just start using the kernel as its own bootloader,

03:04SPEAKERthen it probably fixes some reliability problems with the boot process.

03:08SPEAKERIt solves some problems.

03:09SPEAKERNow, the real irony of this whole thing is that when I was setting up Linux

03:13SPEAKERthe first time, I chose refined because it was billed as the one

03:17SPEAKERthat would just auto configure itself.

03:18SPEAKERIt would look at all the BFI boot partitions and make you a list of them

03:22SPEAKERwith a graphical interface that looked halfway decent and scaled right

03:25SPEAKERon laptops with high resolution screens. Sure.

03:28SPEAKERAnd so I picked that one because it didn't require manual configuration.

03:33SPEAKERAnd then the thing that I ended up having to create the manual

03:37SPEAKERconfiguration for it, for creating snapper snapshots,

03:41SPEAKERan interface for booting into an older snapper snapshot.

03:45SPEAKERTurns out, if I just used grub, that would have been supported automatically

03:48SPEAKERbecause it's the default.

03:49SPEAKERHa ha. So, you know, I kind of was hoisted by my own Picard here.

03:53SPEAKERYeah. Yep. Yep.

03:54SPEAKERI know the feeling. I know the feeling.

04:18SPEAKERWelcome to Brad and Will Made a Tech Pod. I'm Will.

04:28SPEAKERI'm Brad. Yeah, Picard's.

04:31SPEAKERWe have a we have a banger of an episode for you here this week.

04:36SPEAKERYes. And Linux is not really mentioned even once, as far as I remember.

04:40SPEAKERSo you're in luck.

04:41SPEAKERYou mentioned it when we talked about thread directing.

04:43SPEAKEROh, well, barely. Yeah.

04:45SPEAKERBut a couple of weeks ago, I got to go down

04:48SPEAKERto Chandler, Arizona, to Intel's fab 53,

04:54SPEAKERwhere the new 18A process is being made and learn all about the first chip

04:58SPEAKERto roll out with that, which is called Panther Lake.

05:00SPEAKERAnd I went there with my friend and colleague from PC World, Adam Patrick

05:05SPEAKERMurray, to talk about to to learn about this stuff.

05:09SPEAKERSo Adam, we managed to get Adam in here for a return visit on the show.

05:11SPEAKERI think this is, I don't know, nth visit to the tech pod.

05:15SPEAKERWelcome to the show, Adam Patrick Murray.

05:17SPEAKERLet's talk about Intel's Panther Lake.

05:20SPEAKERI'm back doing weird voices today, I guess.

05:23SPEAKERGlad to get you guys on a podcast together.

05:25SPEAKERFinally. Yeah, Will, it's great meeting you, Brad.

05:28SPEAKERLong time no see.

05:29SPEAKERI heard you needed a Linux expert, so I'm glad you brought me on.

05:33SPEAKERGlad you workshop that one a few minutes ago and compiled any good journals lately.

05:38SPEAKEROh, yeah, I'm doing a left and right.

05:40SPEAKERThat's just what I do.

05:42SPEAKERIt's my thing.

05:42SPEAKERLook at Adam's favorite thing in the world is installing new Linux distros.

05:45SPEAKERIt turns out he just he just can't stop.

05:48SPEAKERI literally did it twice today already.

05:51SPEAKERGreat. What do you got on deck?

05:53SPEAKERI need to know who to go to.

05:55SPEAKERYeah. All right.

05:57SPEAKERAlso, we have an upcoming video on on PC World's YouTube channel

06:01SPEAKERabout the bee box, and I was reminded that that haiku OS is out there.

06:06SPEAKERAnd I was like, you know what? Maybe I should try that one as well.

06:08SPEAKERYeah, do you forget forget the Linux podcasts?

06:11SPEAKERWhen are we going to get the haiku podcast?

06:13SPEAKERThat's season two, man.

06:14SPEAKERSeason two of The Dual Boot Diaries, haiku and Linux.

06:17SPEAKERYou know, trust me, trust me.

06:18SPEAKERI had strong feelings about your conversation about, hey,

06:20SPEAKERwhich OS will we go to from here?

06:22SPEAKERYeah. Look, what was the new contender?

06:25SPEAKERSomebody sent in the one that's all virtualized.

06:28SPEAKERDo you remember? Oh, EXVM.

06:32SPEAKERI don't know. That's a that's a show about ESXI.

06:36SPEAKERYes, I was the VMware thing.

06:40SPEAKERNo, man, I can't remember what it was called.

06:42SPEAKERIt was the thing that somebody sent into the letters episode last week.

06:46SPEAKERAnd when I looked at it, it was like everything's running in a virtualized container.

06:50SPEAKEROh, oh, yeah.

06:51SPEAKERThe EXWM is the the emacs where it's like it's all just emacs.

06:56SPEAKERYeah, the whole desktop manager is emacs.

06:59SPEAKERThe whole operating system running emacs that, yeah, OK, that checks out.

07:02SPEAKERAnyway, but anyway, yeah, we're here we're here today to talk about Intel.

07:06SPEAKERAdam and I went Intel Intel brought us to Arizona

07:11SPEAKERa couple of weeks ago to see the their new fab

07:14SPEAKERand to find out what's coming up with their next gen mobile and data center processors.

07:18SPEAKERAnd like spoiler alert, we skipped the data center processor bit

07:21SPEAKERand went to the fab and learned about laptops and then,

07:25SPEAKERyou know, did other stuff the data center day because we're fine there.

07:29SPEAKERYeah, hands are like, I guess that's probably appropriate for your audiences.

07:33SPEAKERI mean, is that Xeon still is that is that or the dance?

07:37SPEAKERAll right. That's like Clearwater Lake or whatever.

07:39SPEAKERRight. Clear Forest is Clearwater Forest.

07:42SPEAKERIt's all it's all code name stuff right now.

07:43SPEAKERYeah. Clearwater Forest is data center.

07:45SPEAKERAnd it's like for the kind of CPUs that that it seemed like the kind of CPUs

07:50SPEAKERthat Nvidia did the deal to get access to in their inference machines

07:53SPEAKERthat they're using to train new models with.

07:56SPEAKERBut but like I said, I went to the first like 30 minutes of that session.

07:59SPEAKERI was like, we're never going to talk about this in a professional context.

08:01SPEAKERI went did something useful. That's not consumer stuff.

08:04SPEAKERBut yeah, you did see consumer stuff.

08:06SPEAKERWe did. We talk about that.

08:07SPEAKERCan we talk about the junket real fast before we get to the CPUs?

08:10SPEAKERYeah. Yeah. They don't call it a junket.

08:13SPEAKERAs somebody who has been on endless numbers of video game

08:16SPEAKERjunkets over the last 20 years, although not so much these days. Yeah.

08:20SPEAKERLike what is a CPU press trip like?

08:24SPEAKERIt probably smells the same.

08:26SPEAKEROK, the game. Yeah, I'm just kidding.

08:28SPEAKERNo, they're sure everyone has smelled great. Yeah.

08:32SPEAKERWell, this one, I would say specifically,

08:35SPEAKERa lot of the junkets around like technology deep dives like this are

08:41SPEAKERare around trade shows, right?

08:44SPEAKERSo CS, Computex, things like that. So.

08:48SPEAKERThe best case scenario is something like this where you can have time dedicated.

08:53SPEAKERYou don't want too much time.

08:54SPEAKERI feel like this one actually went on a little longer than needed to.

08:57SPEAKERBut for the most part, you want time to just sit there

09:01SPEAKERand absorb and get to chat with people, not only from the company

09:05SPEAKERthat you're visiting, but also other press and analysts around.

09:09SPEAKERSo this one, like when we're at trade shows,

09:12SPEAKERthere's a whole mixture of analysts, press influencers,

09:17SPEAKERthe whole the whole gamut.

09:19SPEAKERAt this one, there was a lot of analysts,

09:22SPEAKERa lot of press, but more traditional press, I will say.

09:25SPEAKERAnd then there were only a couple that I would consider influencers.

09:31SPEAKERSo it was it was a lot more.

09:34SPEAKERI mean, I was there, man.

09:35SPEAKERClassic. Yeah. Oh, yeah. I'm sorry.

09:36SPEAKERYeah. Will will the influencer.

09:38SPEAKERI was the biggest influencer there.

09:40SPEAKERPretty, pretty serious.

09:41SPEAKEROK. Did did they at any point try to put you in a helicopter or perhaps a dune buggy?

09:47SPEAKERNo, they put it. They let's go inside the fab.

09:49SPEAKEROK. All right.

09:50SPEAKERThen this is different than a lot of video game press strips.

09:54SPEAKERThe other big difference between most of the video game stuff I was exposed to

09:58SPEAKERat Maximum PC and later on is that they also brought the actual people

10:03SPEAKERthat do stuff here.

10:05SPEAKERSo like most of the video game stuff, you'd have the people

10:07SPEAKERwho were trained to talk about the video game, and that was maybe it.

10:10SPEAKERAnd there'd be people like that.

10:12SPEAKERYou'd be talking to somebody.

10:13SPEAKERYou'd be like, oh, yeah, I designed all the combat for the multi

10:15SPEAKERfor this multiplayer mode when they were doing the demos.

10:18SPEAKERBut they didn't like make those people readily available for you to talk to.

10:22SPEAKERAnd like at this, we sat down with with Tom Peterson Tap.

10:26SPEAKERTalked about GPUs.

10:27SPEAKERHe was just walking around and like hanging out.

10:29SPEAKERWe could sit down with him at dinner.

10:30SPEAKERStephen Robinson, who's in charge of X86, was there.

10:34SPEAKERAnd we could talk to him about X86 architecture stuff like the smart people.

10:40SPEAKERYeah. Tap like essentially spearheaded G-Sync.

10:44SPEAKERThe whole G-Sync initiative at Nvidia.

10:46SPEAKERThis the the X86 architect guy.

10:49SPEAKERI mean, he was a nice guy.

10:50SPEAKERHe just looked like a normal guy.

10:52SPEAKERYou know, he's like, oh, OK, cool.

10:54SPEAKERWe're just hanging out.

10:54SPEAKERI was like, what? What hot new instructions are coming down the pike?

10:57SPEAKERHe's like, we're not talking about forward looking products right here.

10:59SPEAKERBut we talked to the person who the

11:03SPEAKERboth Stephen and Tap and Rajshree Chabuskar,

11:08SPEAKERwho is the they're all Intel fellows, which is like

11:13SPEAKERI think when Tap was on the the full nerd a few weeks ago,

11:15SPEAKERhe described it as the person who's able to kind of cut through the

11:18SPEAKERthe the kind of the back and forth arguing that engineers like to do by saying,

11:23SPEAKERyou know, I think that this is probably the right way to go.

11:25SPEAKERSo they have a little bit more weight.

11:26SPEAKERBut like Rajshree is in charge of thread director. Right.

11:30SPEAKERThe thing that determines what goes on PCOR is what goes on e-cores

11:33SPEAKERand how they like how that communicates with the OS's.

11:36SPEAKERAnd so we were able to have much deeper conversations

11:40SPEAKERabout how things actually work with the people who really know

11:43SPEAKERinstead of being filtered through the PR and marketing team,

11:46SPEAKERwhich is like those people are great, but.

11:49SPEAKERThey they don't have the same level of knowledge as.

11:53SPEAKERHey, I'm in charge of X86 right now.

11:55SPEAKERYeah, I mean, especially with such an intensely technical subject.

11:58SPEAKERYeah, like I was I was over my head talking to Stephen just to be like, I

12:02SPEAKERwas like, I'm really excited to talk to you.

12:05SPEAKERI have no idea what to ask because like this is a level of how this works

12:08SPEAKERthat goes beyond what I've ever really cared about.

12:11SPEAKERI didn't say it that way, but like that's the reality.

12:14SPEAKERWell, also, like I would say,

12:16SPEAKERyou know, you're looking for like extracurricular activities.

12:20SPEAKERI think this one, because of the tours through the fab,

12:23SPEAKERlike they were going literally for three days straight, almost four days.

12:27SPEAKERWhat? Oh, four. Yeah.

12:29SPEAKERAnd like 10 hours a day or 12 hours. I can't remember exactly.

12:32SPEAKERSo like, I think they were so busy coordinating all that stuff

12:36SPEAKERthat they just didn't worry too much about the other stuff.

12:39SPEAKERThat's fine. I mean, I'm mostly joking.

12:41SPEAKERTo be clear, we usually skipped the extracurricular activities

12:44SPEAKERbecause we had to work. But yeah, well, I'll say the Ryzen 7000 launch,

12:49SPEAKERwhich was like the first junket that Gordon and I went to

12:52SPEAKERafter a gourd mong, the late gourd mong.

12:56SPEAKERWe went to it like kind of post pandemic.

12:58SPEAKERIt was like the first one of those.

13:00SPEAKERAnd it was at a resort in outside of Austin, Texas.

13:05SPEAKERAnd they did have like an ax throwing time that you can go through some axes,

13:09SPEAKERyou know, things like that sounds like the classic junket.

13:12SPEAKERSo those exist still.

13:14SPEAKERAnd like the meals and stuff,

13:17SPEAKERit was all just catered by the hotel and was just like it was nice hotel food,

13:22SPEAKERbut it wasn't anything super like it was hot, hot table buffet

13:26SPEAKERstyle everywhere we went and like cold sandwich boxes and stuff.

13:30SPEAKERLike it was the kind of stuff you get it

13:31SPEAKERlike it was better than when you get a convention center

13:34SPEAKERin like the meeting room or something like that, but not like not fancy.

13:39SPEAKERYeah, that sounds reasonable.

13:41SPEAKERThey did have hold on. They did. This is really important.

13:43SPEAKERThey did have miniature horses delivering beer one night.

13:46SPEAKERWhat? Yes. Oh, yeah.

13:49SPEAKERThis was during that is the evening dinner time. Yeah.

13:54SPEAKERYeah. So the minute the miniature horses were taco and Corona were their names

13:58SPEAKERand they had like coolers on either side of their back

14:01SPEAKERthat were kind of balanced against each other with like a little napkin

14:04SPEAKERholder and an umbrella and limes.

14:06SPEAKERAnd then they have tiny sombreros into servitude.

14:10SPEAKERI mean, look, what are who am I to say

14:13SPEAKERthat we shouldn't take control of the natural world, Brad?

14:16SPEAKERThat is more bacchanalian than anything I ever saw on a video game trip.

14:19SPEAKERI got to I petted Corona and I gave him a snack and he was we're best friends now.

14:25SPEAKERSo I'm over here and he's besmirching of taco or Corona.

14:28SPEAKERThe rate that same evening, they also had a couple of people set up

14:31SPEAKERdoing leather work so you could make your own leather suitcase badge.

14:36SPEAKEROK, all right. Yeah. So, yeah.

14:39SPEAKERAll right. Well, we're here to talk about CPU.

14:41SPEAKERSo, yeah, see, you got to use central processing units.

14:45SPEAKERWell, actually, there it's it's not just the CPU.

14:47SPEAKERThere's APS or I'm sorry, GPUs, NPs.

14:51SPEAKERAll of the P's were talked about.

14:53SPEAKERWell, the big thing is that like this is weird.

14:57SPEAKERThis is now like the kind of like the second full generation

15:01SPEAKERof Intel's disaggregated design.

15:03SPEAKERSo Lunar Lake and Arrow Lake and Meteor Lake,

15:06SPEAKERwhich are kind of different paths down the same trail,

15:10SPEAKERall have triplets in the same way that the Ryzans have had for a long time.

15:14SPEAKERAnd they were talking about what they learned.

15:17SPEAKERThe big thing about Panther Lake is it's the first thing on Intel's 18A process,

15:21SPEAKERwhich is their like next generation cutting edge fab process.

15:28SPEAKERLike intended to be a TSMC competitor, right?

15:30SPEAKERI mean, I'm sure that's what they would like to say.

15:32SPEAKERBut is that does that seem realistic?

15:35SPEAKERWell, yeah, I mean, they they they were very clear the whole time being like,

15:40SPEAKERwe believe this is the most advanced process in the world.

15:44SPEAKERAnd it's produced right here in the great US of A.

15:48SPEAKERInteresting. They were they were hammering that home.

15:51SPEAKERSo and not to mention the previous CEO

15:55SPEAKERwas kind of the the spearhead of this this whole effort.

15:59SPEAKERSo he doesn't get to see the fruits of of this labor.

16:02SPEAKERI mean, who knows how long 18A was actually in development?

16:05SPEAKERMaybe it was way beyond

16:07SPEAKERPat's time. But yeah. Yeah.

16:10SPEAKERSo, OK, so the big stuff on Panther Lake, it's like I said, it's a

16:14SPEAKERit's a heterogeneous there's multiple processes.

16:17SPEAKERThey're almost all Intel processes, except for a couple of the GPU core

16:22SPEAKERoptions are TSMC, all of the GPU.

16:27SPEAKEROh, I thought the four was an Intel one. OK.

16:33SPEAKERBut yeah, so there's basically the cores are the tiles on this core

16:36SPEAKERand they're calling these S.O.C.s still.

16:38SPEAKERBut it doesn't stand for system on a chip anymore.

16:42SPEAKERThat was its system of chips now is what they're calling.

16:45SPEAKERYeah, I'm sorry. What?

16:46SPEAKERYeah, it was a growing audience when that one went out. Yeah. Yeah.

16:52SPEAKERBut but the main.

16:53SPEAKEROK, so the big stuff that's changed, smaller transistor size,

16:57SPEAKERthe nanometer scale of the transistors seems to be getting kind of wobbly

17:01SPEAKERand in a way that you report the number because everybody's.

17:06SPEAKERLike it seems like maybe that's not going down anymore.

17:08SPEAKERNow we're doing new stuff to make the machines run more

17:12SPEAKERbetter at lower temperatures and voltages.

17:15SPEAKERThe big jump here is it seems like they're moving beyond

17:20SPEAKERFinFET gates, which was the thing that rolled out about 10 years ago.

17:24SPEAKERAnd that was so if you think about the way a transistor gate works,

17:28SPEAKERthere's a conductor that goes through something and then there's a

17:31SPEAKERgate around it that prevents that signal for me.

17:34SPEAKERIt either stops the signal from going through or lets the signal go through.

17:37SPEAKERAnd FinFET is called a FinFET because the conductor that goes through

17:41SPEAKERwas a vertical like vertical poking out of the semiconductor base. Right.

17:46SPEAKERSo it generally is pretty good.

17:49SPEAKERBut if you put too much voltage through, there'd be leakage

17:51SPEAKERaround the edges of the gate,

17:52SPEAKERaround the edges of the conductor that don't go through the gate. Right.

17:56SPEAKERWhat they're doing now is they're lifting that that conductor out of the stratum

18:01SPEAKERand making the gate a three dimensional loop around the conductor.

18:06SPEAKERAnd at that point, it's that's why they call it ribbon fed because it goes

18:09SPEAKERthe conductor is completely encircled by gate now.

18:13SPEAKERSo they're there.

18:14SPEAKERWhat they're saying is there's a lot less leakage or maybe no leakage.

18:17SPEAKERThat part was kind of unclear.

18:19SPEAKERThey didn't give they were there as a general rule.

18:22SPEAKERWe should we should start this out.

18:24SPEAKERThey wouldn't talk about yields at all,

18:25SPEAKERno matter how we tried to trick them into asking answering those questions.

18:29SPEAKERWell, it's still early, right?

18:31SPEAKERI mean, they announced full production. Yeah, they announced full production.

18:35SPEAKEROh, they are. OK, it's not early.

18:38SPEAKERWell, the rubber has met the road then the day that we were there,

18:42SPEAKERlike they maybe four days before we were there,

18:45SPEAKERthey announced that the fab was fully operational on October 9th.

18:48SPEAKERSo it's like like we were not.

18:52SPEAKERIt is still early in production.

18:53SPEAKERAnd as they said, they're still ramping production.

18:56SPEAKERLike we when we were on the fab tour, we'll talk about this later.

18:59SPEAKERThere were places where there were holes, where it was very clear

19:01SPEAKERthat there were going to be some really expensive machines

19:03SPEAKERthat just haven't been delivered or installed yet. I see.

19:07SPEAKERAnd I can I I want to digress too much here.

19:10SPEAKERThis is like the most basic like transistor one on one stuff.

19:12SPEAKERBut do you know what the mechanism of action is

19:15SPEAKERthat causes the gate to interrupt signal through the transistor?

19:20SPEAKERI mean, it's not obviously it's not mechanical at a nanometer scale.

19:24SPEAKERLike what do you know?

19:25SPEAKERDo you know what what is what is what is the operation there

19:27SPEAKERthat is interrupting the electrical current going through or not?

19:31SPEAKERYeah. When when basically there's signals

19:35SPEAKERthat go into the transistor base that apply into the gate,

19:39SPEAKERinto the base of the gate that make it more or less

19:43SPEAKERelectrically conductive based on how much electricity goes into it. Got it.

19:47SPEAKERSo if there's current at the base that it's open and if there's not,

19:50SPEAKERthen it's closed. OK.

19:52SPEAKERAnd that's the extent of my knowledge on how this works.

19:55SPEAKEROK, good enough.

19:57SPEAKERPlease don't ask follow up questions.

19:59SPEAKERNext question. Yeah. So.

20:04SPEAKERThe next thing is the power of is

20:07SPEAKERso the way that the power comes into the way

20:10SPEAKERfor the things that turn those gates open and closed.

20:13SPEAKERAnd those are kind of buses, is my understanding, kind of like a

20:17SPEAKERin the in the sense of like a system bus like that, like the power comes in on lanes

20:21SPEAKERand then they can open or closed based on state of the chip.

20:24SPEAKERWhat's being used to not is that those have come in on the front side

20:29SPEAKERfor all eternity.

20:31SPEAKERAnd now they're making those poke through the chip

20:34SPEAKERand coming in from the backside now, which will improve cooling

20:38SPEAKERand reduce power loss reduction and basically allow them for scaling

20:43SPEAKERto scale more power into the chip, which gives them better performance.

20:48SPEAKERAnd they specifically like they wouldn't talk about performance.

20:51SPEAKERThey wouldn't talk about yields at all, really, other than to give

20:54SPEAKERreally general numbers like it's 15 percent performance

20:58SPEAKERper watt improvement between 18 a and Intel three, which means nothing. Right.

21:04SPEAKERIn what? Yeah.

21:05SPEAKERBut performance per watt, Adam, performance per watt.

21:10SPEAKERAnd then, like many of the talks, this this whole talk about

21:13SPEAKERwhat fundamentally has changed at the low level with this process was

21:17SPEAKERthis is the most advanced semiconductor facility in the world.

21:20SPEAKERLike they said that three or four times throughout the throughout

21:23SPEAKERthese presentations.

21:24SPEAKERI mean, do you think, OK, this is a very big question

21:27SPEAKERthat they certainly would not have addressed?

21:28SPEAKERI mean, are we going to see iPhones?

21:30SPEAKERYou know, we're going to see like the A20 or whatever the next.

21:33SPEAKERI mean, maybe not the next one, but like three iPhones from now.

21:36SPEAKERIs Apple going to be fabbing SOCs in this place?

21:39SPEAKERAdam's making a face.

21:41SPEAKERYeah. I mean, I'm sure Intel would love

21:44SPEAKERto expand their partnership to work with Apple.

21:48SPEAKERSo I think that's going to happen.

21:51SPEAKERYou know, I mean, we will see.

21:53SPEAKERThey showed an arm chip that they built on 18 a

21:58SPEAKERthe my understanding is that the big

22:01SPEAKERthe big thing keeping Intel,

22:05SPEAKERIntel fabs, Intel foundry fabs accessible only to Intel as a customer

22:10SPEAKERis that the rest of the world uses a certain kind of tool chain

22:14SPEAKERand process to design chips and Intel uses their own thing.

22:18SPEAKERThis is I talked to Dr.

22:21SPEAKERIan Cutrus, who is tech tech potato on YouTube about this a few weeks ago,

22:25SPEAKERa few months ago now, I guess, or last year even.

22:28SPEAKERAnd his his answer is that

22:32SPEAKERIntel does something, does this one way.

22:34SPEAKEREverybody else does it the other way.

22:36SPEAKERPart of the reason that Intel has been building chips on TSMC

22:40SPEAKERbeyond the fact that they didn't have competitive processes for those chips

22:44SPEAKERis that they needed to build that internal knowledge

22:48SPEAKERso that these U.S. fabs can be competitive with the TSMC fabs in Taiwan.

22:54SPEAKERAnd whether they're successful at that, we're not going to know for a while.

22:59SPEAKERSo, yeah, I don't know part.

23:01SPEAKERThe other part of it, though, that's interesting is because of the way

23:03SPEAKERthe packaging works on these chips with the foveros and all that stuff.

23:07SPEAKERLike there's nothing to say that Nvidia can't make a chiplet

23:10SPEAKERat TSMC in Taiwan, ship it to Chandler to be assembled

23:14SPEAKERand put that on with Intel 18 a CPU dies and other stuff, right?

23:21SPEAKERI mean, yeah, after after that announcement of the partnership with Intel

23:24SPEAKERand Nvidia working together in some capacity, which it's still too early.

23:29SPEAKERThey keep telling us this is just the announcement, the intent to work together.

23:33SPEAKERWe probably won't see anything for years.

23:36SPEAKERYeah, I think it's the packaging part that Nvidia is very keen on,

23:41SPEAKERthat this foveros technology is is probably

23:45SPEAKERwhere Intel is going to be able to to hang their hat on.

23:48SPEAKERShould we talk about that, like what I've heard the term foveros

23:51SPEAKERa million times, and I still don't quite know what it is.

23:54SPEAKERSo foveros is the is the kind of glue that holds all these chips together.

23:58SPEAKEROK, it's it's it's a it's a general interconnect material

24:05SPEAKERthat has varying bump pictures for the connection points for the chiplets.

24:11SPEAKERAnd they talked about, like, for example, the data center

24:14SPEAKERparts have a different density than the than the consumer parts,

24:16SPEAKERjust because they want way more chip to chip chiplet to chiplet interconnection

24:21SPEAKERand communication than you get, than you need on the on the

24:24SPEAKERon the mobile parts where there's only basically one compute tile.

24:29SPEAKERSo foveros S, I think

24:33SPEAKERis the 36 micron bump pitch

24:37SPEAKERthat's on the current chipping parts.

24:41SPEAKERThe that's going to go to 10 microns now.

24:45SPEAKERSo seven hundred and seventy two to sixteen hundred bumps per millimeter

24:49SPEAKERmillimeter squared to 10 K bumps per millimeter squared.

24:55SPEAKERAnd then Clearwater Forest is going to be nine microns on that pitch.

25:00SPEAKERAnd what that basically means is how many connections each chiplet can have.

25:03SPEAKERSo that's the gate of that's the gate for chiplet to chiplet communication.

25:09SPEAKERThey also did a bunch of other stuff that we should talk about.

25:10SPEAKERLike we should talk about

25:13SPEAKERthe the S.O.C. chip, the one that has all the IO

25:16SPEAKERand all the other stuff on it that makes the CPU work.

25:18SPEAKERThat isn't an 18A chip.

25:20SPEAKERSo there's the 18A chip.

25:22SPEAKERThere's the S.O.C. chip, which is like all the chip housekeeping

25:26SPEAKERmemory controllers and that kind of stuff, as well as sometimes some e-cores.

25:29SPEAKERAnd then there's compute die and GPU die.

25:34SPEAKERBefore we move on, like interconnect is actually like one of the main things

25:38SPEAKERI wanted to talk to you guys about, because like that's been one of the big

25:40SPEAKERweaknesses on the horizon for so long since since they went to Chiplets.

25:44SPEAKERRight. Like it's the relatively slow infinity fabric between Chiplets

25:48SPEAKERis kind of the big problem when you get it's the whole reason we are still living

25:51SPEAKERwith core parking, right?

25:53SPEAKERBecause when you get a process running on one die, trying to hit the cache

25:56SPEAKERon the other die or whatever it is like talking between those two chips,

26:00SPEAKERChiplets is relatively slow enough to cause hitching and games

26:04SPEAKERand other issues like that.

26:05SPEAKERLike, did you get any sense if if like the Foveros tech

26:09SPEAKERwill alleviate that issue or won't face that issue to the same degree?

26:13SPEAKERI mean, were they talking in such real world terms about things like that?

26:17SPEAKERI asked about that when I was talking to Stephen.

26:20SPEAKERDid you ask at all?

26:22SPEAKERNo, I didn't talk too much about that.

26:25SPEAKERSo the answer, they wouldn't give a clear answer on that

26:28SPEAKERbecause they're not doing consumer multi chip

26:33SPEAKERmulti compute chip products at this point.

26:35SPEAKERNone of those were on the announce like everything that's being announced

26:39SPEAKERlike there were like I said, there was an eight core, a 16 core

26:42SPEAKERand a 16 core with a big GPU model all talked about.

26:46SPEAKEROK, now all of those are one compute tile.

26:51SPEAKERRight. Like Arrow Lake is all single compute tile.

26:54SPEAKERI think that the rumors I mean, this is this is getting further down the road,

26:57SPEAKERlike the rumors swirling about Nova Lake

26:59SPEAKERor that they are going to start shipping to compute tile designs.

27:02SPEAKERBut that's I mean, that's 18 years away now.

27:06SPEAKERYeah, it's actually not fully accurate because the the main tile

27:09SPEAKERthat has the IO and stuff like that has low power e-cores

27:13SPEAKERwhere that you can expand the bigger 16 core option actually adds

27:18SPEAKERa whole nother section with with other processor cores, e-cores on it.

27:23SPEAKERSo I'm getting out over my skis here now.

27:26SPEAKERI've read that like those LPE cores are not on the ring bus,

27:29SPEAKERbut I barely know what the ring bus actually is.

27:31SPEAKERI don't know what the effects of that are.

27:34SPEAKERI don't I don't remember the actual layout.

27:37SPEAKERI'd have to bring up the slides. I do. OK.

27:40SPEAKERSo the the layout is that the four p cores share an L2.

27:45SPEAKERThe four p cores are all connected via L3 cache.

27:49SPEAKEREverything's everything on the computer is connected via L3 cache.

27:53SPEAKERThe e-cores are connected via shared L2 in clumps of four.

27:58SPEAKERSo if you have four e-cores and four p-cores on that compute tile,

28:05SPEAKERthen the e-cores have L2 to L2

28:09SPEAKERcan talk across the L2 cache that's shared between them.

28:13SPEAKERThe p-cores can talk across the L3 cache shared between them.

28:16SPEAKERAnd if you need to dump something out to those LP, those low power e-cores

28:21SPEAKERthat are on the on the on the on the on the main IO die,

28:26SPEAKERthen that's a really expensive process.

28:29SPEAKERAnd we talked about I talked about this with Rajshree a lot,

28:32SPEAKERbecause it actually comes up in benchmarking like the thread

28:34SPEAKERdirector performance and the way

28:37SPEAKERthe way the system and the OS know whether it's worth computationally speaking,

28:42SPEAKERmoving a thread from a p-core to an e-core and e-core to a p-core

28:46SPEAKERis is very dependent on like how long it takes to transfer the data.

28:50SPEAKERLike how many clock cycles it takes to transfer the data.

28:53SPEAKERAnd she she was talking specifically about like twenty thousand clock cycles

28:57SPEAKERas an expensive as an expensive switchover,

29:00SPEAKERwhich is like one two hundred thousandth of a second.

29:03SPEAKERJust to be clear, right?

29:04SPEAKERWe're we're talking about infinitesimal slices of time.

29:09SPEAKERBut and I'm not going to get into that whole thread director thing.

29:12SPEAKERYou should like the conversation I have with her.

29:14SPEAKERI think that video is on the full nerd channel, right, Adam?

29:16SPEAKERThe full nerd network YouTube channel.

29:18SPEAKERYeah, is was fast.

29:20SPEAKERLike it was I learned so much about how that stuff works

29:23SPEAKERand why it's hard. I knew it was one of the big computer science problems

29:27SPEAKERgoing into that conversation.

29:28SPEAKERAnd I came out with a much better understanding of why it's difficult

29:32SPEAKERand a better understanding of kind of where we are with that stuff,

29:35SPEAKERbecause it's it really plagues all of the modern architectures.

29:39SPEAKERApple, Intel, AMD, everybody is having the same problems.

29:43SPEAKERWell, an AMD is is working on this as well.

29:47SPEAKERLike you reference the the chip design that has made them kind of

29:52SPEAKERmade their mark on the industry and changed a lot of things with Strix Halo.

29:57SPEAKERI was watching a video by a gentleman

30:00SPEAKERwe actually met while we were there that his YouTube is called High Yield.

30:04SPEAKERThe he essentially does a real deep dive architecture stuff.

30:07SPEAKERI've learned a lot from him. I'm no expert.

30:09SPEAKERIt was it was great to chat with him, but he actually did a really good

30:14SPEAKERdeep dive of the of like the I guess topology might be the right word

30:19SPEAKERof Strix Halo and how AMD seems to be using that

30:24SPEAKERto experiment with a new type of interconnect that fuses the huge GPU

30:30SPEAKERthat they have on on that processor with the the rest of the processing.

30:35SPEAKERAnd like, at least according to him, this is like super advanced stuff.

30:39SPEAKERSo that's why I think

30:42SPEAKERStrix Halo is important.

30:44SPEAKERAnd then this this Fovaro stuff is definitely important on the Intel side.

30:47SPEAKERSo one of the things they talked about, like TAP in his talk,

30:51SPEAKERwhen he's talking about GPU stuff, talked about using,

30:54SPEAKERespecially for like games that are built for consoles,

30:57SPEAKERusing the 8E cores probably makes more sense than using the 4P cores

31:01SPEAKERbecause the 8E cores have that shared L2 cache.

31:05SPEAKERAnd also they're they're comparable in terms of like instruction sets

31:08SPEAKERand functionality to the era of current generation CPU cores.

31:15SPEAKERSo you don't get any like it's not like there's AVX2 instructions

31:18SPEAKERor something like that in the in these games that are going to take

31:21SPEAKERthat the P cores are going to take advantage of, right?

31:23SPEAKERBecause the current current consoles have Zen 2 cores in them.

31:26SPEAKERSo it's not like you need a lot of modern desktop cores to equal that.

31:30SPEAKERThat's exactly it, actually.

31:31SPEAKERYeah. When I talked to I had a dinner

31:35SPEAKERset up between AMD and Sony about Project Amethyst.

31:39SPEAKERMark Cerny was there.

31:41SPEAKERThat was the first time I got to chat with him.

31:43SPEAKERAnd he was saying at least, you know, I mean,

31:46SPEAKERthey weren't talking about forward looking features,

31:48SPEAKERbut he was talking about how even on the PlayStation 5 Pro,

31:51SPEAKERthe Zen 2 based CPU cores in there were in there.

31:54SPEAKERHe didn't feel like that was a limiting factor,

31:56SPEAKERthat it was more other things that that they're working on.

32:01SPEAKERAnd yeah, the same.

32:03SPEAKERAnd the other thing to think about is that these are these are very

32:07SPEAKERcompact APUs, like

32:10SPEAKERall the all the wattage going into that needs to be shared.

32:13SPEAKERAll the heat dissipation is is makes a huge difference compared to desktop.

32:18SPEAKERRight. These these are packages that are going to be put in laptops

32:23SPEAKERthan light devices, handheld gaming PCs.

32:25SPEAKERSo the more the more perform or the the the more juice

32:30SPEAKERyou put you put into high performance cores might take up some of that

32:36SPEAKERoverhead when you'd rather be sending all that juice to the GPU. So.

32:40SPEAKERAnd and the big like the big thing that's happening with this

32:45SPEAKERthis generation, like last generation, we saw Meteor Lake

32:49SPEAKERand Lunar Lake and Lunar Lake was kind of the mainstream part.

32:53SPEAKERAnd Meteor Lake was like, sorry, rather

32:58SPEAKERwas in the like desktop replacement kind of workstation mobile parts. Right.

33:03SPEAKERWith this, Panther Lake is going to take over all of that.

33:06SPEAKERAnd both the mainstream.

33:09SPEAKERWell, not necessarily, because they in all their slides,

33:13SPEAKERwhich once again, a lot of them are vague for a lot of reasons.

33:17SPEAKERThey never mentioned Air Lake HX competitors.

33:20SPEAKERSo I was just graphics are like the highest end,

33:24SPEAKERlike real thick, chunky, eight pound gaming laptops.

33:28SPEAKERUsually it's you know, you got a low power

33:30SPEAKERCPU, kind of a mid range and then a really high power one.

33:34SPEAKERThey seem to indicate that the HX this would

33:38SPEAKERlike the top end configuration of the core count would would replace that.

33:43SPEAKERBut they never set it on any of the slides.

33:45SPEAKERAll they were doing was comparing it versus Lunar Lake and Air Lake H.

33:50SPEAKERSo I don't know if that means anything, but there might be another spot

33:53SPEAKERin the stack.

33:55SPEAKERI still believe that that Air Lake on the desktop is going to get a refresh.

34:00SPEAKERAnd I think maybe the highest end laptops will also get an Air Lake refresh.

34:04SPEAKERBut yeah, we don't know.

34:05SPEAKERThose are, you know, there's other benchmarks for the I7 270K just pop

34:10SPEAKERplus just popped up or whatever they're calling it.

34:12SPEAKERBut it's not like Panther Lake is explicitly mobile only. Right.

34:15SPEAKERThere is no desktop component to this next generation that we're talking about.

34:19SPEAKERRight.

34:20SPEAKERYeah, unless you fuse it to a mini ITX motherboard or something.

34:23SPEAKERI mean, we've seen some of this stuff, but yeah, this is this is meant for

34:27SPEAKERfor mobile computing of various sources.

34:30SPEAKEROK, got it. So that's kind of the high points.

34:33SPEAKERThey talked about the GPU stuff a little bit.

34:35SPEAKERIt's still like it's Intel, they're calling it XE3,

34:38SPEAKERbut it's basically the same core as the as the B series desktop GPUs.

34:45SPEAKERWhat are you saying?

34:45SPEAKERThat's what will be in this the next CPUs like in Panther Lake?

34:48SPEAKERThat's what's in Panther Lake.

34:50SPEAKERThat's the same technology.

34:51SPEAKERIt's not the same architecture, maybe a smaller,

34:56SPEAKERsmaller total grunt or whatever.

34:58SPEAKERBut and so that's what's in like the current arc cards is what you're saying.

35:02SPEAKERThat's in the arc B series cards. Yeah, yeah.

35:04SPEAKERSo that battle mage, you know, like B5AD.

35:07SPEAKEROK, well, that seems respectable for an integrated GPU.

35:11SPEAKERYeah, they've done stuff like added multi frame gen.

35:13SPEAKERThey're not doing the kind of last second mouse

35:17SPEAKERwarp mouse movement warping that video is doing with their five series cards.

35:22SPEAKERBut like the performance, it seems fine.

35:25SPEAKERLike we got to I got to play the new painkiller game

35:28SPEAKERon one of their dev boards, one of their dev machines.

35:32SPEAKERBut it was a pretty rough demo just just because I was playing

35:36SPEAKERwith a mouse and keyboard on a 40 inch TV about eight inches from my face.

35:40SPEAKERSo it was it was it was non nonoptimal.

35:44SPEAKERThe what I will say is that the multi frame gen on that

35:48SPEAKERon that demo, like even doing something simple, like trying to land the mouse

35:51SPEAKERon the corner of a pillar or something like that was really, really challenging.

35:54SPEAKERSo it was a little floatier than I would want for a mouse and keyboard.

35:57SPEAKERBut it seemed fine on the gamepad.

35:59SPEAKERWait, to be clear, you were playing on an integrated GPU on a panther like machine,

36:02SPEAKERnot a discrete.

36:03SPEAKERYeah, I mean, it was I don't they didn't have the machine out.

36:07SPEAKERThey had a lot of Panther Lake dev units, which were like

36:11SPEAKERI had sorry, Knuck, bigger than Knuck size

36:16SPEAKERopen boxes with small coolers.

36:19SPEAKEROh, these 40 watt coolers on them are these are these like prototype,

36:23SPEAKERlike not for production, like, yeah, kind of similar to when we would see consoles

36:28SPEAKERbefore they came out and it was not the console.

36:30SPEAKERIt was some giant cobbled together.

36:32SPEAKERThey had a mixture like they did have like

36:35SPEAKERthey had one of the same things that I actually saw at Computex,

36:38SPEAKERwhich was kind of like a huge dev box board, like not pretty at all.

36:43SPEAKERYes, I love seeing utilitarian.

36:45SPEAKERYeah, I think hardware like that.

36:47SPEAKERThey did show off the first, yeah, Knuck style development kit,

36:51SPEAKERwhich is a little more polished, but still meant, you know, for developer usage.

36:56SPEAKERIt was not I didn't see one that was powered on.

36:59SPEAKERThe only stuff that I saw running that was powered on were reference laptops,

37:04SPEAKERwhich are usually taken from an OEM and just kind of configured

37:07SPEAKERwith a different platform.

37:09SPEAKERSo I I don't know.

37:11SPEAKERDo you remember seeing any of the the.

37:13SPEAKERSo the painkiller one wasn't out on the shelf.

37:15SPEAKERIt was running on something that was in a cabinet.

37:17SPEAKERSo I don't know what it was.

37:19SPEAKERThe other ones were almost all I think that they had one of those dev kit

37:22SPEAKERboxes running something in the middle, but I don't remember for sure.

37:27SPEAKERAnyway, it doesn't matter like the demos were like they had some premiere demos.

37:30SPEAKERThey had a bunch of A.I.

37:31SPEAKERcrap that was kind of iffy.

37:33SPEAKERI just I also think like it is.

37:36SPEAKERI mean, it's hard to set up an event like this, right?

37:38SPEAKERIt takes it takes a lot.

37:39SPEAKERYou know, my hat's off to people who have to organize this kind of stuff

37:42SPEAKERbecause I'm sure it's a lot.

37:44SPEAKERBut unfortunately, and you kind of pointed out a little bit

37:47SPEAKERsometimes like the environment that you try something out in

37:51SPEAKERjust like kind of like sets the tone, if that makes sense.

37:55SPEAKERLike I listen to Digital Foundry a lot.

37:57SPEAKERAnd very often they'll say they'll go to something like

38:01SPEAKERrecently they went to what was before you for Gamescom.

38:05SPEAKERAnd they took some demos at Gamescom.

38:07SPEAKERBut like the TV had like motion smoothing on.

38:11SPEAKERAnd it was just not the HDR wasn't calibrated right.

38:14SPEAKERAnd, you know, like they're like, guys, I can't

38:17SPEAKERI can't really pay attention to what you're showing here

38:19SPEAKERwhen the when the environment that I'm supposed to see it on

38:22SPEAKERis not doing it justice.

38:23SPEAKERSo I think part of it was that that for will.

38:26SPEAKERBut I'm sure you deal with that all the time with with games, right?

38:29SPEAKERWhen you go to see something.

38:31SPEAKERI think that was like the infamous games beat Cuphead footage thing, right?

38:35SPEAKERWas that like wasn't he wasn't he like kind of Dean Takahashi?

38:39SPEAKERI think didn't he say he was like

38:40SPEAKERlike talking to a PR person and kind of half paying attention to the game

38:44SPEAKERwhen they were recording that? Yeah.

38:45SPEAKERI mean, the error there was then going and recording posting that footage.

38:49SPEAKERLike, I mean, is it like that's where you maybe practice

38:53SPEAKERsome editorial discretion and say, oh, this footage is maybe not usable,

38:56SPEAKERbut but it absolutely happens all the time that like, hey, this

38:59SPEAKERthis environment that I'm checking this out in is compromised in one

39:02SPEAKERto three different ways.

39:04SPEAKERAnd I've got to kind of

39:06SPEAKERintegrate that into my impressions.

39:08SPEAKERThat that's why it's hard because it's just like like not that I'm

39:12SPEAKERand I'm not saying that your experience was not valid.

39:14SPEAKERWell, it's just more like like that is not how not representative.

39:19SPEAKERSo like when when it's like, oh, yeah, we got to try something,

39:22SPEAKERbut it was for all intents and purposes, not worth anything

39:27SPEAKERbecause it was, you know, just a bad experience.

39:30SPEAKERIt was prerelease hardware, prerelease software on a game

39:33SPEAKERthat just barely got announced all that kind of stuff.

39:35SPEAKERAnd it's just like, oh.

39:36SPEAKERWhen you talk about organizing these events and how herculean

39:39SPEAKERthat task must be like, I think about they probably had to commit

39:42SPEAKERto the footprint of that demo station way before they knew

39:46SPEAKERif that demo would actually be ready to show or not.

39:48SPEAKERRight. Like they had to

39:49SPEAKERthey had to build out a list of here's what we're going to show with this thing.

39:52SPEAKERProbably weeks in advance of saying, hey, this demo is actually functional.

39:56SPEAKERI'm a hundred percent sure.

39:58SPEAKERLike like like the actual demos, I'm sure are all coming in as hot as possible.

40:02SPEAKERAnd that's why like I'm always so like I like going to these things

40:06SPEAKERand learning about the technology and, you know, growing my

40:10SPEAKERmy knowledge set, because I'm still pretty, pretty fairly early in my journey here.

40:15SPEAKERBut like really the fun part is like, OK, when when can I get it

40:20SPEAKERand try it out for myself, have my own experiences in a in in,

40:24SPEAKERyou know, a controlled environment?

40:26SPEAKERYou know, we always say wait for independent testing because it isn't.

40:30SPEAKERIt is important in a lot of ways, but that's what I'm always excited.

40:33SPEAKERIt's like, OK, cool.

40:34SPEAKERThank you for showing me all this stuff.

40:35SPEAKERBut man, when can I get my hands on it?

40:37SPEAKERBecause, yeah, I just I want to I want to see it.

40:40SPEAKERI want to feel it.

40:41SPEAKERI want to see how how it works.

40:44SPEAKERSo, OK, so I did.

40:46SPEAKERI do have an update on something that I'd forgotten.

40:48SPEAKERWe asked during the architecture roundtable when Stephen was talking

40:52SPEAKERand we asked what the bandwidth between the CPU and GPU tile was like,

40:58SPEAKERbecause they're both connected via that Phovros backhaul.

41:01SPEAKERAnd he specifically said it can be as high as they want.

41:04SPEAKERLike there's it depends on how many of the the connections you choose to use

41:10SPEAKERup to like like you can do

41:13SPEAKERyou can bump up the frequency and you can do up to like one hundred

41:15SPEAKERand fifty gigabits per second.

41:18SPEAKERBut the thing he said is once you do that, then you start shifting bottlenecks

41:21SPEAKERaround and there's definitely sweet spots at different places that require

41:27SPEAKERthat are difficult to predict because of the way you're moving the bottlenecks.

41:31SPEAKERThe takeaway I had from this is that when they're working inside these SOCs,

41:35SPEAKERespecially when they're talking about doing bigger GPUs in some places

41:39SPEAKERthan others, is that you can get kind of the same benefits

41:43SPEAKERthat we see by using APUs on consoles where they share different parts of the

41:48SPEAKERof the of the they share some high speed connections

41:52SPEAKERand then some low speed connections out to memory and stuff like that.

41:54SPEAKERSo you can end up in these weird hybrid architectures that are

41:58SPEAKERsimilar to desktop, but not exactly analogous

42:02SPEAKERand have some real strengths over the change

42:06SPEAKERwhere the bottlenecks are on these on these new processors.

42:09SPEAKERAnd I thought I thought that was kind of interesting.

42:11SPEAKERBut also they're not at a point where like the GPU can dig into L3 cache

42:15SPEAKERon the compute die, although it can do like some he said snooping,

42:20SPEAKERwhich I don't know if that's a technical term or not. But anyway.

42:24SPEAKERSpeaking of gaming.

42:26SPEAKERYeah, I was surprised by how much they talked about handheld gaming.

42:31SPEAKERThat's obviously one of the new areas.

42:33SPEAKERAnd I know, Brad, you have not really messed around with it too much.

42:37SPEAKERBut then also talked a lot about AI,

42:40SPEAKERyou know, multi frame gen and all that kind of stuff.

42:43SPEAKERWhen TAP was on the full nerd,

42:44SPEAKERwe had a very long discussion about where he thought AI was going.

42:48SPEAKERBut then also ray tracing, like the ray tracing performance in this GPU

42:52SPEAKERis one of the places where they wanted to really boost performance

42:55SPEAKERbecause it is such a big thing.

42:57SPEAKERLike, man, there's been so much talk

43:00SPEAKERrecently about the next gen consoles.

43:03SPEAKERI know you guys have talked about it over on next lander, stuff like that.

43:07SPEAKERIt's it's always interesting to hear

43:11SPEAKERyou guys not or or maybe I don't know how to phrase this.

43:14SPEAKERYou guys talk way less about ray tracing over a next lander

43:18SPEAKERand like that kind of stuff than what a lot of these junkets

43:22SPEAKERwill talk about and like, yeah, I've always wanted to like

43:26SPEAKERwrite in a question to next lander and be like, you know, are you

43:30SPEAKERdo you guys notice the ray tracing stuff?

43:32SPEAKERHow do you guys feel about the ray tracing stuff?

43:34SPEAKERObviously, they're building that into the hardware.

43:37SPEAKERMark Cerny is coming out saying, you know, that it's it's an important part.

43:41SPEAKERYeah, I mean, I personally I remain incredibly bullish about the role

43:46SPEAKERof ray tracing and development and graphics in the long term.

43:50SPEAKERBut I think we're still like on the in the middle of the

43:54SPEAKERyou know, whatever we're defining as long term, right?

43:56SPEAKERLike I felt like when real time ray tracing first became a thing, it was like, OK,

44:00SPEAKERthis is going to take like 10 to 15 years

44:02SPEAKERbefore this is absolutely integral and everywhere.

44:05SPEAKERAnd I would add that I think that is also kind of not a popular

44:09SPEAKERtopic among the mainstream right now. Yeah.

44:13SPEAKERBecause we're not to that point yet.

44:15SPEAKERLike a lot of people still just see it as a a feature that offers

44:18SPEAKERdiminishing returns for a huge performance hit.

44:22SPEAKERAnd also, I feel like a lot of people don't know that they're seeing it right.

44:26SPEAKERI mean, especially like on PC, there's usually a toggle, some sort of,

44:30SPEAKERyou know, hey, do you want to turn ray tracing on?

44:32SPEAKERDo you want path tracing, things like that on console?

44:35SPEAKERYou're usually picking between different quality modes.

44:38SPEAKERAnd sometimes the quality mode will say, oh, yeah, this involves ray tracing.

44:42SPEAKERBut sometimes you have to wait for the digital foundry video

44:45SPEAKERwhere they're like, oh, by the way, this part of this pipeline is using ray tracing.

44:51SPEAKERYou know, it's one of those things where it's like, man,

44:53SPEAKERhow do you how do you really sell that?

44:55SPEAKERHow do you get excited for next consoles when it's like, man,

44:57SPEAKERcheck out this ray tracing performance?

44:59SPEAKERI don't think they're going to have to sell it so much

45:01SPEAKERas it's just going to become a fact of life, because when the next consoles

45:04SPEAKERcome out, they are going to be so capable of ray tracing

45:06SPEAKERthat everybody's just going to adopt it into their pipeline.

45:08SPEAKERAnd I think it'll just it'll just be there.

45:11SPEAKERI love that optimism.

45:13SPEAKERWell, I mean, you're seeing it now, you're seeing it like like Indiana

45:16SPEAKERJones and Doom were the big games this year that you can't run

45:20SPEAKERon non ray tracing hardware. Like that's true.

45:22SPEAKERI think there are going to be tremendously more games.

45:24SPEAKERIn fact, I think just about every game.

45:26SPEAKERBut again, that's that's 10 years from now is what I'm saying.

45:28SPEAKERIt's like it's not going to be until the next consoles have a pretty significant

45:32SPEAKERinstall base that then developers because that's how it's always gone since

45:37SPEAKERsince the 3D, 2D transition, since the PS2 generation, if not before.

45:41SPEAKERIt's always been the consoles, what the consoles can do dictates

45:44SPEAKERwhat trickles down to PCs.

45:47SPEAKERAnd so once every console on the market, like even the switch

45:50SPEAKERtwo has capable ray tracing, right?

45:52SPEAKERI mean, once every every console on the market can do

45:55SPEAKERsignificant ray tracing, then developers are going to be like, OK, well,

45:58SPEAKERnow that we are doing this by default on the set top boxes,

46:02SPEAKERit's also going to come to every PC game.

46:05SPEAKERYeah. And that it's it's it's just interesting, right?

46:08SPEAKERIt's like you always hear a bunch of buzzwords

46:10SPEAKERand the companies are like, to a certain degree, we have to put the cart

46:13SPEAKERbefore the horse and we need we need to use in millions of of CPUs before

46:20SPEAKERor processors before, you know, you can see anything

46:23SPEAKERthat actually uses it day to day. Right.

46:26SPEAKERIt's interesting getting capable ray tracing into,

46:30SPEAKERyou know, mainstream laptop hardware, the kind of computers

46:34SPEAKERthat most people buy is part of that. Right.

46:37SPEAKERAnd I think that's that's that's a step in in that direction,

46:41SPEAKERwhether it's it's obviously not going to perform as well as a discrete

46:45SPEAKERstandalone GPU or console even.

46:47SPEAKERBut but it's a step in the direction.

46:49SPEAKERYeah, I mean, it's it's it's it's the kind of thing where, like, you know,

46:52SPEAKERiPhones also have ray tracing hardware in them.

46:54SPEAKERNow, like at some point in the not too distant future,

46:57SPEAKERevery current shipping product is going to have some amount of ray tracing

47:00SPEAKERand some amount of machine learning capability.

47:03SPEAKERAnd it's just going to become a mundane fact of software development.

47:07SPEAKERLike everything else.

47:08SPEAKERAnd then the other the other part of that is that as the comes more capable,

47:12SPEAKERthen we'll see software that actually takes advantage of it.

47:14SPEAKERLike there were a handful. Yeah, that's what I mean.

47:16SPEAKERThere were a handful of puzzles in Indiana Jones

47:18SPEAKERthat used the light rendering to be as part of the puzzle.

47:22SPEAKERBut it was kind of it was it wasn't super common in that game. Right.

47:25SPEAKERThere weren't a bazillion of those.

47:27SPEAKERSo and also, if you didn't know how it worked,

47:32SPEAKERthen it was functionally no different than like a light reflection puzzle

47:35SPEAKERin Breath of the Wild or Tears of the Kingdom 10 years ago. Yeah. And and like.

47:40SPEAKERLike, I thought it was cool because I recognized it and I saw it,

47:43SPEAKERbut everybody else is just like, oh, OK, whatever.

47:46SPEAKERI mean, that's always been part of the hard sell with ray tracing, right?

47:48SPEAKERIs that the fakery?

47:49SPEAKERThe fakery got so good that telling the difference between fake lighting

47:54SPEAKERand quote unquote real or realistic, physically realistic lighting.

47:58SPEAKERLook, it's hard for a lot of people.

47:59SPEAKERIt's the same industry, though, that counts the reflections on sweat

48:02SPEAKERon people's faces, on console jumps.

48:06SPEAKERSo like, hey, look, look at your nose.

48:08SPEAKERShadowing the side of your face should be an easy talking point

48:11SPEAKERafter puddle gate anyway.

48:14SPEAKERTo me, to me, the ray tracing thing has always been a dark alleyway phenomenon,

48:18SPEAKERmeaning that like every game that has fake lighting

48:20SPEAKERgets a lot of nice artistic attention in the critical path.

48:25SPEAKERBut like dark little alleyways off to the side, they don't bother doing

48:28SPEAKERall the hand tuned lighting to make everything look super great, right?

48:32SPEAKERLike the thing with ray tracing is like, hey, everything is the same lighting now.

48:35SPEAKERLike it's like they don't have to go in and hand place lights

48:37SPEAKERand an art direct lighting to make things look the way they want.

48:42SPEAKERLike like now algorithmically,

48:43SPEAKERjust everything gets the same kind of realistic lighting globally.

48:47SPEAKERThe early ray tracing demos, right, that were touted out

48:50SPEAKERwhere there is battlefield or even control where they're like, oh, well,

48:53SPEAKERlook at these reflections. Wow, these reflections.

48:55SPEAKERAnd even I was like, it's cool stuff, but and it wasn't until

48:59SPEAKERlike the global illumination stuff like Metro was just like, oh, wow.

49:04SPEAKERThat like that is impressive for me because you could actually like

49:08SPEAKERlike real light bouncing around.

49:10SPEAKERYes, that's exactly that kind of stuff.

49:12SPEAKERThat's exactly it.

49:13SPEAKERI got I got kind of bummed out when ray tracing became the even for people

49:16SPEAKERwho were into it, it kind of became the like better reflections tech,

49:19SPEAKERquote unquote, for a long time, because like bounce lighting is the actual

49:23SPEAKERsuper interesting, like brain tickling one for me is like,

49:28SPEAKERyou know, light shines on this apple

49:30SPEAKERand then it picks up red in its in its bounces.

49:32SPEAKERYou know, now now you've got like a red glow on the thing is bouncing out, too.

49:36SPEAKERLike that's the kind of stuff you didn't really see before.

49:39SPEAKERWell, and there's secondary and tertiary effects as well.

49:42SPEAKERSo not to make this whole thing a treatise

49:45SPEAKERon why ray tracing rules and the Internet's wrong.

49:48SPEAKERBut but but like because the cost of moving lights

49:52SPEAKERand changing things that affect the light rendering in a game

49:55SPEAKERgoes down for developers when they're iterating,

49:58SPEAKERit means they're able to iterate more and they're able to make more changes.

50:01SPEAKERAnd they spend less time on that and more time on other things.

50:04SPEAKERAnd then there's the whole tertiary effect of, oh, by the way,

50:07SPEAKERif we have hardware that's really good at tracing rays,

50:10SPEAKERwe can make more realistic sound models and things like that.

50:13SPEAKERAnd we because we can use that same hardware to do to do sound renders.

50:16SPEAKERYes, I have spoken to at least one game developer

50:19SPEAKERwho has talked about how excited they are to use the.

50:23SPEAKERUse the properties of ray tracing hardware for things other than lighting

50:26SPEAKERand sound sound was one of those like A.I., not not machine learning A.I.,

50:31SPEAKERbut enemy A.I. was another one like they basically were very excited about,

50:35SPEAKERlike, you know, it's because it's just hardware for modeling interactions

50:38SPEAKERin a physical space.

50:39SPEAKERSo you can think of all the other ways that that would be useful beyond lighting.

50:43SPEAKERWell, it's funny.

50:44SPEAKERI talked to somebody the other day who's building a

50:46SPEAKERpathing system for their game that uses the ray tracing hardware. Right.

50:50SPEAKERBecause like the expensive and hard thing about generating what's pathable

50:54SPEAKERand what's not is that you run this thing that takes hours and hours to run.

50:57SPEAKERBut if you have ray tracing hardware in the machine,

50:59SPEAKERyou know, you're going to have that.

51:00SPEAKERYou can change that dynamically based on changes in the environment

51:03SPEAKERin a way that's unscripted.

51:05SPEAKERAnd like like anyway, there's there's a bunch of cool knock on effects

51:08SPEAKERthat come with ray tracing hardware that aren't reflections or shadows or

51:12SPEAKERor reflected light even.

51:14SPEAKERSo anyway, yeah, my my my feeling is it's going to be like next consoles

51:18SPEAKERplus three years or so before we really start seeing like

51:21SPEAKERkind of universal effects.

51:23SPEAKERI think that's showing up in games maybe four years.

51:26SPEAKERI don't know. I mean, anyway, the good thing is, I think I think

51:30SPEAKERI think it's going to be based on the sales rate and how long the

51:33SPEAKERlike the the kind of transitory part of the generation is.

51:37SPEAKERLike if we look at last go round, there was a long period

51:40SPEAKERfrom launch of PS5 and Xbox Series X

51:44SPEAKERto the games not being released on PS4 and Xbox One. Yes.

51:49SPEAKERAnd I think like that being so long kind of held back

51:52SPEAKERthat first part of the generation.

51:54SPEAKERYeah, that's totally right.

51:55SPEAKERAnd that's going to be the gate on this, because the thing is

51:59SPEAKERthe hardware has been there for a while.

52:00SPEAKERUE5 is good at ray tracing both in software and in hardware.

52:04SPEAKERAnd we'll like the technologies there.

52:07SPEAKERThe people who've been working on this will be able to do it.

52:09SPEAKERThe question is, does it make a market sense

52:10SPEAKERbecause they want to be able to sell games on PS5?

52:12SPEAKERI don't think they care about the Series X.

52:13SPEAKERSo that's half of that problem solved anyway.

52:17SPEAKERShould we talk about the fab?

52:18SPEAKERYeah, because I think this was this was the cool.

52:21SPEAKERLike this is why I went to this trip.

52:23SPEAKERI could have looked at the slides when I got back and gotten most.

52:26SPEAKERI mean, actually, that's not true.

52:28SPEAKERTalking to the Intel Fellows was it was incredibly informative

52:32SPEAKERand super worth making the trip to Arizona for.

52:35SPEAKERThe fab tour was kind of icing on the cake for me.

52:38SPEAKERDid you guys put on bunny suits?

52:40SPEAKERWe did. Oh, yeah.

52:42SPEAKERHead to toe. We had a we had a beard nets.

52:46SPEAKERYeah. As well as hair nets.

52:48SPEAKERWell, back when I had a beard back in those days.

52:51SPEAKERYeah. No. Yeah.

52:52SPEAKERI felt like because we had small groups.

52:55SPEAKERI think it was like 12 of us that was going in one at a time.

52:58SPEAKERYeah, that's right.

52:58SPEAKERTen or twelve one at a time, like the group at a time.

53:01SPEAKERAnd yeah, they I felt like I was a piece of meat getting shut through

53:05SPEAKERbecause they would give you your suit, your whole your whole setup,

53:09SPEAKERand then they would have somebody like help you do it step by step.

53:13SPEAKERAnd it was intense.

53:14SPEAKERThey kind of warned you it was going to be warm inside the suit.

53:18SPEAKERBut also in some ways, like then later when we were in it,

53:22SPEAKERlike we could kind of stick out our nose and they were like, oh, yeah, it's fine.

53:26SPEAKERYou can have your nose exposed.

53:28SPEAKERWe want you to be comfortable.

53:29SPEAKERIt's like, wait, I thought you just had us go through all this stuff to minimize

53:33SPEAKERdust particles in the air, skin particles.

53:36SPEAKERDoes that officially qualify as a clean room, did they say?

53:38SPEAKEROh, yeah.

53:39SPEAKERSo there's ISO standards for clean rooms that are based on how much particulate

53:43SPEAKERthere is per square foot or square meter of air or cubic meter, rather, of air.

53:48SPEAKERAnd this one is not.

53:50SPEAKERSo this fab, the 18A fab that we were in and which was fab, what, 50, 53?

53:57SPEAKEROh, I don't remember the numbers.

53:59SPEAKERI don't remember. Anyway,

54:01SPEAKERthe we we went through basically one fab to get to the new fab.

54:05SPEAKERAnd both of those use these things called foops that are enclosed wafer holders.

54:11SPEAKERThey're like, we've seen them at the flea market before.

54:14SPEAKERThey're like big plastic containers that are the ones they were kind of orangey,

54:18SPEAKERI assume, to keep certain wavelengths of light out.

54:21SPEAKERAnd they carry the wafers from machine to machine.

54:24SPEAKERAnd these caddies that ride on rails above the floor of the fab, like 30 feet up.

54:29SPEAKERAnd so they have a lower standard than some other fabs

54:33SPEAKERthat use exposed wafers where they need, you know, much lower particulate.

54:38SPEAKERThey wear fully enclosed heads or rebreathers or masks or whatever.

54:42SPEAKERBut yeah, we so we walked into the the like the clean room

54:45SPEAKERto get into the clean room area.

54:46SPEAKERWe had to be badged in through a gate.

54:48SPEAKERAnd once we were on the other side of the gate, we stepped in and put

54:53SPEAKERlike foot coverings on shoe coverings on.

54:55SPEAKERI was waiting to hear about the booties.

54:56SPEAKERI knew that they had to be in there.

54:58SPEAKERYeah. So we put booties on to keep dirt from our shoes off of the floor

55:02SPEAKERin the clean room gowning area.

55:04SPEAKERAnd then we put on beard nets

55:05SPEAKERand hair nets to keep the particulate from our face,

55:08SPEAKERbecause like the humans, the grossest thing in the whole fab, right?

55:11SPEAKERWe're just constantly shedding and anywhere really, not just that.

55:16SPEAKERAnd so then we we got we went in, they sized us up for the bunny suits.

55:21SPEAKERThe people there immediately knew what we were doing and fit us perfectly.

55:25SPEAKERAnd we got the bunny suits and then outer outer.

55:28SPEAKEROh, and we put on gloves before we got into the clean room area, too.

55:32SPEAKERLike two clothes. Yeah.

55:34SPEAKERSo we put on the gloves.

55:36SPEAKERWe put on we got boots, we got a hood, we got a bunny suit

55:41SPEAKERand we put the bunny suit on, then put the boots on on top of those.

55:45SPEAKERBut the hood, which got tucked into the bunny suit,

55:48SPEAKERhad our hair nets and beard nets on already and put on like

55:54SPEAKERlatex gloves on top of the cotton gloves.

55:56SPEAKERIs that right? Or was it the other way around?

55:57SPEAKERLatex first and then cotton. Yeah.

55:59SPEAKERSo night trial, we use night trials, latex allergy people didn't die.

56:05SPEAKERBut we didn't have to wear a mask like the hood kind of covered

56:08SPEAKERyour mouth, kind of covered your nose if you wanted to.

56:12SPEAKERAnd we put goggles on on top of the glasses we were wearing

56:16SPEAKERor if you wore eyeglasses.

56:19SPEAKERAnd then and then once that was all done, we like put our little badges on.

56:23SPEAKERSo they knew we were part of a tour and not to be trifled with

56:26SPEAKERand kind of got let out into the fab floor, which was almost like

56:32SPEAKERthe thing that struck me immediately was how much movement there was.

56:35SPEAKERThere was there was.

56:38SPEAKERA lot of people, a lot of machines scurrying,

56:41SPEAKERlike the caddies that carried the foops around were constantly going

56:46SPEAKERand they were carrying the what?

56:47SPEAKERThe foops, the things that carried the fiber, the the the wafers, the wafers.

56:54SPEAKERAnd I thought, yeah, they look like little plastic containers

56:57SPEAKERthat just have a like a stack of wafers in it.

57:01SPEAKERAnd like that, that's sealed as well.

57:03SPEAKERSo like, I think that's why like they have multiple layers of ceiling.

57:07SPEAKERLike I always thought of a clean room is like a like a like

57:10SPEAKERkind of like a medical space. Right.

57:12SPEAKERAnd you got very limited people in there and they're doing stuff real carefully.

57:15SPEAKERThis was like, I mean, this is a huge, huge building,

57:19SPEAKERyou know, with with dozens and dozens of people all doing all this

57:22SPEAKERdifferent kind of stuff.

57:24SPEAKERAnd they were talking about how the advance, the filtration system

57:27SPEAKERwas in the building, like the air, the the air was moving from top to bottom.

57:33SPEAKERAnd the bottom grading was like vented.

57:35SPEAKERAnd it was just like, yeah.

57:37SPEAKERSo even if you shed a particle, it was going to be sucked into the floor

57:40SPEAKERwithin a matter of seconds.

57:42SPEAKERAnd like you could feel it, just to be clear,

57:44SPEAKERyou could feel the breeze even inside the bunny suit.

57:48SPEAKERAnd so so you walk into this room and there's like a bunch of hallways

57:53SPEAKERthat I think were food like drop points so they could drop those caddies.

57:57SPEAKERI sent you a link to a way for a way for world article

57:59SPEAKERthat explains what hoops are and doesn't exactly show them.

58:02SPEAKERBut it stands for kind of kind of looks like a record holder.

58:05SPEAKERYeah. It's almost like a big CD changer, honestly, a CD changer cartridge.

58:11SPEAKERBut they're they're there.

58:13SPEAKERIt stands for front opening unified pod or front opening universal pod.

58:18SPEAKERAnd the idea is that they're sealed containers

58:20SPEAKERthat carry the wafers from machine to machine.

58:23SPEAKERSo like these caddies are zooping along 20 or 30 miles an hour, 30 feet in the air.

58:27SPEAKERAnd there's a bunch of tubes and stuff coming up out of the floor.

58:31SPEAKERAnd the whole thing is set up.

58:32SPEAKERAnd I talked to the guy was like, do your guys play Factorio when they go home?

58:36SPEAKERIs this like the shit that they love?

58:38SPEAKERBecause the whole thing is set up

58:40SPEAKERso that the really expensive machines are never sitting idle, right?

58:43SPEAKERLike as soon as one of the foops gets lifted off, the next one gets put in place

58:48SPEAKERso that it's fooping constantly like like it's foop to foop, man.

58:53SPEAKERNever stops me fooping. Yeah.

58:57SPEAKERSo so yeah.

58:59SPEAKERSo you see these things I've cracked up at them.

59:02SPEAKERI'm sorry. You see these things done.

59:04SPEAKERYou see these things flying around and then they get to the machine

59:07SPEAKERthat they're above and it lowers down the caddy on like two wires

59:11SPEAKERfrom this from this thing 30 feet in the air and drops it directly

59:15SPEAKERon the foop mounting port.

59:17SPEAKERAnd then the machine docks with it and sucks the the wafers in does its thing

59:22SPEAKERand then spits them back out, probably in the same same place. Right.

59:27SPEAKERAnd and all of the machines are lined up.

59:30SPEAKERSo those foop docks are even with each other and directly below the tracks.

59:33SPEAKERThat that seems to be the main defining characteristic.

59:36SPEAKERAnd then the guy who gave us the tour, Bill, something or other,

59:41SPEAKERwas saying that they're they're

59:45SPEAKERlike that whole structure is something like they have they have buffers

59:49SPEAKERfor the slower machines so that the expensive machines have

59:53SPEAKERhave have wafers ready to go that are like X number of minutes deep

59:57SPEAKERin case one of the machines goes down.

59:59SPEAKERSo that this thing that costs half a billion dollars is never sitting idle,

1:00:02SPEAKERbasically. Sure. Yeah. That makes sense.

1:00:04SPEAKERYeah. It was kind of crazy tense.

1:00:06SPEAKERYeah. Like in that that was the weird part is that, yeah, they were I mean,

1:00:10SPEAKERwhat, like 40 feet, 30 feet over? I think 30 feet probably. Yeah.

1:00:13SPEAKERYeah. Like really high. And when they would go down into the machine,

1:00:17SPEAKERthey're coming down just by like a couple of wires.

1:00:19SPEAKERAnd they're not they're not like swaying in the wind or something like that.

1:00:22SPEAKERIt just looked like they came straight down, came back up

1:00:25SPEAKERthree miles an hour, like kind of seems pretty fast

1:00:27SPEAKERfor delicate stuff like that to be whizzing around overhead.

1:00:30SPEAKERBut it's all I mean, it's all sealed and it's like all that.

1:00:33SPEAKERThere's no it's not like rattling around in there.

1:00:35SPEAKERYeah. I want to ask real quick in your notes here.

1:00:39SPEAKERYeah. You have you have listed that this land was bought in the 70s.

1:00:42SPEAKERYes. And they sorry.

1:00:45SPEAKERAnd then they broke ground in the 90s.

1:00:47SPEAKERIs this a is this a brand new facility?

1:00:49SPEAKERSo or is there was there have there been parts

1:00:52SPEAKERof this facility that were already in operation prior to this?

1:00:55SPEAKERThis this fab is new like this. This fab is.

1:00:59SPEAKERYeah, because it's huge.

1:01:00SPEAKERIt's bigger than any of the ones I've done before.

1:01:02SPEAKERSo so my actual ultimate question here,

1:01:04SPEAKERhave they been working on this facility since the 90s?

1:01:07SPEAKERNo, no, no. There's other fabs at this facility.

1:01:10SPEAKERThat's what I was. So, OK.

1:01:11SPEAKERSo there's a lot of other processes going on with other,

1:01:14SPEAKERyou know, process technologies.

1:01:16SPEAKERSo, OK, I think that they had I think they said they have a million

1:01:20SPEAKERor two million square feet of fab space at this at this campus.

1:01:25SPEAKERThe clean room area you like you can walk from one fab to the other,

1:01:29SPEAKERall in clean room.

1:01:31SPEAKERYou can move flops from clean room to clean room.

1:01:34SPEAKERBut I mean, I don't it seemed like they don't do a whole ton of that.

1:01:39SPEAKERThey did anybody ask why they set up a water

1:01:42SPEAKERintensive industrial process in the desert?

1:01:45SPEAKERYeah, I did ask that.

1:01:46SPEAKERThat was one of the first things I said.

1:01:48SPEAKERThis is fab 52, by the way, not not.

1:01:52SPEAKERSo the reason they did in the desert is because being tectonically stable

1:01:56SPEAKERand having reliable power is like the water problem is tractable

1:02:01SPEAKERand they can reclaim a fairly enormous amount of water.

1:02:05SPEAKERI think they said they go through nine million gallons of water a day

1:02:08SPEAKERand they treat some enormous percentage of it.

1:02:11SPEAKERI don't think they have an exact number,

1:02:15SPEAKERbut the power and tectonic stability is much more important.

1:02:18SPEAKERAs the person who's giving us the tour said, if they had known

1:02:22SPEAKERthat the organ fabs are on a huge fault

1:02:26SPEAKERat the time that they built them, they probably wouldn't have built them there.

1:02:29SPEAKEROh, OK. And it's interesting because, of course, Taiwan has the other big fabs

1:02:33SPEAKERlike this, and it's also on like pure Ring of Fire out there.

1:02:36SPEAKERSo I asked what happens if there's an earthquake, because they do

1:02:41SPEAKERoccasionally have earthquakes that Southern California earthquakes

1:02:44SPEAKERsometimes make it through to Arizona.

1:02:46SPEAKERAnd basically all of the seismically sensitive machines

1:02:50SPEAKERhave the equivalent of like old hard drive head parking technology

1:02:53SPEAKERwhere they just immediately freeze and shut down.

1:02:56SPEAKERIf there's any kind of ground trembler

1:03:00SPEAKERand also the building is somewhat seismically isolated, apparently.

1:03:05SPEAKERThe other thing that Adam and we didn't talk about at all

1:03:07SPEAKERis that there's subfloor beneath the area where this is all built.

1:03:12SPEAKERWhere like a lot of the machines, you see the tops poking out,

1:03:15SPEAKERbut they're like icebergs and the rest of them are like that.

1:03:18SPEAKER75 percent of the machine is in the subfloor underneath where like water

1:03:21SPEAKERand chemicals and whatever, you know, ions and stuff come through.

1:03:26SPEAKERLike I've been in a couple of places that are mechanically

1:03:29SPEAKERincredibly complicated before, like car factories and stuff like that.

1:03:32SPEAKERAnd I've never seen anything that had this level of density

1:03:36SPEAKERwhen it comes to technology and kind of what you see.

1:03:40SPEAKERSo it's pretty wild.

1:03:42SPEAKERYeah, no, I mean, the actual fab tour was definitely

1:03:47SPEAKERonce of a lifetime kind of thing.

1:03:48SPEAKERThere's been other opportunities in the past and I've never been able to attend.

1:03:52SPEAKERSo I appreciate going and checking it out.

1:03:55SPEAKERAnd just like it is it is crazy.

1:03:58SPEAKERBut not that I take any of this PC hardware for granted,

1:04:02SPEAKERbut at the end of the day, when you're playing a game or,

1:04:05SPEAKERyou know, working on Excel or something like that, you're not thinking about

1:04:08SPEAKERall the crazy technologies that it took just to get that that CPU into your hands.

1:04:14SPEAKERYeah, I mean, the thing about the thing about,

1:04:17SPEAKERyou know, semiconductor fabrication to me is that it's so elemental, you know,

1:04:21SPEAKERit's just metal, it's just giant disks of metal that they are turning

1:04:24SPEAKERinto something that can do all the things a computer can do.

1:04:27SPEAKERLike, it's just hard to wrap your head around that process sometimes.

1:04:31SPEAKERWell, in these.

1:04:33SPEAKERSo this this fab in particular takes wafers in and spits wafers out.

1:04:38SPEAKERThere's no like they're not cutting the chips out here or anything like that.

1:04:41SPEAKERThey're only doing those compute tiles here right now.

1:04:43SPEAKERThey're not doing the other.

1:04:45SPEAKERThey're stamping the wafers. Yeah.

1:04:47SPEAKEREtching whatever the technical term is in smartening.

1:04:51SPEAKERYeah. And then and then the other thing is that it operates 24/7, 365.

1:04:57SPEAKERAnd there's always around 4000 people on campus, which I didn't like.

1:05:01SPEAKERA lot of the other fab situations that I've been familiar

1:05:04SPEAKERwith are a little more automated and a little less human intensive.

1:05:08SPEAKERBut like there were a lot load of people around here all the time.

1:05:11SPEAKERKind of kind of crazy to think about

1:05:13SPEAKERpunching the clock on the graveyard shift at the at the 18 a fab.

1:05:17SPEAKERI mean, you wouldn't you wouldn't be able to tell there's no windows.

1:05:20SPEAKERThere's nothing like the lighting was had to be weird anyway,

1:05:24SPEAKERfor whatever reason of, you know, light waves.

1:05:28SPEAKERBut yeah, it's got to be intense for sure.

1:05:31SPEAKERI mean, so the other the couple other things

1:05:33SPEAKERthat I thought were kind of weird and interesting.

1:05:35SPEAKERApparently, if you work with copper in the fab,

1:05:38SPEAKERbecause it's a contaminant at a molecular level,

1:05:43SPEAKERyou only can work on like, yeah, they make you wear a different suit,

1:05:47SPEAKERa different color suit, and you use different tools.

1:05:49SPEAKERYou come in through a different door.

1:05:52SPEAKERAnd every time they walk by the non copper people, they give you a noogie.

1:05:56SPEAKERLike, yeah, like, man, it is really, really classist over there.

1:05:59SPEAKERDifferent. Oh, go ahead.

1:06:01SPEAKERSo I'm sorry. You mean contaminant for the fabrication process?

1:06:04SPEAKERI assume not for human beings.

1:06:06SPEAKERYeah. Contaminant for the fabrication process.

1:06:08SPEAKERGot it. They they like they really like the machines

1:06:12SPEAKEReither had copper, OK, or no copper written on the side of all of them.

1:06:17SPEAKERAnd it was because like a handful of atoms of copper

1:06:20SPEAKERcan contaminate a whole wafer or a whole machine.

1:06:23SPEAKERApparently, it seemed real, real serious.

1:06:25SPEAKERThey were very serious about that.

1:06:27SPEAKERI guess that makes sense.

1:06:28SPEAKERIt would it would alter the conductivity properties, right?

1:06:30SPEAKERTo have a different metal mix presumably. Yeah.

1:06:33SPEAKERThey didn't we weren't allowed to wear a hair product or

1:06:38SPEAKERlike makeup, cologne, makeup,

1:06:41SPEAKERnothing, nothing with volatile esters in it.

1:06:44SPEAKERBasically, they did tell us we could wear deodorant, which was I was grateful for.

1:06:49SPEAKERI skipped it just for fun.

1:06:50SPEAKERYeah. I mean, I thought about it,

1:06:51SPEAKERbut I decided for your benefit, I would I would I would do it or not.

1:06:56SPEAKERAnd yeah, like it was I think we were inside for maybe 45 minutes

1:06:59SPEAKERand walked like, what, a half a mile or something.

1:07:01SPEAKERWe couldn't take any electronics in both from an E.M.

1:07:05SPEAKERinterference perspective, because a lot of the machines are E.M. sensitive,

1:07:09SPEAKERbut also obviously because they didn't want us taking pictures.

1:07:11SPEAKERAnd they like it was funny because we saw like machines

1:07:15SPEAKERrunning old software, like old versions of windows and stuff

1:07:18SPEAKERhooked up to different things.

1:07:20SPEAKERAnd when we asked what those machines were, they were like, well,

1:07:21SPEAKERwe we're not talking about that machine today.

1:07:24SPEAKERWe're not, you know, blah, blah, blah.

1:07:26SPEAKERSneak some spy photos onto a rage bait YouTube channel.

1:07:29SPEAKERIntel's running 18A on Windows seven.

1:07:33SPEAKERAbsolutely. We're running Windows seven.

1:07:36SPEAKERWe did see the EUV machine, which is the half billion dollar

1:07:40SPEAKERextreme lithography machine that's like the size of three school buses.

1:07:45SPEAKERAnd we get to touch it.

1:07:46SPEAKERThey didn't let us touch it.

1:07:48SPEAKERYeah, it's big. Yeah, it's big.

1:07:50SPEAKERIt's nondescript.

1:07:51SPEAKERYou could walk by it and just think it was nothing.

1:07:53SPEAKERBut yeah, it was it was intense.

1:07:55SPEAKERAnd there was space for at least two more of those next to it.

1:07:58SPEAKERSo, you know, are those the ones that I keep seeing stuff

1:08:02SPEAKERabout the supply chain for those is like a single point of failure

1:08:05SPEAKERwhere it's like one company in Germany makes the mirrors,

1:08:08SPEAKERlike one company in Denmark makes this and that. I don't know.

1:08:12SPEAKERIt seems like they're all mostly bespoke, but they also really specialized.

1:08:17SPEAKERYeah, like they did.

1:08:18SPEAKERI don't think that machine even had logos on it for whoever made it.

1:08:21SPEAKERIt's like it's like it's like if you have to ask

1:08:24SPEAKERwhere you get an EUV machine, you're probably out of luck, I guess.

1:08:28SPEAKERYou're probably dealing with some nation state level approvals or something.

1:08:31SPEAKERYeah. To be clear, they do this.

1:08:35SPEAKERThis this machine operates in vacuum

1:08:37SPEAKERand you can't use glass for mirrors because the EUV light,

1:08:42SPEAKERthe extreme ultraviolet light is absorbed by both air and glass.

1:08:47SPEAKERSo they did they had to do all sorts of like

1:08:50SPEAKERcrazy material science stuff to make these things even possible.

1:08:54SPEAKERAnyway, that that was kind of the hype.

1:08:58SPEAKERWe stopped at the EUV machine and they made us turn around and go back.

1:09:01SPEAKERWe didn't get to see what's on the other side of that.

1:09:03SPEAKERAnd it's unclear to me if that machine it was really long.

1:09:06SPEAKERI don't know if like the wafers went in one end

1:09:09SPEAKERand they got blasted by the thing at the end of it,

1:09:11SPEAKERor if it was like going on a little line down it and stuff was happening as it went.

1:09:14SPEAKERSo who knows? Is it possible to say I don't even think we could tell

1:09:18SPEAKERlike where the food centered that I didn't see a food poll on those now.

1:09:24SPEAKERYeah, and there wasn't like people manning it either.

1:09:26SPEAKERLike nobody was like standing around it.

1:09:28SPEAKERIt was just just sitting there doing its thing.

1:09:30SPEAKERBut like talking about how they kept keep that thing fed

1:09:34SPEAKERwas giving me real, real factorio and satisfactory flashbacks to like,

1:09:40SPEAKERoh, OK, so you need a hopper that stores

1:09:42SPEAKERfoops that are ready for this that's nearby.

1:09:45SPEAKERAnd and the guy Bill, who's taking us on the tour, was like, yeah,

1:09:49SPEAKERall that stuff kind of happens like the systems that drive that just know that

1:09:53SPEAKERthe the poop that they're carrying has this wafer on it at this stage.

1:09:57SPEAKERAnd it needs to go to this area next.

1:09:59SPEAKERAnd it knows that there's slack space at these five places.

1:10:02SPEAKERAnd it just puts it the one that's closest.

1:10:04SPEAKERAnd he's like, it's not it's it's almost like a packetized delivery network,

1:10:09SPEAKERlike a network, like an IP network, but for wafers,

1:10:13SPEAKERwhich which like when he explained it that way, it made sense.

1:10:15SPEAKERSo that's wild. Yeah, that's all wild.

1:10:19SPEAKERIt was it was like being in Willy Wonka's

1:10:22SPEAKERchocolate factory, but less Hugh Grant and more foops.

1:10:27SPEAKERSnowsberries taste like snowsberries.

1:10:29SPEAKERYeah. Adam didn't let us taste snowsberries.

1:10:32SPEAKERAdam didn't pick any USB keys off the floor either.

1:10:34SPEAKERSo I was looking. Yeah.

1:10:36SPEAKERNot that I wouldn't have.

1:10:38SPEAKERYeah, probably for the best.

1:10:39SPEAKERYeah. Yeah. Is this a good place to ask some wrap up questions?

1:10:43SPEAKERLet's do it. So as somebody who wants somebody

1:10:48SPEAKERwho's ready for a new desktop CPU, there is nothing here for me yet.

1:10:51SPEAKERRight. Like this is Panther Lake is all mobile, like Nova Lake.

1:10:54SPEAKERThey're not even saying the words right.

1:10:55SPEAKERLike there's nothing for me to grasp on to yet.

1:10:58SPEAKERAnd they have confirmed no Panther Lake on desktop. Right.

1:11:01SPEAKERSo. Right. OK.

1:11:02SPEAKERThey haven't said what's next.

1:11:04SPEAKERBut like I said, I I believe that it's going to be an Air Lake refresh.

1:11:07SPEAKERThat's yes. But but as far as as far as like a fully next generation

1:11:13SPEAKERdesktop part on 18A and stuff, that's all that sounds like that's all a good ways off.

1:11:18SPEAKERI mean, yeah. Or maybe they skip 18A for desktop.

1:11:21SPEAKERI don't even know.

1:11:23SPEAKERI have seen I have seen the term 14A starting to float around.

1:11:27SPEAKERYeah. So I actually don't know.

1:11:30SPEAKERDo you guys have any sense of like how did Intel get on this leapfrog

1:11:34SPEAKERHayden Swear Lake Meteor Lake and Lunar Lake were mobile only.

1:11:38SPEAKERAnd then like Air Lake got a desktop part and now Panther is mobile only again.

1:11:41SPEAKERIs it is it purely a function of laptops or what cell now?

1:11:44SPEAKERAnd they just are not emphasizing desktop parts as much or what?

1:11:48SPEAKERNo, they've done that in the past before.

1:11:50SPEAKERLike Ice Lake was mobile only.

1:11:52SPEAKEROK. And that was around like 10th Gen.

1:11:56SPEAKERYeah. Yeah. So like they do that every once in a while.

1:11:59SPEAKERAnd and I think that's the thing is that like where AMD takes pride

1:12:03SPEAKERon sale scaling the same core from mobile to all the way up to the server.

1:12:09SPEAKERIntel usually has different platforms for different market segments.

1:12:14SPEAKERI think I think it's more like, oh, hey, lunar lake.

1:12:18SPEAKERWe really wanted something that was power efficient

1:12:21SPEAKERthat had the NPU that could hit copilot plus.

1:12:25SPEAKERWe can't or there's no point in scaling this technology to a desktop part.

1:12:29SPEAKERLet's have something else do that.

1:12:31SPEAKERLike Intel, Intel kind of plays mix and match,

1:12:34SPEAKERwhich is why they're really leaning into Foveros kind of stuff.

1:12:38SPEAKERAnd so something that they've always said before.

1:12:42SPEAKERWhether it's around NPU or why they got rid of hyper threading

1:12:47SPEAKERor why they've used TSMC sometimes and their own process

1:12:51SPEAKERsometimes is like, oh, we use the right part for the right time

1:12:55SPEAKERor, you know, something along those lines.

1:12:56SPEAKERSo like they they they pretty much mix and match.

1:12:59SPEAKERSo this isn't unprecedented to have architectures

1:13:03SPEAKERthat that are on one area and not the other.

1:13:06SPEAKEROK, that makes sense.

1:13:07SPEAKERThe kind of feeling I wanted I got was that they knew

1:13:11SPEAKERthat Lunar Lake was going to be really good early on,

1:13:15SPEAKERlike the performance was going to be there for mainstream parts.

1:13:17SPEAKERIt was capped at 32 gigs because of the memories on the on the chip,

1:13:22SPEAKER32 gigs of RAM.

1:13:24SPEAKERBut like so.

1:13:26SPEAKERSo that bought them headroom to do more with this.

1:13:29SPEAKERAlthough this also still feels like a fairly concerned,

1:13:33SPEAKERlike this feels like the small bump rather than the big bump.

1:13:36SPEAKERIf we're looking at the kind of old tick tock metaphor

1:13:39SPEAKERthat every time you mentioned while you were there,

1:13:42SPEAKERhe's like, well, we really don't like that metaphor anymore.

1:13:44SPEAKERBut this one's definitely a tick, not a tock.

1:13:46SPEAKERAnd we were like, OK, thanks.

1:13:48SPEAKERThat clears this up a great deal.

1:13:49SPEAKERSo I always thought tick tock was backwards anyway,

1:13:53SPEAKERbecause isn't tock supposed to be the...

1:13:56SPEAKEROh, no, let's not get through this.

1:13:58SPEAKERYeah, I think we got into that with the Intel people.

1:14:00SPEAKERYeah, people were trying to argue that the tock was the next generation

1:14:03SPEAKERand the tick was the evolved version.

1:14:05SPEAKEROr like, what are you talking about? That is backwards.

1:14:07SPEAKERWe had the stupidest conversation with a really smart person

1:14:10SPEAKERabout this at one point.

1:14:11SPEAKERAnd I don't I like at the end of it, everybody was like, I have regrets.

1:14:14SPEAKERLet's just never mention this again.

1:14:16SPEAKERYeah, yeah, yeah, let's talk about something else.

1:14:18SPEAKERYeah, I think the thing is, is that for for many years,

1:14:22SPEAKERI mean, like the way I step back and look at it, when Ryzen when Ryzen hit, we had

1:14:29SPEAKERlike we had like core wars of like, hey, how many cores can you get?

1:14:33SPEAKERRight. And I think we're we've I mean, there's maybe talk about

1:14:38SPEAKERwhatever the next version might be able to scale up to more cores.

1:14:42SPEAKERBut for the most part, we've settled at there's a lot of cores.

1:14:46SPEAKERWe have a lot of cores.

1:14:47SPEAKERWe should probably figure out efficiency.

1:14:50SPEAKERWe should figure out scheduling, things like that.

1:14:52SPEAKERSo I feel like a lot of the a lot of that is swung back in that direction.

1:14:57SPEAKERLunar Lake was a big part of it.

1:14:59SPEAKERObviously, Qualcomm, you know, with

1:15:03SPEAKERwith their Snapdragon elite stuff was pushing for it.

1:15:07SPEAKERAMD, you know, definitely continues to refine that stuff.

1:15:11SPEAKERSo I yeah, I think like that's what we're looking at for for a good while.

1:15:18SPEAKERIt's kind of refining efficiency, which is a big deal.

1:15:22SPEAKERAnd honestly, like, yes, there was a lot of stuff

1:15:25SPEAKERthat scratched heads around Air Lake on the desktop.

1:15:29SPEAKERBut if there's one thing you can say is that, hey, it was more efficient.

1:15:34SPEAKERSo, yes, maybe it wasn't the top of the gaming charts,

1:15:37SPEAKERbut it used way less power and used less heat or produced less heat.

1:15:43SPEAKERAnd like like, yeah, that doesn't look sexy on the back of a box.

1:15:46SPEAKERBut I mean, that's I'm OK with that.

1:15:48SPEAKERThat's totally that's a big deal for me.

1:15:50SPEAKERSo I will absolutely I'll absolutely take like 90 percent of the performance

1:15:54SPEAKERor like 60, 70 percent of the power consumption of the of the high end,

1:15:57SPEAKERlike the market leader, like 100 percent all the time now. Yeah.

1:16:01SPEAKERLike that's that's that's exciting to hear,

1:16:03SPEAKERbecause those have been the pain points of the last few years. Right.

1:16:06SPEAKERLike scheduling being a mess on on Ryzen's and Intel.

1:16:11SPEAKERIt's a mess everywhere through the roof.

1:16:12SPEAKERLike CPUs have kind of not been in a great place this decade

1:16:15SPEAKERin a lot of ways to me.

1:16:16SPEAKERAnd it sounds like these are the growing pains that needed to happen

1:16:20SPEAKERfor these new paradigms to really take hold and mature.

1:16:24SPEAKERYeah, the big two, like I asked Rajshree about the thread scheduler.

1:16:28SPEAKEROne of the things I hadn't really fully understood

1:16:30SPEAKERis that that's a kind of conversation between the CPU and the OS,

1:16:33SPEAKERnot just the CPU saying this, because no one has complete

1:16:38SPEAKERa complete picture, like neither side has a complete picture.

1:16:42SPEAKERAnd and her take was that they were pretty long way

1:16:45SPEAKERalong the way of making that process.

1:16:48SPEAKERLike they've gotten most of the pain points out of that from their perspective.

1:16:52SPEAKERAlso, at the end of the day, though, and they kept hammering home,

1:16:55SPEAKERlike they're they're just providing hints to the OS. Yeah.

1:16:59SPEAKERHints. That's not a conversation.

1:17:01SPEAKERThat's that's a hint. That's true.

1:17:02SPEAKERLike like like that. That feels very.

1:17:04SPEAKERYeah. Like, oh, you know what?

1:17:05SPEAKERThis is what we think you should do.

1:17:06SPEAKERBut they did say at the end of the day, the scheduling is on the operating system.

1:17:12SPEAKERAnd because Will and I are our Linux experts now,

1:17:15SPEAKERwe actually did get to ask some questions of like, hey, does

1:17:17SPEAKERhow does the scheduling work on Linux versus Windows and things like that?

1:17:23SPEAKERHer answer was it really depends on which version of the of the kernel

1:17:26SPEAKERyou're using, because the newer ones are pretty good and the older ones are kind of bad.

1:17:30SPEAKERSo, yeah, I can I can tell you as somebody who followed the kernel mailing list

1:17:33SPEAKERfor several months, watching Intel engineers send patches for for big little core scheduling.

1:17:39SPEAKERThat's yeah. Yeah.

1:17:41SPEAKERSo it was it was really fascinating.

1:17:43SPEAKERI have a better understanding of how this stuff works.

1:17:45SPEAKERI came out with the big takeaway for me

1:17:49SPEAKERis that they're still sticking with the big little cores.

1:17:52SPEAKERThe little cores are getting more competent all the time.

1:17:55SPEAKERAnd the big cores are, you know,

1:17:59SPEAKERlike I used to think the big cores were the interesting thing.

1:18:01SPEAKERNow I'm kind of on team little core for the way that I work most of the time.

1:18:05SPEAKERAnd like I almost given the opportunity, I would probably buy an oops,

1:18:09SPEAKERall ecores CPU before these weird hybrid architectures to simplify things.

1:18:15SPEAKERYeah, like I'm very excited about the prospects of Wildcat Lake,

1:18:18SPEAKERwhich I know that's all rumors.

1:18:20SPEAKERYou guys can't speak to that because obviously Intel isn't going to confirm.

1:18:22SPEAKERThey didn't talk about any of that.

1:18:24SPEAKERYeah, but the rumors about that.

1:18:25SPEAKERI'm looking at that for like a router or whatever and thinking, like,

1:18:28SPEAKERcould you just disable the p cores on that and just use the LPE cores?

1:18:32SPEAKERBecause those are plenty fast now and.

1:18:34SPEAKERWell, and remember when Alder Lake came out and it was the first time

1:18:38SPEAKERthey had the performance cores and efficiency cores.

1:18:44SPEAKERI can't remember who had brought it up,

1:18:46SPEAKERbut somebody said that they equated the ecores

1:18:50SPEAKERto the performance of a 9900K p core.

1:18:55SPEAKERAnd I don't remember the generation.

1:18:57SPEAKERIt was like Skylake.

1:18:58SPEAKERIt was that was the thing I always heard was that the Alder Lake ecores

1:19:02SPEAKERwere roughly the performance of like the cores of a 6700K or a 6500K,

1:19:06SPEAKERwhich is like, yeah, that's still not that old of a CPU.

1:19:11SPEAKERAnd I mean, yeah, that's what's in my machine.

1:19:13SPEAKERRight. Is it 9900K? Yeah. Yeah.

1:19:16SPEAKERAnd I did try to ask and I don't remember the answer.

1:19:19SPEAKERNow I'd have to go back and look at my notes.

1:19:20SPEAKERI don't have them right here, but I did try to ask, like, hey, what are what is

1:19:24SPEAKERthe what is the the processing grunt of an ecore right now equate to?

1:19:28SPEAKERLike how many generations back do we need to go to?

1:19:31SPEAKERThat's a great question.

1:19:32SPEAKERThe p core. And I think he said Alder Lake p core.

1:19:36SPEAKERNo, he said he said the IPC is Raptor Lake.

1:19:41SPEAKEROh, no, no. I didn't write down what the answer was.

1:19:45SPEAKERI have I have your question.

1:19:47SPEAKERAnd I did not write down the answer because I'm a dope.

1:19:49SPEAKERWhat you're saying is basically

1:19:50SPEAKERwhich generations p core is equal to the current generation ecore. Exactly.

1:19:54SPEAKERYeah. If I remember correctly, it was somewhere around like 12th gen.

1:19:57SPEAKERLike, yeah, that's crazy.

1:19:59SPEAKERThat is crazy. Yeah.

1:20:00SPEAKERAnd on a smaller process.

1:20:03SPEAKERYeah. No, you know, way more like the idea of

1:20:06SPEAKERa way more power efficient, say, like four core little box

1:20:09SPEAKERthat is as fast as an Alder Lake would have been.

1:20:11SPEAKERBut with way, way less power consumption like that, we're really in like a

1:20:15SPEAKERwe're a golden age for like home networking tinkerer types. Oh, yeah.

1:20:20SPEAKERBuy a bunch of cheap little powerful low power boxes and do fun stuff with them.

1:20:24SPEAKERWell, and really like and that's like imagine

1:20:28SPEAKERimagine chips that are all oops, all low power ecores. Yes. Right.

1:20:32SPEAKERLike you actually have three sizes, of course.

1:20:35SPEAKERSo that's that's that's the rumor of what wildcat.

1:20:38SPEAKERThat's if you look at all the wildcat like leaks.

1:20:41SPEAKERIt's like two pea cores, four LPE cores.

1:20:44SPEAKERAnd that's why I'm thinking, can you just disable those pea cores

1:20:46SPEAKERand just roll with a quad core that's like close to an older lake, apparently.

1:20:51SPEAKERSo those LPE cores are all on the are on the I/O die,

1:20:55SPEAKERwhich is where the memory controller and all that, the MPU and all that crap is.

1:20:58SPEAKERAnd they're clocked lower.

1:20:59SPEAKERYeah, you know, the next next two or three years

1:21:01SPEAKERgoing to be very exciting for new CPUs from the sound of things.

1:21:04SPEAKERI think so. Yeah.

1:21:05SPEAKERWell, except for at the end of the day.

1:21:07SPEAKERAnd this is what we're talking about with this thread scheduling thing.

1:21:10SPEAKERIt all depends on the OS, because if Windows still has issues with scheduling,

1:21:16SPEAKERyou know, there's only so much Intel and AMD and Qualcomm can do.

1:21:21SPEAKERI've heard there are other operating systems that might be able to help.

1:21:24SPEAKERYeah. Yeah. Run cache 2025.

1:21:26SPEAKERI'm talking about free BSD now.

1:21:28SPEAKERYear of Linux, baby.

1:21:31SPEAKERYou know, actually, the year of the free BSD desktop is coming.

1:21:35SPEAKERYeah. No. Well, actually, this is a real quick side before we get out of here.

1:21:39SPEAKERWe had a friend of the show, Wendell, from level one techs on on Tool Boot Diaries.

1:21:45SPEAKERTrust me, I listened to that episode.

1:21:47SPEAKERI love Wendell. I love Wendell.

1:21:49SPEAKERI have to speak to Wendell now because half the shit he was talking about

1:21:52SPEAKERis like extremely stuff I want to pick his brain about.

1:21:55SPEAKERYou have to talk to Wendell. Yeah.

1:21:58SPEAKERAnd one of the things he was talking about was that

1:22:01SPEAKERbecause, you know, he does a lot of consulting and stuff for corporations.

1:22:06SPEAKERThat's kind of his main stuff.

1:22:08SPEAKERThe YouTube, the level one stuff is just the fun side stuff, which is crazy.

1:22:13SPEAKERBut he was saying that actually because of all this stuff

1:22:16SPEAKERthat's going on with with Windows and Microsoft kind of,

1:22:20SPEAKERyou know, hitting the rake really hard, that

1:22:24SPEAKERstepping on the rake or what is the rake like stepping on the rug?

1:22:28SPEAKERYeah. And then it just.

1:22:29SPEAKEROh, OK. Yeah, sorry.

1:22:32SPEAKERThat that that actually, you know, it's not so much the year of Linux,

1:22:36SPEAKERwhich I mean, we are seeing some some play on that, obviously.

1:22:40SPEAKERBut he's seeing a lot of people move over to Mac OS that secretly this is actually

1:22:45SPEAKERthis could be the year of Mac OS, because that that's a better

1:22:49SPEAKERor that's a safer place for people to land because they probably

1:22:53SPEAKERalready have an iPhone.

1:22:54SPEAKERSo they're like, oh, I hate this Windows stuff.

1:22:57SPEAKERI'm just going to go over to Mac.

1:22:58SPEAKERSo what I'm saying is that we need to make a Mac podcast, apparently.

1:23:02SPEAKEROh, yeah. I use I use and quite enjoy a MacBook Pro.

1:23:07SPEAKERMac OS. I don't want to go back.

1:23:09SPEAKERIt's for professionals.

1:23:10SPEAKERI don't want to go back. Mac OS is an excellent offer.

1:23:12SPEAKERAnd my Mac OS is more Unix than Linux is.

1:23:14SPEAKERYou're more low. OK, well, now I'm in. Yeah.

1:23:17SPEAKERHave you ever tried to load Linux on a Mac?

1:23:20SPEAKERSomebody else? Yes. Yeah, I've got I've got Fedora running on my old.

1:23:23SPEAKERI have a very old 2011 MacBook Air that's running Fedora just fine.

1:23:27SPEAKERMy girlfriend uses for light tasks that are pretty good.

1:23:31SPEAKERWe should have her on the Dolby Diaries.

1:23:33SPEAKERMy Mac is a it's an M1, which compatibility

1:23:37SPEAKERwith with the ARM Macs is a little bit more limited right now.

1:23:40SPEAKERLike there's a Sahi Linux, which is a good chunk of the way there.

1:23:43SPEAKERBut there's still quite a bit of hardware

1:23:45SPEAKERthat doesn't work with ARM Macs and Linux yet.

1:23:48SPEAKERBut again, I mean, whatever we're getting way off topic here and you have to go.

1:23:51SPEAKERBut like Mac OS has enough of the common POSIX Unix like stuff in it.

1:23:56SPEAKERBut I don't need Linux on a MacBook like Mac OS is good for that.

1:24:00SPEAKERIt does all the same shit.

1:24:01SPEAKERDolby Diaries season three, baby. That's right. That's right.

1:24:04SPEAKERWe'll talk about the Mac.

1:24:06SPEAKERLet's talk about my Mac OS package manager.

1:24:09SPEAKERHaiku year two, Mac OS year three.

1:24:14SPEAKERBSD year four, year four.

1:24:16SPEAKERAnd then compile your own kernel year five.

1:24:19SPEAKERLike getting to roll your own Linux year five. Yeah, perfect.

1:24:23SPEAKERGreat. Thanks for coming by, Adam.

1:24:25SPEAKERAlways a pleasure to see you.

1:24:26SPEAKERIt's nice to finally get to do a podcast with you.

1:24:28SPEAKERI've been a big fan for a long time and really appreciate your time

1:24:32SPEAKERover there, PC World and The Full Nerd.

1:24:35SPEAKERFor folks who don't know, because I'm bad at telling them

1:24:37SPEAKERwhere can they find you and your work?

1:24:40SPEAKERYeah, so I am the head of video at PC World and The Full Nerd.

1:24:45SPEAKERSo go over to PC World's YouTube channel.

1:24:47SPEAKERGo over to The Full Nerd Network, which is a series of podcasts.

1:24:52SPEAKERWe have The Full Nerd podcast where we talk about

1:24:55SPEAKERDIY desktop computing and all that fun stuff.

1:24:58SPEAKERI have a podcast with Will talking about Linux called The Dual Boot Diaries.

1:25:02SPEAKERI have a third podcast coming up talking about tiny computing platforms.

1:25:08SPEAKERSo handhelds, mini PCs, laptops, retro devices, things like that.

1:25:14SPEAKERSo look forward to that.

1:25:16SPEAKERIs the retro stuff that we did with

1:25:19SPEAKERI don't want to say who, but is that going on Full Nerd or PC World?

1:25:22SPEAKERThat's going on PC World, Retro Roadshow.

1:25:25SPEAKERYeah, Retro Roadshow brought a TI 99 4A and a B-Box,

1:25:29SPEAKERthe PowerPC B-Box, BOS box buy, so we could get hands on with those.

1:25:35SPEAKERAnd there's videos of both of those coming shortly.

1:25:37SPEAKERI just watched the newest cut today.

1:25:40SPEAKERIs it good? Yeah, they turn it out real good.

1:25:42SPEAKERI think people are really going to like these historical deep dives.

1:25:45SPEAKERI really got to nerd out on stuff I'd never.

1:25:47SPEAKERWell, one of those machines I spent a lot of time with, one of them

1:25:49SPEAKERI spent no time at all with and was really excited about.

1:25:51SPEAKERSo but it started your career, started my career.

1:25:53SPEAKERYeah, exactly. Look forward to that.

1:25:55SPEAKERAnyway, thanks. Thanks, Adam.

1:25:56SPEAKERThanks for coming by.

1:25:57SPEAKERAnd we'll see you see you next week on the Full Nerd.

1:26:00SPEAKERBye. Next time I'll have a beard.

1:26:02SPEAKERHopefully, I hope so.

1:26:03SPEAKERThere's no video here, so I don't know if you have a beard or not.

1:26:06SPEAKERYeah. Thanks, dude.

1:26:10SPEAKERThanks for coming by, Adam.

1:26:12SPEAKERPlease check out the Full Nerd and Dual Boot Diaries.

1:26:14SPEAKERThey're both, you know, I think they're good podcasts.

1:26:17SPEAKERI'm on both of them, but but definitely go check out

1:26:21SPEAKERthe stuff on the PC World YouTube channel if you like computer stuff.

1:26:24SPEAKERI listened to DBD Weekly at this point, it seems.

1:26:28SPEAKERIt's I feel like for so many years, the rule was, oh,

1:26:31SPEAKERyou don't have to listen to your friend's podcast.

1:26:33SPEAKERAnd then I never do.

1:26:35SPEAKERYeah, I almost never do.

1:26:36SPEAKERLike I'm trying to think like

1:26:37SPEAKERI listen to the Andor run of a more civilized age earlier this year.

1:26:41SPEAKERYeah. And now I'm listening to this and that might be it.

1:26:45SPEAKERBut apologies to all of my other friends who do podcasts.

1:26:48SPEAKERBut I think those are the only friend casts I have ever listened to.

1:26:52SPEAKERI listen to more friend casts when I'm in the car a lot,

1:26:54SPEAKERwhen I have a lot of commuting to do or when I'm driving to L.A. or something.

1:26:58SPEAKERBut but yeah, for the most part, it's it's AMCA.

1:27:02SPEAKERAnd then a couple of tech podcasts.

1:27:03SPEAKERBut if you if you like the show and you want to support the show,

1:27:07SPEAKERwe are listening to support it.

1:27:08SPEAKERWe don't take money for ads. We don't take money other places.

1:27:12SPEAKERWe only take money from you, the listener, because that's who we want to work for.

1:27:15SPEAKERAnd you can do that.

1:27:16SPEAKERYou can give us money by going to patreon dot com slash tech pod

1:27:20SPEAKERwhere five bucks will buy you access to the Discord,

1:27:24SPEAKERwhich is full of nerds that talk about stuff like we talk about on the tech pod

1:27:27SPEAKERall day, as well as the monthly patron exclusive episode where this month

1:27:33SPEAKERI feel like we recorded that yesterday

1:27:35SPEAKERand we just kind of talked about like Halloween decorations and other like

1:27:40SPEAKERwe were kind of all over the place, but it was a fun, rambly conversation.

1:27:43SPEAKERFixing inflatable decoration.

1:27:45SPEAKERYeah, probably some Linux in there.

1:27:48SPEAKERHomeschooling. Yeah. Yeah.

1:27:51SPEAKERWorking on your house. Yeah.

1:27:53SPEAKERHouse projects. Yeah, that's probably up by now.

1:27:56SPEAKEROh, yeah. Yeah, you can get that now.

1:27:57SPEAKERIf you if you suddenly get all of them, there's like 70 of them.

1:28:00SPEAKERIf you want more tech fun.

1:28:01SPEAKERYeah. So that'll do it for us.

1:28:04SPEAKERAs always, thanks to everybody who subscribes at patreon dot com slash check.

1:28:08SPEAKERBut a very special thank you to our executive producer, dear patrons,

1:28:11SPEAKERincluding Jason Lee and Felicitous Rips.

1:28:14SPEAKERAndrew Slosky, Jordan Lippett, celebrating 40 years of bunny crimes.

1:28:19SPEAKERCongratulations. Bunny crimes.

1:28:20SPEAKERTwinkle Twinkie, David Allen, James Cammack and Pantheon

1:28:23SPEAKERmakers of the HS three high speed 3D printer.

1:28:25SPEAKERThank you all so, so much. Thank you.

1:28:28SPEAKERAnd that'll do it for us this week.

1:28:29SPEAKERWe'll be back next week with another edition of the tech pod.

1:28:31SPEAKERUntil then, please consider the environment before printing this podcast,

1:28:35SPEAKERwhich we also talked about last week.

1:28:41SPEAKER[BLANK_AUDIO]