Timeline for answer to Is there a theoretical possibility of having a full computer on a silicon wafer instead of a motherboard? by GNA
Current License: CC BY-SA 4.0
Post Revisions
29 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 14, 2021 at 23:42 | comment | added | user0193 | @GNA awsome! Thanks a lot! | |
| Jan 14, 2021 at 23:25 | comment | added | GNA | @JhonnyS Here it is. I found it. I've set hte timestamp to the moment you can see the T&H-Chips. youtu.be/U3w_EWgGQuk?t=1513 However, my memory served me a little wrong. it is not molded. But an interesting package anyway. | |
| Jan 14, 2021 at 23:11 | comment | added | user0193 | @GNA if you ever find the youtube video on the high frequency oscilloscope you were talking about please do care to share in comments. | |
| Jan 5, 2021 at 16:05 | comment | added | Yakk | So, to be fair, the idea of an entire CPU (or even more than a transistor) on one piece of silicone was once practically impossible as well. Almost everything you mention is engineering problems, and not fundamental ones; the advantage of fewer manufacturing steps may eventually lead towards this. Just wire a chip up a battery and some USB ports and go. | |
| S Jan 5, 2021 at 13:44 | history | suggested | Robert Harvey | CC BY-SA 4.0 |
Spelling, grammatical and stylistic fixes.
|
| Jan 5, 2021 at 13:01 | review | Suggested edits | |||
| S Jan 5, 2021 at 13:44 | |||||
| Jan 5, 2021 at 8:25 | comment | added | mathrick | It doesn't invalidate your answer in any way, but I thought you might find this interesting: cerebras.net. Disclaimer: this is my employer, and the product is, in fact, a single chip per wafer :) Not a computer, mind you, just a chip, which needs a lot of support from surrounding components that are almost as full of interesting engineering challenges as the chip itself. But it's manufactured using current fabrication technology. | |
| S Jan 5, 2021 at 8:23 | history | suggested | Jonathan Johansen | CC BY-SA 4.0 |
Spelling, full stops and two wikipedia links.
|
| Jan 5, 2021 at 7:44 | comment | added | bandybabboon | superuser.com/questions/277655/… | |
| Jan 4, 2021 at 23:12 | review | Suggested edits | |||
| S Jan 5, 2021 at 8:23 | |||||
| Jan 4, 2021 at 19:13 | comment | added | GNA | @PeterCordes There are power modules in a single mold package. These have exactly what you talk about. a power inductor molded into a single black blob with the control asic and decoupling caps. Intel's Enpirion pwer series for example. Other manufacturers like TI and LT/AD also build these modules. Additionally, I remember a tear-down video on youtube of a high frequency oscilloscope. The track-and-hold Chip hat a coax-connector somehow molded into it to allow a very direct connection from the input to the ASIC. However, I don't find the youtube video right now. | |
| Jan 4, 2021 at 19:03 | comment | added | GNA | @Shamtam: Yes. I also used to develop for an GaAs HEMT technology. The foundry specified a yield of like ~15%. And those were only the chips that didn't instantly break down. Those with working high frequency parts were even less. It was guite a pain... | |
| Jan 4, 2021 at 6:40 | comment | added | Peter Cordes | You could maybe imagine putting some inductors inside the same plastic package that houses the SoC. If it was going to be big enough to have an AC wall plug, or a 20VDC plug for an external power supply, and USB-C (and maybe HDMI) for keyboard and video anyway, it would have some room inside the package. Or the required external HW would include some external inductors and maybe capacitors, although I think the OP is trying to avoid a mobo entirely? Anyway, just thought it was amusing to picture a modern CPU with it's >7 layers of whisker thin metal wires, and also a power supply inductor :P | |
| Jan 4, 2021 at 3:55 | comment | added | Shamtam | Great answer. I also work in semiconductor, and it always makes me laugh how "20 chips at most" loss is reasonable in silicon, whereas in III-V "only 20 chips yielded over the entire wafer" is not unheard of 🤣 | |
| Jan 3, 2021 at 23:35 | history | edited | GNA | CC BY-SA 4.0 |
deleted 2 characters in body
|
| Jan 3, 2021 at 23:32 | comment | added | chux | "let's say hour headphone" --> "our"? | |
| Jan 3, 2021 at 22:32 | comment | added | GNA | Of course we can start comparing modern emiconductor technology with the capabilities of the '80s. But that's not very useful. Maybe in future it will be possible to integrate all of a now "modern" PC in a single chip. But current performance cannot be achieved by current technologies using a single chip. Small joke: Even old '70s tech can't be implemented on a single chip. Where do you put in your punch-cards? | |
| Jan 3, 2021 at 22:25 | comment | added | GNA | All x86 computers that I know of have DC/DC regulators for the core voltage. These core voltage regulators are on PC mainboards for ages. Saying they're "fairly recent" is not very accurate. They've been the standard for more than two decades. I have an old pentium 4 mainboard laying around from the year 2000. And it uses DC/DC converters for the core supply. Even the pentium 4 CPU required almost 100 Amps on its core voltage. intel.com/content/dam/support/us/en/documents/processors/… Table 9 page 22. ~1V with ~100A without a DC/DC converted would be crazy. | |
| Jan 3, 2021 at 22:16 | comment | added | Chris Stratton | Plenty of computers don't have a DC/DC converter. They're a fairly recent idea. Historically the power supply has been entirely distinct from the motherboard, to which it provided the precise voltages needed. You started to get on-board regulators when logic voltage standards diverged, when core voltages dropped, and especially when you got to core which needed different supply voltages when running at different speeds. Plenty of "computers" exceeding legacy PCs still have none of these, or in fact accomplish it with on-chip linear regulation. | |
| Jan 3, 2021 at 22:14 | comment | added | user0193 |
@ChrisStratton yes, I do agree but having read DC/DC regulators cannot be implemented on-chip, Manufacturing a headphone amp in a 7 nm fin-fet technology is not ideal, Inductors on Chips are quite a pain. They're usually only done for > 1 GHz transformers for signals coupling plus other concrete cases such as CMOS technology etc. gave very strong intuition to me.
|
|
| Jan 3, 2021 at 22:04 | history | edited | GNA | CC BY-SA 4.0 |
added 7 characters in body
|
| Jan 3, 2021 at 22:02 | comment | added | GNA | Let's just clarify: TSMC is a manufacturer. | |
| Jan 3, 2021 at 22:02 | comment | added | Chris Stratton | @JhonnyS indeed, for something on the scale of today's full-"computers" that is true (though it is not for smaller systems well exceeding the functionality of the PC's of the past) - but many many people already explained that quite earlier in the lifecycle of this question. | |
| Jan 3, 2021 at 22:00 | vote | accept | user0193 | ||
| Jan 3, 2021 at 21:59 | comment | added | user0193 | I am accepting this as a formal answer and I think this is most direct answer to what I was looking! Putting everthing in single wafer is not possible mainly due to the difference in manufacturing process like CMOS and TSMC for various components! | |
| Jan 3, 2021 at 21:54 | comment | added | GNA | No. Because this allows you to manufacture these circuits in different technologies. It also improves your yield. If you take the 7 or 5 nm TSMC technology AMD uses to manufacture its high end CPUs, you wouldn't be able to manufacture a decent Ethernet-Chip or an audio chipset with this technology. | |
| Jan 3, 2021 at 21:49 | history | edited | GNA | CC BY-SA 4.0 |
added 435 characters in body
|
| Jan 3, 2021 at 21:49 | comment | added | user0193 |
Can you please clarify It is still better to combine multiple dedicated chips than trying to put everything on one die, isnt combining same as putting them in same die?
|
|
| Jan 3, 2021 at 21:39 | history | answered | GNA | CC BY-SA 4.0 |