I've recently heard of a concept called Open Source Hardware. From what I understand it is about making Open source CPUs and GPUs (the designs) and maybe also Motherboards. Does anybody know anything about this? Does this have potential? That is the performance when compared to other (commercial) CPUs from Intel/AMD? GPUs (ATI/ NVidea)? Are these compatible with other products (not motherboards but more like software... Windows, X86, are they 32-bit for CPUs and things like Open GL, prob not DirectX, and modern games)? Do these things mean that anyone can just go to Fabs (like Fabless companies, example: Agiea, ATi, Nvidea...) and get them made without royalties? This is interesting as it could help my company, but I wanted to see what it really is!
Open Source Hardware
- Thread starter Ggamer
- Start date
Well... I wouldn't expect cutting-edge CPU or GPU performance, but a stripped-down CPU that just did 32-bit x86 (without extensions like MMX or SSE) would be a lot easier to design. Or a RISC CPU that ran the POWER instruction set or something like that. If you step back a few generations in terms of hardware (like a chip built on a 110nm or 130nm process at a lower clock speed), designs get a lot easier.
Prices would be high for small runs, but much less for larger ones. You also might be able to implement a simple enough CPU design on something like a Xilinx FPGA. However, I doubt it would be worth it for a general-purpose CPU; there are many good, cheap ones on the market today. But if you wanted to build some sort of signal-processing board or specialized add-in processor (like an encryption or compression/decompression engine), however...
If you had a totally open-source design, yes, you could in theory take it to a fabricator and have them built. However, you would likely need a pretty big order (thousands, if not tens of thousands of chips) for a major company to take it on, and the startup costs would be very high (probably millions of dollars up front to have the lithography done). Unless you have some compelling reason to not use an existing CPU, and you can't use an FPGA (due to cost or speed constraints), it is unlikely to be worth it for most projects.
Prices would be high for small runs, but much less for larger ones. You also might be able to implement a simple enough CPU design on something like a Xilinx FPGA. However, I doubt it would be worth it for a general-purpose CPU; there are many good, cheap ones on the market today. But if you wanted to build some sort of signal-processing board or specialized add-in processor (like an encryption or compression/decompression engine), however...
If you had a totally open-source design, yes, you could in theory take it to a fabricator and have them built. However, you would likely need a pretty big order (thousands, if not tens of thousands of chips) for a major company to take it on, and the startup costs would be very high (probably millions of dollars up front to have the lithography done). Unless you have some compelling reason to not use an existing CPU, and you can't use an FPGA (due to cost or speed constraints), it is unlikely to be worth it for most projects.
So there is no way any of these designs are anywhere close to modern (even Pentium 4 level) performance? I guess I got my hopes up to high... Open Source software has reached such levels I thought hardware might also be that good.
I actually have no idea what may or may not be available out there.
But no, I wouldn't expect anything close to what latest-gen CPU/GPUs can do. AMD and Intel spend hundreds of millions every year on cutting-edge R&D. Anyone with a computer, some books, and a compiler can do software development; hardware ASIC development requires in-depth knowledge of electrical engineering, and generally pricey simulation packages.
Look for the "smaller" things first. I know video is being worked on currently (er, at least last time I checked...). And I'm sure it won't be "top of the line," but it'll be something useful for the rest of us.
Open Source software can be compiled on a $200 PC.
Open Source hardware would require a $200,000,000 facility.
Small difference in scale.
I see about as much future for open source hardware as open source supertankers. You can get the design done for free using a collaborative model, but someone has to build the thing eventually, and that takes capital.
OK, well there's a bit more potential than for open source ships, but it's only really viable for items that have very low manufacturing costs. Things that can reasonably made by a DIYer or such. Anyone who tries to sell such a product is going to have a lot of possible liability for defective product lawsuits that he is going to have to hire his own engineers to double check the open-source design anyhow.
That's how I see it.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter