4 ms·
John Carmack imagines a future where GPUs don't need PCs
- galleywest200 2y ago> Imagine connecting a graphics card to a display and a power source and running diagnostics before installing it or running it standalone later when troubleshooting. Perhaps the card could have a compact Linux distro onboard, mused the iconic Doom developer. So basically a Raspberry Pi glued to an eGPU?
- tlhunter 2y agoYeah, sans Raspberry Pi ;)
- geor9e 2y agoYou're quoting a blogger's interpretation - here's the tweet quote - "put a tiny linux system running busybox on your command processor". Command processor is the part of a GPU that schedules the hundreds of smaller cores.
- anotherhue 2y agoThe Pi (at least the earlier versions I know) actually is mostly a GPU with a cpu bolted on.
- umvi 2y agoCan we change title to say "John Carmack"? Otherwise it's click bait
- LorenDB 2y agoDone!
- theevilsharpie 2y agoBeing able to run some graphics diagnostics without a host CPU would be useful. However, if we're talking about putting a general-purpose processor/RAM onto a GPU that has enough power to run Linux, at that point, you've essentially created a game console. May as well go all the way and add a display and battery too, for portability. I'm sure that would be a hit! /s
- VTimofeenko 2y agoTangentially related, but GPUs with embedded m2 drives mentioned in the article (IIUC they effectively pass m2 drives through to the system by using unused PCI lanes) are a very interesting idea for a low volume transcoding machine.
- jsheard 2y agoIt's a clever idea but unfortunately limited by the fact that it requires the motherboard to support PCIe slot bifurcation, which isn't very common on consumer boards. I assume the target market is OEM integrators who can go out of their way to pair it with a suitable motherboard.
- undersuit 2y agoMy dirt cheap Asrock A320m-ITX and my mid-level Gigabyte Aorus Elite x570 support bifurcation, did the feature become much less common in the last couple years?
- jsheard 2y agoI think the situation is generally better on AMD platforms than Intel ones, but still YMMV. https://forum.level1techs.com/t/any-lga1700-motherboards-with-x4x4x4x4-bifurcation/186037 https://forum.level1techs.com/t/any-lga1700-motherboards-wit... Some boards will only bifurcate the x16 slots into x8/x8 mode, which is sufficient for an x8 GPU with one SSD attached to the spare lanes, but that still leaves 4 lanes unutilized. Adding two SSDs to fill up the bus requires support for x8/x4/x4 mode, and an x16 carrier for four SSDs requires support for x4/x4/x4/x4 mode.
- undersuit 2y agoWell the cheaper one does lack the 4x/4x/4x/4x mode in the bios, wonder how much of that is a hardware limitation.
- boulos 2y agoI mean Larrabee did already back in 2008-ish, but with BSD. That was a real advantage of the reuse of some x86 cores. On the flip side, until somewhat recently, NVIDIA GPUs didn't have the kind of hardware protections you'd want to build a modern multitenant OS (e.g., hardware page table and TLB support). In the actual tweet (https://x.com/ID_AA_Carmack/status/1865127952712905197 https://x.com/ID_AA_Carmack/status/1865127952712905197) he mostly seems to be talking about RLO style diagnostics with busybox on the device much like Dell servers with fancy BMCs: > I still think today’s GPUs should be able to operate without host CPUs if they have a private link. Chains of accelerators are a legitimate use, but it would just be fun if GPUs made their own video signal with diagnostic information when you apply power outside of a host system. You could go farther and put a tiny linux system running busybox on your command processor, and backchannel keyboard input through the display port if you don’t have a USB port.
- mepian 2y agoThis is blogspam, just submit his tweet directly instead.
- TheAceOfHearts 2y agoHonestly, what an absolutely minimal effort blog post. It would've been so much cooler if the author actually interviewed John and fleshed out or explored the idea a bit more.
- nox101 2y agoGPUs are slow for single threaded things. Often 20x+ slower than your CPU. They get all of their speed by massive parallelism and some specialized hardware (like samplers). If your problem can't be easily parallelized then they aren't a solution.
- nox101 2y agoWould love to know why this was downvoted. It's factually true and easily provable.