Rust wgpu now has a
-
@amarioguy @MaddieM4 i am far from the only person to think about this. someone will do it eventually. the question is when we hit one that has riscv levels of Go Juice.
@mcc @amarioguy @MaddieM4 there's also a thing where...
our friends who do CPU stuff don't seem to think riscv is going to be outside of the systems of oppression that cause trouble, and we do for sure see how stuff like core layout and fabrication are bottlenecks that allow hegemony to form...
by all means a GPU that's as open as riscv is would be awesome, but there's more to aspire to
-
@lispi314 @erincandescent one problem is i want to *create* games and those are by nature new. and that creates nonzero complications
@mcc @erincandescent It certainly does indeed, though it does also put you in the position to evaluate how hostile the engine or framework you intend to use is to older hardware.
(Granted depending on what you want them to do exactly, there are limits to that.)
Presumably given your no-slop & older computing habits, you'd be intent on something quite friendly.
That does leave you with the problem that the groundwork needed to get started needs to be done. A rather unfortunate setback. -
@mcc@mastodon.social Redox depending on implementation-defined Rust where the main tooling is being slopped if not Rust itself does seem like a problem.
Have or are they considering pinning the Rust dependency to a pre-slop version?
(Honestly the language needs a hard fork that solidifies on a proper specification and then goes specification-driven. These problems will keep happening while it hasn’t.)
@lispi314 I don't think they're discussing this. This is I believe a very big problem but not my concern at the current time because (1) you need a programming language; that's not optional (1.5) i no longer acknowledge the existence of C (2) there are already at least four Rust backends and at least two frontends (3) i am much more confident i can create an independent Rust implementation than I am confident that I can do literally *any* of the projects described in the thread above.
-
@lispi314 I don't think they're discussing this. This is I believe a very big problem but not my concern at the current time because (1) you need a programming language; that's not optional (1.5) i no longer acknowledge the existence of C (2) there are already at least four Rust backends and at least two frontends (3) i am much more confident i can create an independent Rust implementation than I am confident that I can do literally *any* of the projects described in the thread above.
@lispi314 I do think something is wrong with our framing of the disussion that people are perpetually concerned about slop in Rust but are not concerned about same in C when GCC and clang are both slopped and, in fact, LLVM is the main slop contributor to the Rust implementation at the current time. "But Rust?" is not really a useful objection unless you plan to write an OS in Zig. And as described above, we don't even have an OS in Rust yet.
-
@lispi314 I do think something is wrong with our framing of the disussion that people are perpetually concerned about slop in Rust but are not concerned about same in C when GCC and clang are both slopped and, in fact, LLVM is the main slop contributor to the Rust implementation at the current time. "But Rust?" is not really a useful objection unless you plan to write an OS in Zig. And as described above, we don't even have an OS in Rust yet.
@mcc@mastodon.social Oh GCC is fucked as well.
We (disgruntled GCC users) are talking about alternative C compilers, and for x86_64/i686 the situation is relatively not that bad if one doesn’t use a lot of GCC/CLANG extensions.
(Other hardware though? Not good.)
The problem has also reached some Common Lisp implementations. But I don’t need to use SBCL, when ECL, ABCL & others exist and will work with any compliant program written with portable libraries (it’s seen quite negatively to write implementation-specific libraries unless they’re explicitly named after and targeting a feature of a specific implementation, which will have people avoid it).
Zig is actively looking to move away from LLVM. Zig also being implementation-specified annoys me too. (It does for every language, for me.)
-
@mcc@mastodon.social Oh GCC is fucked as well.
We (disgruntled GCC users) are talking about alternative C compilers, and for x86_64/i686 the situation is relatively not that bad if one doesn’t use a lot of GCC/CLANG extensions.
(Other hardware though? Not good.)
The problem has also reached some Common Lisp implementations. But I don’t need to use SBCL, when ECL, ABCL & others exist and will work with any compliant program written with portable libraries (it’s seen quite negatively to write implementation-specific libraries unless they’re explicitly named after and targeting a feature of a specific implementation, which will have people avoid it).
Zig is actively looking to move away from LLVM. Zig also being implementation-specified annoys me too. (It does for every language, for me.)
@lispi314 well, in the year 2026 i'm willing to write an alternative rust compiler, but i'm not willing to write an alternative c compiler. i think there is no justification for using c in the year 2026
-
@lispi314 well, in the year 2026 i'm willing to write an alternative rust compiler, but i'm not willing to write an alternative c compiler. i think there is no justification for using c in the year 2026
@lispi314 C/C++ is a bad thing we accepted for a long time because there were no alternatives. the last ten years means we have alternatives. i count at least three
and yes, i know zig is moving away from LLVM, that's why i said they're a reasonable alternative
-
@lispi314 well, in the year 2026 i'm willing to write an alternative rust compiler, but i'm not willing to write an alternative c compiler. i think there is no justification for using c in the year 2026
i am much more confident i can create an independent Rust implementation
The thing is, without a specification to implement, you’ll have to stay bug-compatible with some probably-slopped reference implementation.
That’s the problem I see with it.
-
@mcc @amarioguy @MaddieM4 there's also a thing where...
our friends who do CPU stuff don't seem to think riscv is going to be outside of the systems of oppression that cause trouble, and we do for sure see how stuff like core layout and fabrication are bottlenecks that allow hegemony to form...
by all means a GPU that's as open as riscv is would be awesome, but there's more to aspire to
@ireneista @mcc @MaddieM4 tbf i suspect a lot of it is that a very large amount of the tooling needed to make remotely modern CPUs and chips and such (PDKs, etc) is under so much NDA'ed lock and key (seriously, the things i've heard about this are on the border of absurd...) that you basically need %BIG_COMPANY% resources to even enter the sphere....so effectively even if the ISA is open, it's big companies who are driving the standard effectively
-
@lispi314 C/C++ is a bad thing we accepted for a long time because there were no alternatives. the last ten years means we have alternatives. i count at least three
and yes, i know zig is moving away from LLVM, that's why i said they're a reasonable alternative
i count at least three
I’m actually interested in this. I’ve been looking for specification/standard-backed ones.
For high-level languages I mostly have Java and Common Lisp. But for low-level work I’m kind of stuck, as Ada not only doesn’t have a viable bootstrap compiler yet (nlnet project ongoing) but even ignoring that issue GNAT is GCC-based (ah, I really hope GNAT itself hasn’t been slopped either).
C has a proper specification but I’d rather not write C if I can avoid it.
For “scripting”,
shellisPOSIXspecified. Otherwise Scheme is an option, but portable libraries haven’t really been a thing in Scheme culture (recent language specification changes further enabling it may change this). -
@ireneista @mcc @MaddieM4 tbf i suspect a lot of it is that a very large amount of the tooling needed to make remotely modern CPUs and chips and such (PDKs, etc) is under so much NDA'ed lock and key (seriously, the things i've heard about this are on the border of absurd...) that you basically need %BIG_COMPANY% resources to even enter the sphere....so effectively even if the ISA is open, it's big companies who are driving the standard effectively
@amarioguy @ireneista @MaddieM4 RISCV already exists. RISCV is already successful. You're suggesting it's impossible to do something that was already successfully done eight years ago
-
i count at least three
I’m actually interested in this. I’ve been looking for specification/standard-backed ones.
For high-level languages I mostly have Java and Common Lisp. But for low-level work I’m kind of stuck, as Ada not only doesn’t have a viable bootstrap compiler yet (nlnet project ongoing) but even ignoring that issue GNAT is GCC-based (ah, I really hope GNAT itself hasn’t been slopped either).
C has a proper specification but I’d rather not write C if I can avoid it.
For “scripting”,
shellisPOSIXspecified. Otherwise Scheme is an option, but portable libraries haven’t really been a thing in Scheme culture (recent language specification changes further enabling it may change this).@lispi314 rust, zig, go. go isn't well suited for an operating system. but i think i could hack it until it was
-
@mcc @sabrina ...I am not proud of how little progress I've made, as it's not been done at all efficiently. (Rewriting WINGs in Rust while mostly maintaining source-level compatibility with the original C library is probably kinda dumb. But I am learning.)
Most recent status update is that I'm un-mothballing it again and making sense of a backlog of local changes so they can all get pushed to a project repo. You're welcome to look:
https://git.sdf.org/vitrine/wmaker
https://git.sdf.org/trurl/wmaker -
Okay so now what I'm looking at is Redox. Redox is the OS project I find that's interested in being written by humans instead of people who have decided to replace parts of their brain with a corporation. I can live with this. I've got an update on my Redox thread I haven't posted yet, but for these purposes what's important is Redox doesn't have, and considers a far-future goal, video acceleration. That makes sense. But it means there's a gap in the staircase I wish to climb.
-
Rust wgpu now has a .agents/skills, a .claude/skills, an AGENTS.md, and a CLAUDE.md.
Increasingly wondering if my goal of living in a way that I never benefit from the actions of generative AI users and generative AI users never benefit from the actions of me is going to inherently mean never using hardware-accelerated video on a computer again.
@mcc I ended up down the 'cpu bound' route, see you in 100 hours. https://digipres.club/@discatte/117318307216404824
-
@amarioguy @ireneista @MaddieM4 RISCV already exists. RISCV is already successful. You're suggesting it's impossible to do something that was already successfully done eight years ago
@mcc @ireneista @MaddieM4 right, but an open ISA is not the same thing as being able to create silicon with said ISA, for a truly open chip, we do need people who can fabricate said silicon, is more my point
also RISC-V is evolving, usually according to how big companies want it, it's not like RISC-V stood still and just froze, there's a lot of funky stuff being added to it (M-mode being my usual example, it reeks of "intel SMM" levels of mistakes personally)
-
@lispi314 rust, zig, go. go isn't well suited for an operating system. but i think i could hack it until it was
@mcc@mastodon.social Of those, Golang does have a specification target, too.
-
This thread is probably meandering and maybe solipsistic. My point is this: My weird background means I can look at this staircase full of missing stairs, identify one step, and go, "I can fix that". But I don't know if there's enough of me to fix multiple steps. I haven't decided what to do from here. Pick a step and fix it anyway? Pick a "fun" project, like the GPU [which face it, even if I finished it chances are not high more than 3 people would ever use it]? Give up and learn electric bass?
@mcc i have this intermediate representation layer that's kind of webassembly-ish called bismuth, which i can then create high level languages for that target that IR. and i want to make it so that its easy for people to run their own implementation of a bismuth interpreter
then i want to make an OS with it. i think i can do all of this, but i can't make drivers for hardware, so even if slop is barred from the bismuth ecosystem it'll still have to rely on a probably-slop host operating system
-
@mcc i have this intermediate representation layer that's kind of webassembly-ish called bismuth, which i can then create high level languages for that target that IR. and i want to make it so that its easy for people to run their own implementation of a bismuth interpreter
then i want to make an OS with it. i think i can do all of this, but i can't make drivers for hardware, so even if slop is barred from the bismuth ecosystem it'll still have to rely on a probably-slop host operating system
@mcc although freedos is, i think, recently going towards a slop-free AI policy? and freedos can run on bare metal with sound on at least some systems. so maybe that's a possibility
not sure if it's helpful to relate all this to you. i figure you might be glad to know you're not alone trying to construct part of a slop free ecosystem while also trying to balance the limitations of being only one person
-
@mcc @ireneista @MaddieM4 right, but an open ISA is not the same thing as being able to create silicon with said ISA, for a truly open chip, we do need people who can fabricate said silicon, is more my point
also RISC-V is evolving, usually according to how big companies want it, it's not like RISC-V stood still and just froze, there's a lot of funky stuff being added to it (M-mode being my usual example, it reeks of "intel SMM" levels of mistakes personally)
@amarioguy @mcc @MaddieM4 one hopeful note
in our experience when a company is very big on having NDAs around its tooling, it's because there's nothing hard to replicate about the tooling itself, once you know the basic idea of it
we suspect people who do chip design can already basically say, yeah you need these specific types of simulator to eliminate whatever specific classes of problem
that seems like a thing the community could take on, in principle
