@bmastenbrook to the extent that mainstreamish architectures are similar to each other, assembly for *any of them* is a portable assembly
-
-
Replying to @whitequark
@whitequark I *think* people looking for a "portable assembly" want something with straightforward translation, but not to worry about ...1 reply 0 retweets 0 likes -
Replying to @bmastenbrook
@whitequark ... things like branch delay slots that something else can automatically insert. but I could be wrong2 replies 0 retweets 0 likes -
Replying to @bmastenbrook
@bmastenbrook uhm. delay slot stuffing is like the basic job of a risc assembler2 replies 0 retweets 0 likes -
Replying to @whitequark
@bmastenbrook i'm going to go further and say that the fact our assemblers don't do regalloc or cc handling just means they suck2 replies 1 retweet 2 likes -
-
Replying to @johnregehr
@johnregehr@bmastenbrook yes. LLVM is patterned for and after mainstreamish architectures, so it's just as valid here2 replies 0 retweets 0 likes -
Replying to @whitequark
@johnregehr@bmastenbrook ... well, somewhat less valid, because LLVM IR has UB in its semantics. maybe if you don't use opt2 replies 0 retweets 1 like -
Replying to @whitequark
@whitequark@johnregehr@bmastenbrook note: a lot of what people still need *actual* assembly for is clever use of SIMD instructions2 replies 0 retweets 0 likes -
Replying to @elwoz
@elwoz@whitequark@johnregehr@bmastenbrook The only thing ppl _actually_ need asm for is stuff _not representable_ in C.3 replies 0 retweets 0 likes
@elwoz @whitequark @johnregehr @bmastenbrook i.e. things like task/context switch, thread creation, sj/lj impl, code that runs w/o stack,...
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.