[bitc-dev] Reviewing the state of BitC
hramrach at centrum.cz
Wed Mar 10 14:47:35 PST 2010
On 10 March 2010 23:33, Philipp Klaus Krause <pkk at spth.de> wrote:
> Jonathan S. Shapiro schrieb:
>> On Wed, Mar 10, 2010 at 2:09 PM, Philipp Klaus Krause <pkk at spth.de> wrote:
>>>> The problem with compiling to C is that it isn't simple to make things
>>>> work nicely with a garbage collector, and all of the techniques that
>>>> do so impose a significant performance penalty.
>>> So it is decided that bitc shall no be usable for embedded systems?
>> I certainly didn't say that. As a concrete example, I still expect
>> that we would use BitC in the next round of Coyotos implementation.
>> Compiling to C has nothing to do with embedded system use. Indeed,
>> compiling direct to assembler is far better than compiling to C.
> And what do you mean by "compiling to assembler"? I suppose that's
> compiling to x86, and there's lots of systems that are not x86, in the
> embedded world there even are a lot of architectures in use not targeted
> by gcc.
> C is today's (only) portable assembler.
I don't know where you came with the notion that assembly is x86.
Assembly is symbolically named machine instructions and it indeed
differs per machine, and many popular compilers including gcc generate
an architecture specific assembly which is then converted to the
More information about the bitc-dev