Dan Sugalski <[EMAIL PROTECTED]> wrote:
> I was using .locals for the actual variables in the source program,
Well, you know it: .locals aren't vars.
> and $Px for all the temps the compiler generated. I've been migrating
> a lot of the code to use a few .local-defined hashes and indexing
> into
At 6:04 PM +0200 4/22/04, Leopold Toetsch wrote:
Dan Sugalski <[EMAIL PROTECTED]> wrote:
registers needed:I43, N0, S12, P3327
registers in .pasm: I25, N0, S20, P32 - 464 spilled
2007 basic_blocks, 2079 edges
Ouch. Register allocation is spending huge times
Dan Sugalski <[EMAIL PROTECTED]> wrote:
> registers needed:I43, N0, S12, P3327
> registers in .pasm: I25, N0, S20, P32 - 464 spilled
> 2007 basic_blocks, 2079 edges
Ouch. Register allocation is spending huge times during spilling.
Something is definetely wro
At 4:03 PM +0200 4/22/04, Leopold Toetsch wrote:
Dan Sugalski <[EMAIL PROTECTED]> wrote:
Dunno what parrot thinks--it's not done yet. grep says 569 .locals
and 473 temp PMC registers.
I've now enabled some more code that speeds up temp allocation more
(~2.5 for 2000 temps in a unit). This needs
A good register allocation algorithm will always have problems with
big subs, there is nothing that we can do about it.
I think that what "real compilers" do to solve this problem is
implement two different algorithms: one for normal subs which tries
to generate optimal code, and a naive one f
Dan Sugalski <[EMAIL PROTECTED]> wrote:
> Dunno what parrot thinks--it's not done yet. grep says 569 .locals
> and 473 temp PMC registers.
I've now enabled some more code that speeds up temp allocation more
(~2.5 for 2000 temps in a unit). This needs that the usage range is
limited to 10 lines. I
Dan Sugalski wrote:
As I sit here and wait for parrot to churn on the output from compiling
a relatively small program
I've put in another factor ~2.5 change for a unit with 2000 temps.
leo
At 7:55 AM +0200 4/22/04, Leopold Toetsch wrote:
Dan Sugalski <[EMAIL PROTECTED]> wrote:
I think it may be a handy thing if someone'd go through and draw up a
set of rules for the use of temps, and things that'll cause the
register coloring algorithm to go mad. (I'd like to avoid 30 minute
com
At 1:22 PM +0200 4/22/04, Leopold Toetsch wrote:
Dan Sugalski wrote:
... (I'd like to avoid 30 minute compile sessions--it's a tad tedious :)
Should be faster now by some factor.
Cool, thanks. I've an optimized build of parrot going now, and I'll
see what things look like when it's dine.
How many
Dan Sugalski wrote:
... (I'd like to avoid 30 minute compile
sessions--it's a tad tedious :)
Should be faster now by some factor.
How many symbols are in the biggest compilation unit (parrot -v
"registers in .imc")?
leo
Dan Sugalski <[EMAIL PROTECTED]> wrote:
> I think it may be a handy thing if someone'd go through and draw up a
> set of rules for the use of temps, and things that'll cause the
> register coloring algorithm to go mad. (I'd like to avoid 30 minute
> compile sessions--it's a tad tedious :)
Are you
11 matches
Mail list logo
