Hi Alex,
java wlErsatzLisp Pil32Pil64
+---
(fibo 22) | 250.19 0.0150.016
(fibo 23) | 450.25 0.0260.024
(fibo 24) | 690.36 0.0410.039
(fibo
Hi Tomas,
java wlErsatzLisp Pil32Pil64
+---
(fibo 22) | 250.19 0.0150.016
(fibo 23) | 450.25 0.0260.024
(fibo 24) | 690.36 0.0410.039
Hi Edwin,
by the way, thank you for releasing this in the wild. your java
implementation gave me a better understanding how picolisp works under
the hood.
you are welcome;-)
BTW, thanks Alex for mentioning the build problem. I fixed it now so
you can git pull again.
Cheers,
Tomas
--
Hi Tomas,
On Thu, Oct 14, 2010 at 1:43 AM, Tomas Hlavaty t...@logand.com wrote:
In that email, I just wanted to say that I implemented some functions
directly in Java and that speeds up the fibo benchmark significantly.
For getting better picture about interpretation overhead, there are now
If you get the latest wl code (git pull), you'll see even a pure java
'fibo.' function for comparison and getting an idea of the overhead
caused by interpretation.
Oops, forgot to push. Now in the repo.
Cheers,
Tomas
--
UNSUBSCRIBE: mailto:picol...@software-lab.de?subject=unsubscribe
Hi Tomas,
If you get the latest wl code (git pull), you'll see even a pure java
'fibo.' function for comparison and getting an idea of the overhead
caused by interpretation.
Yeah, I know that very well. Did a lot of 'fibo' tests during the last
30 years ;-) Here, locally, I see that a
Hi all,
oops oops! No problem with the mail archive. I sent it to myself:-(
Sorry for the confusion, must have been too tired.
However, what disturbs me more: Where does the above quote come from?
I didn't get that mail. Anybody? Did you really send it? It seems
also not to be in the mail
On 10/06/2010 10:04 PM, Tomas Hlavaty wrote:
Hi Alex,
Perhaps 'fork' can be simulated by copying all runtime variables to a
thread?
not sure whether that's enough. The semantics is completely different,
what about open files and so on? Is it necessary to go that far in
terms of
Hi Tomas,
For production system, all these functions should be coded in Java
instead of lisp, but it vas a great learning exercise this way,
especially in the context of argument evaluation and shallow binding
maintenance.
T
What I actually like most about wl is the Java FFI which is very
Hi Alex,
Perhaps 'fork' can be simulated by copying all runtime variables to a
thread?
not sure whether that's enough. The semantics is completely different,
what about open files and so on? Is it necessary to go that far in
terms of compatibility?
Cheers,
Tomas
--
UNSUBSCRIBE:
Hi Tomas,
On 10/4/10 7:27 PM, Tomas Hlavaty wrote:
..
For example, even simple 'if' is implemented as a picolisp fexpr:
(de if (C . L)
(loop
(T C (up. '@ @) (eval (car L) 1))
(T T (run (cdr L) 1)) ) )
..
I'm just curious, what does that 'up.' do? I can't find it in
Hi Jon,
I'm just curious, what does that 'up.' do? I can't find it in the
PicoLisp ref.
'up.' is an internal function implemented in directly in Java. It was
needed as a functional equivalent to 'up' because I needed something
that would expect quoted agruments. 'up' is not programmable.
//
Hi Alex,
For example, even simple 'if' is implemented as a picolisp fexpr:
(de if (C . L)
(loop
(T C (up. '@ @) (eval (car L) 1))
(T T (run (cdr L) 1)) ) )
one of the significant factors that are slowing it down is all these
environment lookups like the 'up.' and
Hi Tomas,
btw, the strategy of deriving _all_ flow functions from 'loop' is very
elegant!
Cheers,
- Alex
--
UNSUBSCRIBE: mailto:picol...@software-lab.de?subject=unsubscribe
Hi Alex,
btw, the strategy of deriving _all_ flow functions from 'loop' is very
elegant!
thanks! It is indeed elegant and very inefficient;-)
For production system, all these functions should be coded in Java
instead of lisp, but it vas a great learning exercise this way,
especially in the
Hi Tomas,
I have published my implementation of picolisp in java, in case you want
to compare with your implementation.
Thanks! Great!
It runs without problems.
I'm a little bit surprised, however, that it is so extremely slow.
When I tried this
(de fibo (N)
(if ( 2 N)
1
Totally awesome!
I have for a long time wanted Picolisp on other platforms. I think the
best thing about Picolisp is the language itself, not its current libraries.
Thank you!
// Jakob
On 10/04/2010 09:40 AM, Tomas Hlavaty wrote:
Hi Alex,
I have published my implementation of picolisp in
Hi Alex,
It runs without problems.
great!
I'm a little bit surprised, however, that it is so extremely slow.
How can these huge differences be explained?
I wasn't aiming for speed with this first version but rather for
understanding of the underlying concepts, namely how evaluation and
Hi Alex,
and then did
$ time cat x |java wl
ok.
Just to note, you are also measuring startup time of the JVM which might
be considerably slow compared to picolisp. In that sense, your
ErsatzLisp might have unfavourable timings in regard to pure
interpretation.
respond, however, and if
Hi Tomas,
Just to note, you are also measuring startup time of the JVM which might
be considerably slow compared to picolisp. In that sense, your
ErsatzLisp might have unfavourable timings in regard to pure
interpretation.
Yes, I'm aware of that. Still, if I just put (bye) into a file 'y'
Hi Alex,
Did you type anything into the text field? The example is very
simplistic so it expects an integer value in the text field when you
Yes, as I wrote in my previous mail, that field doesn't respond. That
is, it doesn't accept any keystrokes. But now it occurs to me that
these
21 matches
Mail list logo