Hi,
ok, if cython is a wrapper to manually done c code, then Theano is the
faster automatic code regenrator at run time :)
Under linux, mac, the easiest way to change theano flags is via the
command line under bash:
THEANO_FLAGS=linker=c python benchmark.py
would do the trick.
This only work w
Hi,
This is just to note that I did some optimization in Theano related to
this. I did "easy" stuff, like adding c code to some "trivial" op that
where bottleneck in our cases.
It is still slower then your cython example. Is it manually done
cython or generated cython?
Also, I fixed a crash due
Sure, I'll do some benchmarks on that soon.
Jason
moorepants.info
+01 530-601-9791
On Tue, Nov 5, 2013 at 9:27 PM, Aaron S. Meurer wrote:
> Assumedly if it doesn't, then you would be able to tell just by applying
> the sympy cse and seeing if it makes a significant difference in the run
> tim
Assumedly if it doesn't, then you would be able to tell just by applying the
sympy cse and seeing if it makes a significant difference in the run time.
Aaron Meurer
> On Nov 5, 2013, at 6:45 PM, Matthew Rocklin wrote:
>
> Sadly I'm not an expert in C compilers and I no longer have an awesome
Sadly I'm not an expert in C compilers and I no longer have an awesome
compilers-post-doc-office-mate. I don't know what they'll do in this
situation. His no-CSE statement would still hold if the input pointers
point to dynamically allocated memory. I don't recall what he said if they
point to s
Got it.
SymPy's code generators in codegen and autowrap don't support evaluating a
SymPy matrix as the expression. I'd like to improve those so that you can
pass in a sequences of Matrices and it returns a function that evaluates
all of the matrices in a single C/Fortran/etc file.
Here is a simpl
My previous statement was that Theano won over autowrap.ufuncify because
theano did CSE internally (theano's basic data structure is a DAG).
My understanding is that your C friend is correct IF the code is written in
the standard way, declaring each variable individually. I remember looking
at ol
I would still benchmark running cse and not. If cse() is fast enough,
it might out-balance the overhead of the code generation, especially
if it significantly reduces the resulting code size.
Aaron Meurer
On Tue, Nov 5, 2013 at 11:36 AM, Jason Moore wrote:
> My roommate who's a C guy suggested t
My roommate who's a C guy suggested that CSE was a waste of time if you
have a good compiler. He claimed that the compiler will do CSE anyways, so
doing it in the C code was not useful. I need to compare my Cython code
output with and without CSE (it currently uses CSE).
How would I use CSE with t
In my experiments with sympy-theano earlier this year I found that it was
competitive with our Fortran code generation (autowrap.ufuncify) on single
array outputs and outperformed it when CSE were involved. I don't have
anything recorded that profiles scalar operations.
Frederic might have some i
Ok, so I think I'm comparing things correctly then. It is surprising that
the compiled Theano code doesn't seem to execute that fast. The functions
I'm generating are essentially evaluating very long sympy expressions,
maybe Theano isn't great at that...
Jason
moorepants.info
+01 530-601-9791
O
`sympy.printing.theanocode.theano_function` calls `theano.function`
The result of `theano_function` is a compiled theano function.
On Tue, Nov 5, 2013 at 9:27 AM, Jason Moore wrote:
>
>
>
> Jason
> moorepants.info
> +01 530-601-9791
>
>
> On Tue, Nov 5, 2013 at 8:46 AM, Frédéric Bastien wrot
Jason
moorepants.info
+01 530-601-9791
On Tue, Nov 5, 2013 at 8:46 AM, Frédéric Bastien wrote:
> Hi,
>
> Just to know, witch version of Theano did you used? I have speed up a
> little the optimization phase of the compilation and I have a PR that
> will optimize this more.
>
I'm using the mast
Hi,
Just to know, witch version of Theano did you used? I have speed up a
little the optimization phase of the compilation and I have a PR that
will optimize this more.
Also, Theano cache the compilation of the c code. Where this timming
done with a empty theano cache or a filled Theano cache? If
Le 03/11/13 14:19, Jason Moore a écrit :
Ronan,
Thanks for looking at the derivation code. We haven't ever really had
any review of it outside of our mechanical engineer group, so this very
helpful.
I'll review your PR and look into the second two items.
Why do you think the dictionary represe
Ronan,
Thanks for looking at the derivation code. We haven't ever really had any
review of it outside of our mechanical engineer group, so this very helpful.
I'll review your PR and look into the second two items.
Why do you think the dictionary representation will be so much faster?
Jason
moo
Le 01/11/13 11:54, Jason Moore a écrit :
I've been tinkering with code generation for ODE's that
sympy.physics.mechanics spits out and have some results:
http://www.moorepants.info/blog/pydy-code-gen.html
Several people have posted topics on this recently. We need to build in
a code generator f
I've been tinkering with code generation for ODE's that
sympy.physics.mechanics spits out and have some results:
http://www.moorepants.info/blog/pydy-code-gen.html
Several people have posted topics on this recently. We need to build in a
code generator for solving ODE's into SymPy that would play
18 matches
Mail list logo
