Next meeting on Nov 18th

2 views
Skip to first unread message

Erik de Castro Lopo

unread,
Nov 8, 2010, 5:45:18 AM11/8/10
to fp-...@googlegroups.com
Hi all,

Our next meeting (and the last for this year) will be held on Thursday,
November 18th, starting at 6:30pm.

As usual we will be having two presenters, this time Trevor McDonell
and Dominic Verity. Abstracts for their presentation topics are included
below.

The meeting will be held at the Google offices at 48 Pirrama Road,
Pyrmont. I will send out an RSVP form and a map closer to the
meeting date.

Cheers,
Erik


Programming Multicore GPUs in Haskell
-------------------------------------
Speaker: Trevor L. McDonell
University of New South Wales

Modern graphics processors are massively parallel multicore processors
increasingly used for general purpose computations. However, good
performance requires highly idiomatic programs, whose development is
work intensive and requires expert knowledge. Accelerate represents our
efforts to raising the level of abstraction to GPU programming; an
embedded language of array computations for high-performance computing,
in which computations on multidimensional, regular arrays are expressed
in the form of parameterised collective operations such as maps,
reductions, and permutations. In this talk, I shall briefly introduce
the Accelerate EDSL, before focusing on the implementation of a CUDA
backend for parallel execution on NVIDIA GPUs.


Monads, Arrows, Traces and other Categorical Amusements
-------------------------------------------------------
(title with apologies to Martin Gardner)

Speaker: Associate Professor Dominic Verity
Centre of Australian Category Theory and
MU Programming Languages Research Group.
Faculty of Science, Macquarie University

As functional programmers most of us will, at some point, have heard about a
bag of mathematical tools which rejoice under the collective term "category
theory". Maybe we caught the whisper that our favourite functional "design
pattern", the {\bf monad}, had originally been inherited from category
theory. Possibly we were taught that category theory has been shown to be a
convenient tool for organising the formal semantics of programming
languages. If our functional mentors were particularly inclined to such
things, they may even have tried to claim that functional programming was,
in some sense, the "internal language" of category theory.

In this talk, I hope to convince you that all of these claims are indeed
founded in some truth. To do that my intention is to introduce you to just
enough category theory to show that monads and arrows do arise naturally
from purely categorical considerations. Along the way we will find that
functional programming really can be regarded as the language in which (some
parts of) category theory can be written.

Of course, were this talk simply an attempt to illustrate how pragmatic
functional programming practice can be encoded as an exercise in abstract
mathematics, then I suspect that I will win few friends by delivering it.
Instead, its purpose is to illustrate how we may plunder new functional
programming principles from our categorical colleagues.

--
----------------------------------------------------------------------
Erik de Castro Lopo
http://www.mega-nerd.com/

Erik de Castro Lopo

unread,
Nov 15, 2010, 4:13:18 PM11/15/10
to fp-...@googlegroups.com
Hi all,

Our next meeting will be held on Thursday, Novemver 18th (two days
from now), starting at 6:30pm.

As usual we will be having two presenters, this time Trevor L. McDonell
on Haskell and GPUs and Dominic Verity on Category Theory. Abstracts for


their presentation topics are included below.

The meeting will be held at the Google offices at 48 Pirrama Road,

Pyrmont, in a building called Workplace6 (or WP6).

(map:
http://maps.google.com.au/maps?f=q&source=s_q&hl=en&q=48+Pirrama+Rd,+Pyrmont+NSW+2009&sll=-25.335448,135.745076&sspn=46.465145,88.505859&ie=UTF8&cd=1&geocode=FSo8-_0dhw8DCQ&split=0&z=16
)

The RSVP form is here:

http://spreadsheets.google.com/viewform?formkey=cjlDYkJOSkF0RXNoVkwzUDBfUG1Ia2c6MA..

Whether you are coming or not, it would help us gauge numbers if you
could fill in the form.


Cheers,
Erik

> --
> You received this message because you are subscribed to the Google Groups "fp-syd" group.
> To post to this group, send email to fp-...@googlegroups.com.
> To unsubscribe from this group, send email to fp-syd+un...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/fp-syd?hl=en.

Reply all
Reply to author
Forward
0 new messages