From: David Yu <yuorme@gmail.com>
Date: March 7, 2009 12:46:06 GMT+13:00
To: Raazesh Sainudiin <R.Sainudiin@math.canterbury.ac.nz>
Subject: Re: Compiling codemlZ Source
In-Reply-To: <92d15d080903061544j288656f7xe990d5f47e3bea7@mail.gmail.com>
Message-Id: <92d15d080903061546p5a4b487by45ee5efbe075eb73@mail.gmail.com>
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="Boundary_(ID_isQD6gDj7JOyZx5JphrT4g)"
References: <20090217214144.xiejkdn5s048ccsw@webmail.utoronto.ca> <20090219004043.0xpznlk834s4swg8@secure.math.canterbury.ac.nz> <92d15d080903021006l48e86a4fx9500bc46ebfe4234@mail.gmail.com> <E451C78E-B90D-4110-81F4-764F8DF38BCE@math.canterbury.ac.nz> <92d15d080903061544j288656f7xe990d5f47e3bea7@mail.gmail.com>
Domainkey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type; b=BJL7A1gMEEOWRPl7dNMptCwsOz82F3En6P45oXrpLDjAwIpZiqadUOPAKOIxXuydfk qurSbCBAXQEgUIqcHHunRcuvsbjNQF28gQQfy51e6VMWilEMvHu9tdGYnZW4ibd3qYhu +jnbMxTtmUj8xDiB3Vux7wv07xNR3yz1jFuuM=
Content-Length: 725485

Oops, forgot to attach the original files. :)

David

On Fri, Mar 6, 2009 at 6:44 PM, David Yu <yuorme@gmail.com> wrote:
Hi Raazesh,

I duplicated all the example analyses in the PhysChem_6MHC folder and compared the results (as given in the mrf file) using the 'diff' tool in cygwin. Most differences amounted to rounding differences at the 5th decimal place. The lone exception was Omega M8 (see below). A complete summary of the results is shown below and is saved in diffanalysis.txt. I've also attached the original output files so you can take a look at the results yourself. Attribution will be greatly appreciated.

David

Example output vs. Windows binary output
-only the mrf file was compared for all analyses. rub and rst files generated too many minor differences to list
-files with the .txt extensions are generated from my compiled windows binary
-changes are formatted as [parameter] [difference]
-nth indicates the decimal place of the difference between the two outputs

Charge M7
ÊÊÊ seed
ÊÊÊ kappa 5th

Charge M8
ÊÊÊ seed
ÊÊÊ w 5th
ÊÊÊ 
Omega M7
ÊÊÊ seed
ÊÊÊ tree length 5th
ÊÊÊ kappa 5th
ÊÊÊ q 5th 

Omega M8
ÊÊÊ seed
ÊÊÊ lnL 4th
ÊÊÊ tree length 3rd
ÊÊÊ kappa 3rd
ÊÊÊ *all p and q values (1st)
ÊÊÊ 
PolVol4 M7
ÊÊÊ seed
ÊÊÊ q 5th
ÊÊÊ ÊÊÊ 

PolVol4 M8
ÊÊÊ seed
ÊÊÊ kappa 5th
ÊÊÊ q 5th
ÊÊÊ w 5th

Polar2 M7
ÊÊÊ seed
ÊÊÊ tree length (internal branchs) 5th

Polar2 M8
ÊÊÊ seed
ÊÊÊ tree length (internal branchs) 5th
ÊÊÊ w 5th

Volume2 M7
ÊÊÊ seed

Volume2 M8
ÊÊÊ seed


On Mon, Mar 2, 2009 at 7:01 PM, Raazesh Sainudiin <R.Sainudiin@math.canterbury.ac.nz> wrote:
Dear David:
Let me know if you are able to replicate all the results in the paper or examples directory.
I'll distribute the binary you sent me with your attribution.
Thanks,
Raaz


On Mar 3, 2009, at 07:06, David Yu wrote:


Hey Raazesh,

I got a friend to help me with the getting a working binary as you suggested. I've run the executable through a couple of your example datasets and received identical results but I haven't extensively duplicated your results. In case you decide to distribute binaries in the future, ÊI've attached the Êexecutable with this email. So far, I've only had success running it in cygwin - it won't work from the Windows command prompt. Hope this helps.

David

On Wed, Feb 18, 2009 at 6:40 AM, <R.Sainudiin@math.canterbury.ac.nz> wrote:
Hi David:
I am not actively supporting the codemlZ. ÊI have no access to Windows. ÊThe code compiles on Unix/Linux and Mac OS X platforms under Gnu C compiler. ÊI don't distribute binaries. ÊI suggest you ask a local linux geek for a little help. ÊOthers have been able to compile on Cygwin in Windows.
Hope this is useful.
--Raaz

Quoting da.yu@utoronto.ca:

Hi Raazesh,

I'm interested in applying the physico-chemical codon models
implemented in codemlZ to the study of antibiotic clusters in
Streptomyces. Being more of a bench biologist than a computer
scientist, my attempts to compile the source for my Windows XP SP2
machine have come to nought. Thus far, I've tried using both tinycc and
MiniGW to make a working executable. Running the executables produced
by both compilers gave the following error: program too big to fit into
memory. Perhaps for this reason, I also can't delete these executables,
with the system telling me that it's being used even when it's not
running.

I suspect, because of the consistency of the error message, that I'm
hitting my head against something fairly obvious. If it's not too much
trouble, would you be able to help me resolve this issue or
alternatively, provide me with a working binary for Windows XP?

Thanks in advance,
David


----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.


<codemlZ>