Re: Broken applications: Could we be honest?
On 7/8/06, Art Edwards <[EMAIL PROTECTED]> wrote: I have been writing to the list about two applications thatare so broken on the AMD64 distribution that they render thebox pretty useless. I'm sure one could say that two measlyapplications are no big deal. However, if you do scientific computation for a living, and two of the primary tools are broken, you now havea rather clumsy paperweight where a computer should be. You couldargue that we should simply learn new tools, and we could, but weshould really be doing science instead. This brought up the question, who uses 64 bit Linux anyway?Surely gamers do not drive the 64-bit linux community. It can't be the desktopcommunity, seeing that the standard office tool doesn't reallywork for 64-bit. I would think that scientific and engineering users would drive this community. Besides the instruction set, whichcan probably give some speed, but wouldn't justify the cost, the addressspace in 64-bit OS's mean that we can solve much larger problems. Unles you're not doing some heavy-duty, memory-intensive computation, 64-bits seemsto be simply a status symbol.For compute servers the amd64 distribution is fine. All you really need arelanguages (compilers), libraries and decent MPICH. We run our small, 32 bit Beowulf on debian with abandon, and from my experience, I look forward toconverting it to amd64 ... with a 32-bit node where things actually work.Unless such core pieces as the debugging tool (ddd) and the data display tool (xmgrace) are working, it is dishonest to pretend that the 64-bit versionis ready for testing. It would be very nice if you, and other distro's, wereto put appropriate caveats on the websites, saying that 64-bit is really not ready for the prime-time desktop. That way, we could make better purchasingdecisions.At the risk of imposing what we do at our work place onto your work flow, i find that users generally should have access to better debuggers/profilers than what ships with standard gnu distros. presumably if you are doing scientific computations, you probably have access to a commercial compiler? i know that the portland group compilers ship with a fairly good gui debugger if you are not satisfied with gdb (in parallel attached to each running process) also shouldnt users be using programs like xmgrace on their local workstations? again with out trying to impose my workflow to yours, i find sometimes users do silly things on the head node on clusters, and I tend to try and get my users to do post analysis etc... stuff that can run serially on their own desktops whenever possible. Jim-- Jimmy TangTrinity Centre for High Performance Computing,Lloyd Building, Trinity College Dublin.http://www.tchpc.tcd.ie/
Re: Broken applications: Could we be honest?
Hi, On 7/8/06, Oliver Rother <[EMAIL PROTECTED]> wrote: > At the risk of imposing what we do at our work place onto your work > flow, i find that users generally should have access to better > debuggers/profilers than what ships with standard gnu distros. Well, if you intend to start a flame war on the lists... but enough on that. Nope, I have no intentions of any flame wars, i guess i didnt phrase my initial comments very well (i pretty much meant that from a point of view that some of our users fortran90 codes that some of our users need to compile happily breaks different compilers and it just happens we favour a certain toolchain for certain/most things). i apologise for that. also is it necessary to be cross posting to so many lists? > presumably if you are doing scientific computations, you probably have > access to a commercial compiler? Oh, we do. Consider an project with a timeline of many years or even decades of years. would you choose a non onepn source (commercial) compiler/debugger for that project? I'm pretty sure, you won't. again my comments were from the point of view of various codes at our site uses just breaks different compilers or have different requirements. and im quite happy to use whatever program/applications that is available that is suited to a task. i guess im agonistic in my approach at using the appropriate tool for my tasks that i do. > also shouldnt users be using programs like xmgrace Talking about commerical applications from your point of view - why use free software for data analysis when powerful commercial packages like IDL are available? i think you misread that comment. i just made a comment on how users probably should run things like xmgrace/gnuplot etc... at the end of their job runs on their own workstations whenever possible. -- Jimmy Tang Trinity Centre for High Performance Computing, Lloyd Building, Trinity College Dublin. http://www.tchpc.tcd.ie/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Broken applications: Could we be honest?
Hi, On Wed, Jul 12, 2006 at 01:29:37PM -0600, Art Edwards wrote: > Excuse me for chiming in, but I think many places simply look > for the best performance and productivity/dollar(euro). We do use the PGI > compiler, > mostly because gnu had not had a f90-f95 compiler, and partly because > of, maybe, a 10% improvement in speed. heh, I think that whole choice f90/f95 codes and compilers issue a chicken and egg problem. its probably not a idea to get into any religous discussions on why one should use a certain toolchain to build things. > > What I find interesting is that both Fedora and Debian have similar > problems for different reasons. Debian has now stable release for > AMD64 because Sarge was released before AMD64 was really ready. This means > that we are all stuck in the beta test-site pool. It would be really nice > if Debian actually packaged up a "stable-like" version of AMD64 at the > same level as Sarge. Fedora has been moving so quickly, that they have > incorporated the same problems into a nominally stable release. > To be honest, I find that debian amd64 (sarge) is quite stable, we choose to use it simply because we like debian and what apt-get has to offer. In terms of building a compute cluster using debian amd64 (sarge) it's certainly a risk since its a release candidate rather than what debian stable is typically like. I guess if you find packages to be broken or dont work as expected, do as other have suggested on these lists, report the bugs, or fix them yourself and submit the fix to someone so that it gets to become stable quicker :) Jimm -- Jimmy Tang Trinity Centre for High Performance Computing, Lloyd Building, Trinity College Dublin, Dublin 2, Ireland. http://www.tchpc.tcd.ie/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
