Quad Core, I3, I5 And I7-- Bioinformatics Tool
4
0
Entering edit mode
14.0 years ago
Dk Shetty ▴ 240

Hello,

Are the new processors like 'quad core' and Intel 'I', compatible with the bioinformatics tools, even the unix based onces..?

hardware • 5.6k views
ADD COMMENT
1
Entering edit mode

It's not clear what you are asking. The phrase 'bioinformatics tools' is too general to give an answer.

ADD REPLY
0
Entering edit mode

Is the question around multicore, or something else?

ADD REPLY
14
Entering edit mode
14.0 years ago

I do not agree with the explanationless downmod of this question. Yes, it is basic (and maybe phrased plain wrong) but so are a lot of other ones here that do get contructive comments. I think DK Shetty should not generally be discouraged asking these types of things, especially since he is new to the site. On the other hand, he should have put more effort into phrasing a proper question.

As a response to Nico's answer: In my mind, two things can be compatible or not no matter in which order they are put in the sentence. Thus, if A is compatible with B then this implies the other way around. Also, the issue is a little bit more complicated than you outlined here.

Basically, the source code of a programm can be built for a certain architecture (eg. i686 for standard 32bit processors or x86_64 for standard 64bit processors (simply put). The executable (=built program) can be run within the same architecture. How many cores your processor has does not really matter in the case of the mere possibility of code execution.

However, with the rise of multi-core CPUs, programs need to have certain properties in order for them to be executed efficiently. For instance, modern hardware can process a multitude of threads at the same time. If the program spawns only one thread, it does not make full use of the hardware, even if it was compiled on that specific platform.

That being said, most of the up-to-date tools make extensive use of parallel thread execution. In some cases, there are specialized modifications of them in order to run them efficiently in a distributed environment (eg., MPI for clusters) or the the graphics processor (GPU, eg. CUDA).

As to UNIX vs. Windows tools: most of bioinformatics happens on Linux, for a multitude of reasons. Most web servers run some flavor of Linux and so do almost all computing clusters. The "real work" (read: contig assemblies, microarray analysis, and really everything with a large data volume; this is compared to commercial graphical end-user interfaces) is being done here. So the question in most cases is, how good are their Windows ports?

edit: a little rephrasing

ADD COMMENT
1
Entering edit mode

I downvoted it as it is a poor question. I believe that is what the downvote is for. Conversely I upvoted your answer, as it is helpful. The fact that the original question is poorly phrased and poorly scoped means it is unlikely to be of useful to people in the future. I would have happily tolerated a discussion on 32 vs 64bit architectures and software compatibility in bioinformatics. Regardless of whether users are new, I think some standards need to be maintained.

ADD REPLY
0
Entering edit mode

In my mind, processors aren't build to be compatible with software, but software are build to run on many architecture. But you're right my answer was really incomplete :)

ADD REPLY
0
Entering edit mode

Yes, you are right. I should have put "downmod without explanation" there (as Keith's comment wasn't there at the time).

ADD REPLY
0
Entering edit mode

+1 for a good answer. Nonetheless, I do not think "most of the up-to-date tools make extensive use of parallel thread execution" at least in the area of sequence analyses/NGS.

ADD REPLY
0
Entering edit mode

@lh3: yes, I might have overestimated that

ADD REPLY
1
Entering edit mode
14.0 years ago
Ketil 4.1k

We did some study of this some years ago, when we still had other architectures (Sparc and POWER) around. The results was that many tools fail to compile, or fail to run correctly on non-x86 architectures, many tools are single threaded (and will only use one core efficiently), some are multithreaded (can use more than one core, efficiency varies a lot), very few are distributed. Many tools want lots of memory per thread.

So the safest choice is to get the fastest Intel or AMD CPU there is, as many of them as you can afford, in a single box with lots of memory. This is the opposite of typical HPC, which tends to involve clusters of many relatively weak PCs with little RAM each, and a cumbersome queuing system which gets in the way.

Similarly, most programs are developed on Linux, they may have been ported to Mac, and sometimes to Windows. Thus Linux is the safest choice.

Your mileage may vary.

ADD COMMENT
0
Entering edit mode
14.0 years ago
Nico ▴ 180

I think the question is the opposite. Do those software are compatible with new processors ? For unix based software I think you "just" have to compile sources on your new processor. For Windows binaries, if they are compiled on one of the new processor the answer is yes.

ADD COMMENT
0
Entering edit mode
14.0 years ago
Michael 55k

Yes ;)

ADD COMMENT

Login before adding your answer.

Traffic: 2500 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6