CUDA Paper Submission

Hey Guys (and Girls :)),

I’ve spent some time on a CUDA related research and I managed to write it down to a paper.
I want to find place to publish it (don’t want to sell it or something). I’ve tried TOPC on ACM, but got rejected without review (answer was : “idea is good, but the paper is only 7 pages, and we require somewhere between 20 and 25”. Go figure).
TOPC ACM also made a suggestion to write at world scientific - I did so but I have no answer for a quite long time now (these guys are just like gone, no answer at all …) …

Do any of you know about some place, where I can submit for free and get a review fast ?

I am pretty sure the paper is worth it (it targets Nvidia GPUs and the CUDA language features); all I need is someone that does not care about how long it is and can review it in a reasonable time (or at least can answer - something like “We got the paper, the review is coming at some point” will be greatly appreciated …).

Thanks !

Ooh, what’s your paper about?

Also, have you just considered using blogging as a means of distributing your information?

Maybe you can get someone somewhere to start a nvidia cuda advancement/ research forum (as an additional forum)…

Nevertheless, if you want an ‘informal’ review, then sure, follow the informal channels; if you want a ‘formal’ review, stick with the formal channels

I don’t see why you can’t build on and expand your work (7 pages) and re-submit the required 20 pages - it is only about 10 pages more; have you attended a research methodology class before? - this really should help your thinking and writing (to get the extra pages)

Thanks for the answers so far.

I don’t want to blog about it; I believe publishing as paper can get it to more people (and make it easily find-able for a longer period of time; that is why I have spent so much time in formatting it as a paper)

And there is no point making “extra” 10 pages “just-because”. I managed to put everything I want in 7, so adding 10 pages filled with rubbish does not seem like a great idea to me.

I want to follow the formal channels, I just need some orientation which channels might server my needs - most of them are pure theoretical, or math oriented, or CPU oriented, etc.

If you can suggest any channel that I might try (fast response (or at least feedback), no tax required for submission, no “min” page limit (or if so, 7 pages is all I got), CUDA technology research accepted) I will be really thankful !

The problem with following the formal channels would be that formal channels follow a formal method to evaluate paper submissions, quite formally
Your definition/ view of a formal paper may not necessarily correspond and align with their view of a formal paper

For one, I can promise you that they are very likely to look for (signs of) a well formulated research problem, a well formulated research methodology, and systematic research findings, depending on the stage of the research; you likely got turned down, because these areas were lacking, in their view
Research papers imply research, and research is very systematic and methodological - in line with objective science; they judge your paper based on scientific content, not on how good your ideas are or how nice it reads - as the latter may be subjective views
So it is very possible to expand and build on your work, by further “fleshing out” the mentioned sections
I know people who would consider 7 pages as merely the introduction of a research paper
There are actually ropes to learn, when it comes to writing/ producing research papers
I am not saying you should not give it a shot; I am just saying that you will need to follow standard procedure and method

Perhaps focus on the significance of the research - what are the key conclusions of the research, what are the key resolves/ actions that can/ should be taken (not necessarily implying a negative connotation), and which specific people would benefit the most from the research
Then, act accordingly; you may possibly find that you would poorly reach your target audience by actually publishing the paper, and that can better reach your target audience via other means, because your true audience hardly read published papers

little_jimmy is exactly right. There are a set of requirements that submissions need to adhere to in order to have a chance of acceptance. For example, not including a section that explains your experimental methodology in a paper that presents experimental results is a good way to get rejected. These requirements are usually not spelled out explicitly and often vary from one venue to the next.

An easy way to get familiar with a particular venue is to read through the proceedings from the previous year, pick out what you consider to be the best written papers dealing with a similar problem, and try to adhere to the format that they use. It should be clear if you are missing any important sections.

TOPC ACM is a good quality journal. Don not be surprised by becoming rejected, many people are rejected in these types of journals, and it does not mean the idea is wrong or bad, its just that there are some journals that look for almost revolutionary research. Also, journals look for longer papers.
In the early years, one could make a very short and direct paper of 3 pages and publish it on a journal just the sake of the result itself being worth. Now things have changed, i don not know if for good or bad, but they care about how it is written, how important is the result based on their criteria, how figures are designed, what references are cited, they look for unsupported claims and ask why this why that, etc.

If you want a paper of your work soon, i recommend take to try computer science conferences, the papers are still peer reviewed and the accepted ones are presented by the author and later can get published on proceedings, depending on if the conference is good. Also, the amount of pages is around 5 ~ 8 so you are good to go!.

Take a look at this site which has a ranking of conferences, and look for the High Performance or Parallel Computing ones.