IVF and MKL users

[g_f_thomas]

Most posts to the MKL forum are from C programmers working on Linux. Hasany consideration been given to having an MKL forum for Fortran programmers on Windows (sic IVF users) moderated by knowledgeable Fortran-oriented engineers? As it stands the MKL forum is dull and unresponsive to even the most trivial of queries which if posted hereon the IVF forum would be answered in a flash.

Gerry

[eos_pengwern]

Let's test this; I'll try re-posting my unanswered MKL query (posted on that forum three and a half hours ago) to this one...

[Steven_L_Intel1]

And I'll move it to the MKL forum. Perhaps this forum is too good, but three and a half hours is not long enough to wait for a response in my view.

If you feel that MKL issues are not getting needed attention, you can let me know and I'll pass on the complaint, but perhaps you're spoiled by the Fortran forums.

[g_f_thomas]

Let me see, a post of mine dated Mar 26 took until April 7 to elicit a response which wasn't helpful. MKL's implementation of lapack is flawed in some aspects. Matlab don't endorse MKL for lapack. Even the documentation of some Netlib code has been altered to from the original which is clearer and unambiguous.

Allow me to register my dissatisfaction of MKL forum support. Please raise their laggardly response with them.

Gerry

[eos_pengwern]

Touche; no complaint was implied Steve, but I don't think it's ever taken that long to get an initial response in the Fortran forum. I think your record is about 14 minutes! As you say, maybe we're spoiled, but we do appreciate it.

[brianlamm]

That's a tough one. My take is we should remember these forums are ... forums, not requests for technical support. I've been (metaphorically) spanked with that in the past. But it's a double edged sword: when one posts to the forum the whole community could immediately benefit. But I would suggest if you have a problem which impacts you significantly, go to Intel Premier Support and the relevant forum. My experience with IPS is 100% positive. Oh sure, I've had gripes, but the attention and professionalism in response from IPS has always been spot on. Additionally, when one encounters a real problem and goes to IPS, it'll get slated to be fixed, and the whole community again benefits, albeit indirectly and possibly without ever knowing it, or what went on to percipitate the "fix".

I agree this IVF forum is great and one does get spoiled. Just take a look at the 1/Infinity thread: I got a bit sour, but even after that I was still inundated with both Intel insider useful and enlightening responses, as well as the same type responses from clearly expert, at least in that sub-domain, people from "outside-of-Intel". Great stuff.

I would take away that one should post to the forum, and if not satisfied, then go to IPS.

-Brian

[g_f_thomas]

I don't disagree. IIRC the last (or maybe it was second last) time I contactedPS I was asked to explain what I meant by IVF. Really! This is not to say that they don't give excellent service, merely that they are not infallible.

I like this forum for its responsiveness and much of the dialog raised bythe community that's of interest to me. If an issue is raised that Steve feels would be better off handled by PS he says so and often asks that it be assigned to him. It doesn't get much better than that.

Gerry

[Steven_L_Intel1]

Regarding "IVF", it is not too surprising that some support engineers might not recognize that term. We don't call it that internally for the most part. Similarly if you refer to "CVF", some support engineers might not recognize it.

As our user base expands, we expand our support team. As I'm sure you'll understand, it isn't always easy finding people experienced in Fortran, even if they're otherwise technically good. So we do training and cross-support to bring them up to speed.

I pinged my counterpart on MKL and he promised that things would improve in the future.