Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: pgerassi who wrote (74344)	3/13/2002 2:29:43 PM
From: Ali Chen	Read Replies (1) \| Respond to of 275872

Dear Pete: "Did you attempt to take just two points at 1.5 and 1.6 to estimate 2.0? NO!" I am exceedingly sorry that my poor English prevents me from clear communication with you. In my post I said: "Please go ahead and try any two points from that set that are 10% apart, and you will find out that any pair will give you almost the same parameter values. You even can experiment with sensitivity to noise." In my poor language expression, it supposed to imply that I did all these exercises many-many times, otherwise I would not risk asking you to verify this with SHEER ADACITY. You then say: "I can take any multipoint curve fitting algorithm and make it fit the data points and get a 0% error rate for any given data point. That proves nothing other than the fitting procedure worked." The point that escaped your attention is that your polynomial must be of 7-th order to fit all 8 points. My "estimator" is of the first order. I suggest you brush out on spline theory a little bit to understand implications of this. "I used a different method using SPECrate_int scores. They show a different picture." With all due respect, I already have explained to you that the "rate" setup uses entirely different operating arrangement, with multiple copy of tasks, each with its own data sets, competing for room in a single tiny cache on time-shared basis. No wonder the picture is different, and I do not care how much within the frame of our discussion. I never looked into SPECrate and do not care at the moment, and you should not bring this ever again. You go on: "Output checking is also not performed ... What good is a run that yields erroroneous results?" Why are you spewing all this FUD without even any attempt to study the subject? For your information, SPEC might be the only benchmark that performs thorough correctness check. In case you miss it, that was exactly the way somebody discovered problems in earlier versions of Intel complier, and there was a public scandal around this. Since then all SPEC submissions seem to be passing the check, including AMD Athlons-XP Then you conclude: "The reason I and many others do not use the INtel compiler is that for many programs, it fails to either compile or produce good working code. And Intel hasn't been able to fix it for those either. Yet, I could take gcc (and many other commercially good compilers (even MS C/C++)) and it would making good working code first time." I doubt you ever tried it. "Working code first time" is an extremely insightful statement from a professional. You know, there is one definition of an idiot: When someone tries same thing many times and expect different results. I must be that idiot. - Ali