Replies to post #14641 on Intel Corp (INTC)

[Elmer Phud]

SmallPops

don't know what IBM bidded with. This was my point all along. IBM also has IPF designs. How can one (chipguy) know for certain that IBM did not bid with it? If Montecito is sampling right now would you not agree that IBM has samples up and running also? They could of bid with it. The second problem with chipguys post was his claim that Montecito is superior to Power 5. Would you not agree Elmer that both will show strengths and weaknesses relative to each other? Or do you also feel that Montecito will beat Power 5 top to bottom in ALL HPC apps?

I'm not qualified to speak with authority on this topic but I can only wonder if IBM has much commitment left to Itanium when they are pushing Power so hard. So if IBM did bid as the article said they did why would IBM not bid with Power? And if they did bid with Power why did CEA think a Monticeto in the bush was a better choice than a Power in the hand?

[chipguy]

Would you not agree Elmer that both will show strengths and weaknesses relative to each other? Or do you also feel that Montecito will beat Power 5 top to bottom in ALL HPC apps?

You are posing the fallacy of false dilemma. Your second
condition (complete dominance in all benchmarks) just does
not happen in practice even in cases of huge performance
disparity. In the days when Alpha was yielding about 2.5x
higher SPECfp2k performance than Athlon it still marginally
lost to Athlon in one of the 14 apps that make up SPECfp2k.

Why you pose the false dilemma of either being equivalent
or unrealistically dominant is curious. Either you are too
clueless even to know that the question was boobytrapped
or you did know it and you were being deliberately dishonest
in posing it. Please tell us so we know whether to point at
you and laugh or point at you and jeer.

[SemiconEng]

I don't know what IBM bidded with. This was my point all along. IBM also has IPF designs. How can one (chipguy) know for certain that IBM did not bid with it? If Montecito is sampling right now would you not agree that IBM has samples up and running also? They could of bid with it.


The thing is, In that scenario, I'm having trouble believing that with IBM's relative minimal enthusiasm, or support of IPF up to this point, that they would bid with an intel product, instead of their own Power product. It just doesn't make sense, intuitively or logically, to me, that they would do that, when they barely even acknowledge any IBM/IPF existence. As far as which is "better", I agree with you, we'll have to wait and see.