SI
SI
discoversearch

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.
Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?


To: Tenchusatsu who wrote (81085)11/25/1999 4:39:00 PM
From: Charles R  Read Replies (1) | Respond to of 1574523
 
Tenchusatsu,

<Yet AMD is trying to derail the IA-64 train with their x86-64 plans. >

I am not sure if THAT is AMD's intent. I would say AMD's intent probably has a lot more to do with finding a credible raodmap going forward. With x86-64 (or AA-64) they have one.

<But for AMD to say that x86-64 is the right way to go, and throw not-so-subtle hints that IA-64 is the wrong way to go, now that's hogwash.>

Just curious, is it hogwash to say the opposite? If so, why?

<AMD is hoping that x86-64 will bring x86 to IA-64 performance levels>

This comment about IA-64 performance levels may be premature. First of all, no independent benchmark houses have verified what the first instantiation of IA-64 performs like. Also, it is clearly way too early to speculate about IA-64 follow-ons.

So, if you are suggesting that the performance target for Sledgehammer is Merced then I think that would be grossly underestimating competition.

Chuck



To: Tenchusatsu who wrote (81085)11/25/1999 5:33:00 PM
From: dumbmoney  Respond to of 1574523
 
Believe me, Chuck, I've seen the arguments raised at MPF. And I still think that x86-64 is nothing but a fantasy, not because it's a bad idea in itself (it isn't), but that AMD is hoping that x86-64 will bring x86 to IA-64 performance levels. (Yes, Fred Weber from AMD was indeed hinting at this.)

64-bits has nothing to do with performance. Alpha, which was 64-bits from day one, runs faster when executing 32-bit code - simply because the cache hit ratio is a little better with less data.

If you were referring to the new floating point model, that should indeed provide a performance boost. However, I expect that Merced will be the FP king - useless for internet servers, but great for games. :)