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


To: dougSF30 who wrote (183211)1/12/2006 11:49:31 PM
From: Joe NYCRead Replies (1) | Respond to of 275872
 
Doug,

I still wonder about 333 multiple for HT, and any impact on core freq, etc. Or maybe everyone got that wrong, and HT base stays 200, CPU freq's are multiples of that, but memory speeds are multiples of 333? (or 400 for DDR2-800).

Based on my understanding, the base (reference) frequency does not change. It would completely screw up HT, for one. Supplying a second base frequency to the CPU does not make a lot of sense to me.

Besides, the granularity that 200 MHz provides for CPU frequencies is great. Look at the trouble Intel is in with teir increasing FSB as a base for CPU clock speed (especially Mentocito).

As far as memory speeds, you go from CPU speed, divided by divisor to arrive at memory clock speed, which may not land at 333 MHz DDR2 (just like it happened with DDR1), but when we move to DDR2-800, everything is fine again.

Actually, DDR2-667 runs from 166 MHz clock x 4, doesn't it?

One additional improvements with the divisor situation is that with increasing clock speeds (on high end parts that get benchmarked), even if the clock speed does not land exactly land on clock speed producing 667 (hence lower), the higher the processor speed, the narrower the misses will be.

If it worked like Rev E, I'd expect 2.866GHz, with a 333 base. I guess, given that HT spec only has 1.2 and 1.4, perhaps the 333 thing has all been a misunderstanding, and that only applies to the memory controller.

I think it was. Let's wait what the INQ says (and I mean Mike or Charlie).

Joe