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 : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: pgerassi who wrote (134793)	5/13/2001 12:58:03 PM
From: muzosi	Read Replies (2) \| Respond to of 186894

2^32 bits is only 512MB i think for small values of 2 this could be true, but as long as there is no lorenz contraction 2^32 is 4G. NT and linux support files larger than 4T (64 bit offsets). This has nothing to do with how much of such a file can be read into the physical memory. Also most new x86 chips have more than 32 bit address pins so even > 2^32 address space is possible externally even though internal address registers are 32 bit only. But one can use different segments to point to different regions. All x86 CPUs with x87 FPUs can handle 64 bit integers all DPFP's can do this but it is irrelevant especially for x86 where FP register file access sucks. The overhead of getting bits from FP register file to integer registers kills any potential advantage. you seem to be a little bit confused about these details.

To: pgerassi who wrote (134793)	5/13/2001 1:11:50 PM
From: Robert Salasidis	Read Replies (1) \| Respond to of 186894

Yes I meant 2^32 bytes. As far as the FPU handling 64 bit integers, it is at a speed penalty compared to native CPU registers (2-10x if I am reading the P4 datasheets right). The support of large files in a variety of operating systems exists, but again a speed penalty would occur with the use of all pointers to such files. The point is that these would be situations that a 64 bit CPU would excel in, and benchmarks/applications which require this functionality would show the most benefit. The fact that the Itanium will only be released at 800 Mhz will likely not matter significantly if its tasks are chosen wisely.