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

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Scumbria who wrote (49124)2/9/1999 10:24:00 PM
From: Tenchusatsu  Read Replies (3) of 1582936
 
<Do yourself a favor and don't start conversing at the same gutter level as Paul Engel. I think that you are above that.>

Funny you should mention Paul Engel, because with the way you keep repeating the same points over and over again, you are beginning to sound just like him.

<MMX proved that no matter how useless an Intel feature may be, it is still necessary to include it in your design. Otherwise, OEM's will not buy it. Intel FUD is powerful, no doubt.>

Maybe that's why you keep repeating your points over and over again. Perhaps if you can spread even more FUD than Intel, like "All mission-critical apps will be ported to 3DNow" or "Intel is losing money on Celeron" or "RDRAM has lower bandwidth, lower latency, and higher cost than multiple SDRAM channels," you'll be able to overcome Intel's so-called FUD.

Now getting back to the point, perhaps you can show me how dual PC100 SDRAM channels will have lower latency, higher bandwidth, and lower cost compared to a single 800 MHz RDRAM channel. Latency varies from transaction to transaction. Thanks to RDRAM's greater number of banks, average latency will be lower. How much lower depends on how many simultaneous transactions can be sustained at one time in a system. And don't forget that Intel's Pentium III has cache-streaming instructions that can hide some latency. Next, bandwidth will be the same, but once again, RDRAM's ability to pipeline transactions will make better use of that bandwidth. Finally, cost per megabyte may be higher with RDRAM, but the additional cost for a motherboard which supports dual PC100 SDRAM channels will be higher thanks to the increased pin count.

(It's different with servers, because with the large amounts of memory that servers possess, cost-per-megabyte is more important than the cost of the motherboard and chipset.)

There are only two reasons why AMD is going with SDRAM first in their K7 platform. The first reason is obvious. RDRAM is likely to be scarce when the K7 is scheduled to be released. The second reason is that AMD is betting the farm on a timely K7 release, so they can't risk a delay in the K7 schedule. Therefore, they can't afford the risks associated with early RDRAM introduction. That's it.

Tenchusatsu
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext