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)
AMD 218.44+1.1%2:05 PM EST
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: wbmw who wrote (245368)1/4/2008 1:28:07 AM
From: pgerassiRead Replies (1) of 275872
 
Wbmw:

At single threaded apps where 3 cores aren't doing much, I'll agree to that (Barcelona behind faster clocked C2Ds and C2Qs). But when all four cores are heavily loaded with different tasks, it isn't a dog, but better than C2D or C2Q. This leaves out those easy to scale rendering tasks where the code is shared amongst all four cores and the data fits in the cache and not much interaction occurs. Those were put into reviewer's benchmark sets by Intel during the P4 era which was bad at jump laden AI and physics stuff, but good at simple streaming stuff. Where Excel sorts had to be run in the background to allow the P4 to slip ahead of the Athlon which trounced it in most relevant tasks.

Well UT3 does have a number of interacting threads that do not share code or data much. That is a bad workload for a FSB connected quad which is shown by how well the Penryn dual core does over the older Conroe quad cores in that same benchmark. Yet to get higher realism demanded by gamers on these multicore CPUs, this is the direction of future games. This will also be true of applications over time for some of the same reasons. With the current C2D and C2Q platforms, which don't scale well at this sort of load, This will cause Intel trouble in the long run.

Intel is hoping that Nehalem and its successors will do better at this type of work load. Else they will be in for a real bad time, unless they can scale the single thread performance of their CPUs.

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

HomeMailHotSubjectMarksPeopleMarksKeepersSettings
Terms Of UseContact UsCopyright/IP PolicyPrivacy PolicyAbout UsFAQAdvertise on SI
© 2025 Knight Sac Media.  Data provided by Twelve Data, Alpha Vantage, and CityFALCON News