Tom's Hardware > Forum > CPU & Components > CPUs > Full memory benchmark from Xbit lab

Full memory benchmark from Xbit lab

Forum CPU & Components : CPUs - Full memory benchmark from Xbit lab

Tom's Hardware: Over 1.4 million members in 6 different countries available to answer all your high-tech questions. Sign up now! Its free!
Word :    Username :           
 

http://www.xbitlabs.com/articles/c [...] 800_5.html


I875 canterwood

A scaling of RAM lantency on 800FSB
288 NS or clock cycle with DDR 266 (i guess it clock cycle)
278 with DDR 333
248 with DDR 400

The scaling should be in theory linear as there a 33 mghz increase in bus speed and internal speed but in real life it -10 after is -40 a 4X time better scaling.DDR 400 optimization is real.The bandwith figure scale linear with all kind of memory.

On read bandwith not bad as long that access to a open page .

Write perf is still pretty low around 1 GB is bad when compare to RDRAM at close to 1.8GB consider there running on a very old chipset and much slower internal speed.

No write-read figure on a real workload.

PAT offer also much more memory bandwith 600 mb more that I865.The rest is a bit faster on PAT.That can be due to a flaw I865

I865

Is lantency figure is more normal or normal scaling.On bandwith wise it clear that I865 dont offer the same DDR 400 optimization it DDR 400 is at the same read speed that DDR 266.(can be a issue on springdale as it will be offer only in 1 month)

There are mix result with springdale i have conclute that the boards is not ready yet or bios issue.

I845Pe

Does offer a better lantency that flaw I865 but is lantency is very close to I875 only 20 over consider that many allready have Pe boards i dont think it will be so good to change it.Bandwith wise only read speed do change vs I875 few 100 of mb less (still that a lot).A others flaw of DDR single command bus like good old Pc 133/100???

Only cache man bench i found useful Sisoft is flaw if you ask me.Pc memory, ok you should stay with 3Dmark and graphic benchmark


[-peep-] french

Sponsored Links
Register or log in to remove.
Tom's Hardware > Forum > CPU & Components > CPUs > Full memory benchmark from Xbit lab
Go to:

There are 1174 identified and unidentified users. To see the list of identified users, Click here.

Please mind

You are about to answer a thread that has been inactive for more than 6 months.
If you still wish to proceed, please ensure that your posting is original and does not duplicate or overlap any prior responses to this thread.

Add a reply Cancel
Sponsored links
  • Ask the community now
  • Publish
Ad
They won a badge
Join us in greeting them