[U-Boot] about ram test in u-boot

Hi,
maybe this question not greatly related with u-boot.
when I read the wiki of u-boot: http://www.denx.de/wiki/view/DULG/UBootCrashAfterRelocation
I noticed this: Standard memory tests are not effective in identifying this type of problem because they do not cause stressful cache burst read/write operations.
We have a board, the ram works fine before, but recently, we need to change the ram for the lack of the original chip. After I read the datasheet of original chip, I found the parameter tRCD is wrong! In code, tRCD is set to 10ns, but in datasheet, the minimal tRCD is 15ns.
I know in u-boot, there is a command name mtest, it is a simple RAM test. it test fine in two tRCD condition. I also try memtester 10 loops, also fine!
Could somebody give me some advice to determine which value of tRCD is better? thanks.
here is the memtester log:
[root@root]# ./memtester 31M 10 memtester version 4.1.2 (32-bit) Copyright (C) 2009 Charles Cazabon. Licensed under the GNU General Public License version 2 (only).
pagesize is 4096 pagesizemask is 0xfffff000 want 31MB (32505856 bytes) got 31MB (32505856 bytes), trying mlock ...locked. Loop 1---10: Stuck Address : ok Random Value : ok Compare XOR : ok Compare SUB : ok Compare MUL : ok Compare DIV : ok Compare OR : ok Compare AND : ok Sequential Increment: ok Solid Bits : ok Block Sequential : ok Checkerboard : ok Bit Spread : ok Bit Flip : ok Walking Ones : ok Walking Zeroes : ok

Dear IaMaPlAyEr,
In message 2ae8403e0912221835s575a3e31m7578652a6d1e72de@mail.gmail.com you wrote:
After I read the datasheet of original chip, I found the parameter tRCD is wrong! In code, tRCD is set to 10ns, but in datasheet, the minimal tRCD is 15ns.
...
Could somebody give me some advice to determine which value of tRCD is better? thanks.
This depends on what you want to do.
Many people overclock their processors to get a little more performance for the buck. Doing this, they hazard the consequences like reduced reliability and lifetime.
It's the same here: you can go for reliability and lifetime and strictly follow the specifications of the manufacturer, or you can try a hand-optimized set of settings that appears to be working for you (but may fail on other boards oin the same type, or at other temperatures or other operating conditions, etc.).
It's also a matter of warranties - when you have 50,000 systems failing in the field becuase of memory issues you cannot blame the chip manufacturer if you are running their chips out of spec, and you might even sued for punitive damages because of gross negligence by your own customers.
Go figure...
Best regards,
Wolfgang Denk
participants (2)
-
IaMaPlAyEr
-
Wolfgang Denk