[Arm-netbook] Another Interesting Tegra2 Board

Gordan Bobic gordan at bobich.net
Wed May 23 17:20:31 BST 2012


On 05/23/2012 05:04 PM, Bari Ari wrote:
> On 05/23/2012 10:41 AM, Gordan Bobic wrote:
>> You're seriously putting me off just by mentioning Advantech. I bullshit
>> you not, when I queried why they didn't implement CPU core voltage
>> scaling (part of Intel Enhanced Speed Step spec) on an Atom N450 board I
>> bought (tested by the fact that power draw doesn't change when you
>> change the VID at full speed between minimum and maximum - and it also
>> doesn't de-stabilize) I was told it wasn't implemented because it is an
>> "overclocking feature". Not to mention other gems such as options in
>> BIOS that don't actually do anything:
>>
>> 1) There is an FSB toggle between 667 and 800MHz, and regardless of what
>> you set, it always runs at 667MHz.
>>
>> 2) There are options to set the RAM timings manually, but whatever you
>> set them to, it will always use the settings it reads out of the SPD on
>> the DIMM.
>>
>> Of course if you don't test these things, you are likely to assume it
>> just did what you told it to.
>>
>> Did they concede they screwed up and offer a refund? Of course not.
>>
>> If I never hear of them again it will be too soon. Not to mention that
>> the motherboards were substantially overpriced to begin with (no - you
>> don't get anywhere near what you pay for).
>>
>>
> That's why coreboot got started. Broken BIOS by vendors that don't care.
> Or imagine trying to change a BIOS setting on 1,000 or more nodes in a
> cluster and the only way to do it was with keyboards and monitors.

This is actually far worse than just settings that don't work - I asked 
if the broken speedstep support is a BIOS fault and they implied the CPU 
core voltage is actually hard-wired.

Gordan



More information about the arm-netbook mailing list