On Thu, 1 Mar 2012, Ben Hutchings wrote:
On Wed, 2012-02-29 at 21:09 -0800, Shannon Dealy wrote: [...]Actually, a RAM upgrade can most definitely have an impact. Replacing any hardware on the bus changes bus loads, propagation delays and overall timing of the bus.[...]Unfortunately, I'm not current on the hardware/buses involved here, so I don't know how these potential issues might apply to this case.I don't think RAM and PCI cards have ever been on the same bus. And with PCIe there is a separate link to each card.
Unfortunately, that doesn't actually isolate them, though it does usually reduce some of the sources of problems. There are always interface chips tying them together (otherwise DMA transfers would not be possible) and the behavior of these chips will change depending on the load and temperature, so activity on the bus on one side of the chip can significantly affect the timing behavior of the chip on the other side (I track down bugs like this for a living, just not on PC's in a very long time). I have seen bugs in software show up due to a change in manufacturing processes used to produce an interface chip in the system and hardware glitches on an "isolated" bus that were due to transmission line effects that rippled right through the chip that was supposed to be isolating the two buses. Of course there is always the old standby of noise passed between chips via the power supply traces. There are many more examples.
FWIW. Shannon C. Dealy | DeaTech Research Inc. dealy@deatech.com | - Custom Software Development - Phone: (800) 467-5820 | - Natural Building Instruction - or: (541) 929-4089 | www.deatech.com