I got a new (ECS geforcem100) motherboard, amd x2 dual 5600+cpu, and a maxtor 500g sata drive.

I built a w2k system, installed their latest drivers, and ran a small file checksumming program on a 12 gig video file over and over. It would not pass this test and would find a differnt checksum every 3 or 4 times through the file.


But is this really a drive problem? Where is the crc error? reading from the disk into the buffer, or trnasferring the buffer (over the sata ide line) to cpu ram via dma.


I tried everything to make it more reliable: I used 2 different sets of powersupplies, memory, cables, and I installed extra cooling on the drive. I lowered the amd clock to 1.8 which lowered the voltage from 1.3 to 1.1 (automatically). Still I would get checksum errors.

Might this be an indication that ECS doesn't do a good job of udma transfers and that it's not actually a problem in the hard drive itself?

I eventually took the mobo/cpu/drive back to frys. This is the second frys combo (last one was ecs and core 2 duo 4300) that had problems with my "checksum" test. But my old system will run and rerun this checkpoint for days with no errors.

I'm beginning to suspect ecs chipsets can't handle this sort of load.


Any opinions?

All ECS boards are cheap crap, in my opinion.

Would not touch them with a barge pole.

Buy a ASUS or ABIT.

Aopen & Gigabyte are also pretty well sorted system boards. You'll like the core2 CPU though!

_________________
I think ST is a sub company of Lucas

rocket777
UDMA errors occur when you have an overclocked bus/CPU, or if you've got a crappy IDE/SATA cable. In some very rare cases that could mean that the hard drive or the motherboard is damaged.

May also be a RAM or bus issue. Did you run Memtest86?

Thanks for the suggestion, yes I had.

It's been a while and I went with an intel mobo and a core 2 duo. This works w/o error using all the same parts. I've bought my last ECS board.