I reduced the intensities (from 1000 to 800) on my AMD cards and it's run for just over 11 hours on HiveOS (using 1.47) and 10 hours on windows using 1.48. Before I lowered my intensity from 1000 to 750 on the windows machine, I was getting a lot of "the GPU is responding slowly" and it would stop accepting jobs for about 5-6 minutes before finally timing out the card.
You can continue mining even if you get the warning. The default pool timeout is set to 60 seconds, so the miner stop to receive jobs it will fail with a different error message. If the gpu crash, you get a lower power usage. So sounds like you have a gpu crash and the thread in the program also crash because the hashrate hasn't been updated.
A higher intensity might need more power(voltage) to run stable.