Charles Houghton-Webb wrote:
...........
The other thing that bothers me about stress testing my newly repaired machine is if the stress test DOESN'T break it !
I would have, all told, just given my "like new" computer, a hammering for nothing!!
True if "for nothing" includes finding out " the stress test DOESN'T break it !" in lieu of the stress test does break it.
Computers aren't exactly like batteries, where a certain amount of use will render than unusable.
As those who experienced the problem discovered Apple built in some fail safe protection. Early in the failing process our machines would restart occasionally because they got too hot and immediately recovered upon shut down. Later in the failing process our machines would not boot if the discrete GPU was turned on, some remedied that problem by removing AMD drivers from the system library.
So when the machine is stressed we learn if the fans come on appropriately, if the fans have sufficient capacity to keep temperature levels in the safe range, we learn how high the GPU temps will go and operate safely and we learn how fast the GPU temps will recover when unloaded.
Useful information if your suspicious of Apple's repairs based on four years of Apple's history of repairing the same machines repeatedly for this problem.
From comments some have made about getting "new" boards, this suggests that Apple has known for sometime they had a problem, created and tested a solution, produced the solution and filled an inventory supply chain with "new" boards even while they were using up old, highly suspect inventory. It would be reasonable that Apple is now purging their supply chain of old boards. That purge will take time. Somehow Apple can, without taking the machine apart, tell which board is in your machine or they are playing three blind mice.
In the past replaced motherboards didn't fail immediately. So how are you going to know if you have a suspect motherboard or a reliable replacement motherboard? Apple isn't talking. You are gambling that in March 2016 your recent repair is not going to fail again. Testing gives you an insight into whether the repair changed the norms.
I had my machine repaired by a third party. "Normal" GPU temps are 3..5 degrees lower. I have ran GPU loads sufficient to discharge the battery while MagSafe is plugged in, fan speeds never got higher than 5,300 and GPU temp reached 170F for a maybe a minute. Taking the load off cooled the machine to GPU temps in the 90s in less than 15 minutes. To me that indicates the fans are doing their job, the thermo paste and heat sink are doing their job. The machine feels quicker, but I have the feeling that's like your motorcycle in the spring compared to the same bike when you put it away in the fall. Those spring bikes are crotch rockets.