I would like to fine out the best way to cool the GPU, therefore, I tried few difference combinations to install the PCIe card. End up the numbers confirmed that the best way is actually keep the GPU at slot 1, simple.
However, I also found that the power draw characteristic is a bit different when the graphic card is installed in slot 1 or slot 2.
When I install the card in slot 1, even with a "bridge" in between the mini 6pins and the 6+8pin to balance share the loading. The power draw will not be really even distributed.
The situation is very clear during the Luxmark test because of the relatively stable power draw. Clearly there is a power draw difference between the booster A and B.
But when I install the same card in slot 2. With exactly the same power connection. The power draw is almost eventually distributed.
This make me wonder if the PCIe slot 1's 12V line is somehow connected to the booster A or B. And the software / sensors actually unable to tell the exact power draw.
I tried it few times, always the same result. And from memory, someone use the EVGA powerlink also can't perfectly balance the loading. Which doesn't make any sense to me. Because the powerlink should combine all power source and transmit the electricity via a single piece of metal. There is no way for the card to constantly draw more power from one of the booster.
Why I raise this topic? Mainly for fun.
But if anyone interested in it, you may try to install the card in different slot and observe the power draw. Then we can know more about the cMP.
P.S. I also found that install the card in slot 1 will run much cooler, may be due to easier to let the cool air flow through PCIe compartment and push the warm air out. If the card in slot 2, the cool air will flow mainly below the card, which will be much less effective to push away the raised warm air.
However, I also found that the power draw characteristic is a bit different when the graphic card is installed in slot 1 or slot 2.
When I install the card in slot 1, even with a "bridge" in between the mini 6pins and the 6+8pin to balance share the loading. The power draw will not be really even distributed.
The situation is very clear during the Luxmark test because of the relatively stable power draw. Clearly there is a power draw difference between the booster A and B.
But when I install the same card in slot 2. With exactly the same power connection. The power draw is almost eventually distributed.
This make me wonder if the PCIe slot 1's 12V line is somehow connected to the booster A or B. And the software / sensors actually unable to tell the exact power draw.
I tried it few times, always the same result. And from memory, someone use the EVGA powerlink also can't perfectly balance the loading. Which doesn't make any sense to me. Because the powerlink should combine all power source and transmit the electricity via a single piece of metal. There is no way for the card to constantly draw more power from one of the booster.
Why I raise this topic? Mainly for fun.
But if anyone interested in it, you may try to install the card in different slot and observe the power draw. Then we can know more about the cMP.
P.S. I also found that install the card in slot 1 will run much cooler, may be due to easier to let the cool air flow through PCIe compartment and push the warm air out. If the card in slot 2, the cool air will flow mainly below the card, which will be much less effective to push away the raised warm air.
Last edited:
