First thing to test: how does the CPU-370 like to be mounted? All previous microchannel blocks have preferred having the injector slit(s) oriented vertically (pointing to the top of the board and the PCIe slots), with the channels themselves oriented horizontally (pointing toward the RAM slots and the I/O ports). CPU-370 seems to break that trend. Due to changes in the block’s design, however, the barbs are in the same orientation as they were on the CPU-360s. In these tables, “Vertical” and “Horizontal” refer to how the barbs are positioned.
So there is a preference in orientation, but it’s not a huge deal. That said, visually, “Vertical” is the correct orientation due to the way the text is positioned. The oddity is that the microchannels are oriented the wrong way compared to previous blocks. At this point, I can’t say for sure why this is the case, but my guess is that the new bow is best in this orientation.
For reference, here are pictures elucidating the orientation situation.
The CPU-360 rev1.2 could also be in this comparison, but for very good reason, it is not: I don’t have one to test. While I have two CPU-360 rev1.2’s, both are in permanent use for TIM testing. One isn’t even stock any longer, either. Because of this, I was unable to use them for this comparison, but the CPU-370 going head-to-head with the CPU-360 rev1.1 should reveal how the CPU-370 has evolved from the CPU-360s.
Two quick takes from this: the CPU-370 runs away from the CPU-360 rev1.1 thermally (a 3.4C improvement is pretty huge in the CPU Block world) and the flow rate has dropped a little with the CPU-370. Where does that improvement come from, though? From the CPU-360 rev1.1 to the CPU-370, both the top and the base have changed, but they’re also interchangeable. So I threw the CPU-360 rev1.1’s base onto the CPU-370 for a quick test.
And wow, it seems the entire improvement comes from the CPU-370’s top and the bow it induces (and the improved contact that comes from the bow). The CPU-360 rev1.1’s base is better thermally and less restrictive than the CPU-370’s base. To be honest, I’m surprised. I wonder if this also means the CPU-360 rev1.2 lags behind the CPU-360 rev1.1 (the base on the rev1.2 is seemingly identical to the CPU-370’s base).