10 CPU water blocks tested on socket LGA1700 and AM5 in a big roundup including a necessary reappraisal

Lessons learned

Even though this article was actually only started as a review of the original test of the Alphacool Core 1, it has admittedly escalated somewhat with 10 water blocks. It took almost 2 months of my free time and, looking back, I probably should have cut back. Not only because of the time involved, but also because I’m still not 100% sure about the test methodology. Uncontrolled variables such as the HWinfo logging interval or my approach to evaluating the data can certainly still be improved in one place or another. During testing, I have also stumbled across one or two enlightenments or sobering findings.

As described, I used Alphacool Eiszapfen quick-release fasteners so that I don’t always have to empty the entire water circuit when I change the block. To speed things up, I didn’t just use one pair, but several. Is it dawning yet? One QDC fitting had a female G1/4 thread and is equipped with a male-to-male adapter. I also compared this one fitting with another native G1/4 male and found very similar flow rates. It’s just a pity that there is in fact a quick-release fitting that looks like all the others, but eats up 10 l/h more flow than all the others.

I only noticed this when the same block on AM5 suddenly had different flow values than on LGA1700, which of course can’t be the case. So I tested all my quick-release fittings, sort of binned them, and was then able to sort out the one black sheep. The only problem was that I had almost finished all the tests by then and had to repeat all the runs that were originally carried out with this fitting (approx. 30%). So don’t worry, almost all the results in the diagrams have already been corrected.

Only with the Thermaltake Pacific SW1 on LGA1700, whose mounting hardware died after the second attempt, I was no longer able to provide a corrected measurement for. The temperature values in the diagrams should be a maximum of half a Kelvin worse, which is not decisive for the war, but the flow values on LGA1700 are simply not plausible. It’s exactly these kinds of stumbling blocks that you have to fall over first that still worry me a little. What do you think of the test setup and methodology? What would you perhaps do differently or better?

Summary and conclusion

I was very surprised that every block had some kind of mounting problem. Be it badly cut threads, ripped screws, poorly manufactured screw heads, grooves milled too tightly, suboptimal tightening torque or simply modest handling – almost every water block has some kind of ailment. The only blocks that don’t have this are the Hydro X XC7 from Corsair and the older Eisblock XPX blocks from Alphacool, although all of these are significantly behind in terms of cooler performance. Our test did not really produce an “all-round carefree” block with very good performance. The best compromise is still the Heatkiller IV Pro from Watercool. Or am I being too strict with my extremely German understanding of quality? 

We can’t really define a winner or make a clear recommendation in today’s test. On the other side of the field, however, there are water blocks that don’t really get along with Intel’s LGA 1700 socket and which we would advise against. Unfortunately, these include the Thermaltake Pacific SW1 Plus, which became an assembly nightmare with its plastic backplate and inferior screws and was also at the bottom of the charts in terms of performance. We also cannot recommend the acrylic version of the Liquid Extasy No. Uno, as the material simply cannot transfer the forces prevailing on the LGA 1700 socket without bending and risking leaks.

Alphacool’s old generation of XPX blocks are also no match in terms of performance on modern platforms, even with an aftermarket Core backplate. The assembly works reasonably well, but the age of the design is noticeable throughout. EK-WB’s Quantum-Magnitude also suffers from age, and is by far the most expensive water block in today’s test. If you dare to tighten the screws more than specified, the cooling performance is at least acceptable, but the price is still not in proportion to the performance. Of course, there is more to a CPU water block than just the performance: design, RGB lighting, manufacturing quality, modularity, etc.. If price and pure cooling performance are not the top priority, customers may still be happy here.

In general, it can be observed that the mounting of a water block is at least as important for the performance of modern platforms as the internal design. Here again, there are two approaches that manufacturers pursue: Either one tries to make the CPU straight or to prevent bending by the ILM. Or the concavity of the CPU is included in the design of the base plate and it is made at least convex enough so that the center of the base plate also touches the center of the CPU and exerts the highest contact pressure there. In practice, of course, most manufacturers use a combination of both approaches.

It gets tricky when several platforms are involved and, of course, when the CPUs and mainboards of the same platform are curved differently. And then of course there are washer mods, aftermarket ILM frames and people who lap their CPU flat. Slaughtering all these variables with more contact pressure may physically lead to the best contact, but the system also must still post. Finally, there are also specifications from CPU manufacturers on how much contact pressure a CPU cooler can theoretically have, but this also behaves differently in practice. Finding the right compromise here is an insane engineering puzzle.

If we look at the test and the results as a whole and soberly, then the Alphacool Core 1 is one of the best at solving this complex task, closely followed by the Aqua Computer cuplex kryos NEXT and then the Watercool Heatkiller IV Pro. Although the latter is cheaper in the basic version, it requires an additional backplate in order to catch up in terms of performance. Realistically speaking, these 3 blocks are almost on a par and only differ in a few subtleties. In terms of user-friendliness, each block also has its own little stumbling blocks and the visual design is, as we all know, a matter of taste.

In the midfield, the No. Uno from Liquid Extasy is a technically interesting concept with a hint of early access and, at least in the aluminum version, performance on a par with an EK-Quantum Velocity 2 D-RGB. This, in turn, is visually the most unusual and perhaps the prettiest candidate, but all the more modest in terms of mounting – if you want to be beautiful, you have to suffer. The XC7 Pro from Corsair proves to be a solid mid-range block with RGB lighting and simple assembly at a reasonable price.

In conclusion, I hope that I have done justice to the manufacturers and customers of these water blocks with this test. There will certainly be one or two small improvements to the test methodology and I already have various ideas for follow-up pieces – feel free to leave your ideas in the forum thread. But after so much thermal paste application and cleaning, I need a change of pace. New DDR5 ICs don’t overclock themselves after all.

The following water blocks are a test sample from the respective manufacturer: Corsair Hydro X XC7 Pro, Thermaltake Pacific SW1 Plus, Alphacool Eisblock XPX Pro 1U and Liquid Extasy No. Uno.
The following water blocks are retail goods purchased by us: Watercool Heatkiller IV Pro, Aqua Computer cuplex kryos NEXT, EK-Quantum Magnitude, EK-Quantum Velocity 2 D-RGB, Alphacool Core 1 Aurora and Alphacool Eisblock XPX Aurora.