In a new experiment, German publication Igor's Lab explored different methods for improving the operating temperatures on Intel's 12th Generation Alder Lake processors. Thanks to the outlet's efforts, we now have a way to drop temperatures by as much as 5 degrees Celsius.
The transition from the LGA1200 socket to the LGA1700 socket saw a 42% increase in pin count, resulting in a more extended socket. Consequently, Alder Lake chips are larger and more rectangular than Rocket Lake, which is smaller and has a more squared shape. The problem falls on the Independent Loading Mechanism (ILM). The ILM still latches to Alder Lake processors at the same pressure points as Rocket Lake. The result is that the mechanism pushes down on the processor in the center, causing it to sit lower than on the edges. It's the reason why Alder Lake's integrated heat spreader (IHS) eventually takes a concave shape after several hours of operation.
In contrast, Igor's Lab exhibited an Alder Lake processor that has seen several hundreds of hours of operation. You can see that the processor presents a concave form at the top and a convex shape at the bottom. In other words, the processor looks slightly bent along the points where the ILM exerts the pressure.
In a previous investigation, Igor's Lab noted that some motherboard vendors were using materials that didn't meet the required specifications for the socket holder. It ended up warping the processor. The IHS on most Alder Lake processors is flat when out of the box. However, samples with convex or concave IHS show up occasionally, so that doesn't help, either. The outlet discovered that adding a gigantic backplate could prevent the warping problem as long as you get it in there before you install the processor.
Coming back to the issue at hand, the concaveness of the IHS hinders the CPU cooler's baseplate from making direct contact, so the heat has to travel a longer distance through the gap with the thermal compound. The solution is easy, cheap, and quick. You only need to fix the bend on the processor so that the CPU cooler has optimal contact and, as a result, improves cooling performance.
To reduce the mounting pressure, you only need to add some M4 washers to the socket holder. The first step involves removing the four M4 Torx T20 screws that hold the ILM in place. The second step is to place the M4 washers in each screw hole, and it doesn't matter if the washer is metal or plastic. Lastly, you put the four screws back in place and screw them down. Igor's Lab didn't specify how much pressure you'll need to apply. For safe measure, keep them hand-tight.
Igor's Lab measured the temperature of the Core i9-12900K's P-cores with the utility HWiNFO at intervals of 500 ms. The publication used the popular Prime95 software with the Small FFTs preset and AVX2 (FMA3) as a stress test with a duration of five minutes.
| Washer Config | P0 max Δ | P1 max Δ | P2 max Δ | P3 max Δ | P4 max Δ | P5 max Δ | P6 max Δ | P7 max Δ | Average | Improvement |
|---|---|---|---|---|---|---|---|---|---|---|
| none/stock | 69.5 | 82.5 | 73.7 | 86.6 | 75 | 83.6 | 73.7 | 74.8 | 76.64 | N/A |
| 0.5 mm | 66.4 | 79.1 | 70.2 | 83.7 | 72.3 | 79 | 69.3 | 70.7 | 73.84 | -2.8 |
| 0.8 mm | 67.1 | 77.9 | 70.2 | 82.6 | 72.3 | 78.4 | 70.2 | 70.5 | 73.65 | -2.99 |
| 1.0 mm | 63.9 | 74.8 | 67.2 | 79.3 | 69.3 | 77.4 | 67 | 68.1 | 70.88 | -5.76 |
| 1.3 mm | 64.2 | 75.2 | 68.1 | 80.1 | 70.2 | 77.9 | 68.1 | 69 | 71.6 | -5.04 |
The experiment was performed with M4 washers of different sizes; however, the ones with a thickness of 1 mm yielded the best result, lowering the average temperature by 5.76 degrees Celsius. Unfortunately, thicker isn't better, either. The 1.3mm-thick M4 washers only improved the temperatures by 5.04 degrees Celsius. Unfortunately, the outlet couldn't test the 1.8mm M4 washers because the M4 Torx T20 screws' short length holds the ILM to the motherboard.
Currently, Lotes and Foxconn are the leading manufacturers that produce LGA1700 sockets. It'll be interesting to see whether the difference in quality varies from one vendor to another and see if there is an optimal size washer for each socket. Fortunately for us, Igor's Lab is already investigating the matter.
