Intel Core Ultra 9 285K cooling testing: How much does it take to keep Arrow Lake cool in MSI's MPG Gungnir 300R Airflow PC Case?

Our most recent test bench for cooling utilized Intel’s i7-14700K and it (along with its i9 sibling) was one of the most difficult-to-cool CPUs in recent history. As a result, I recommended liquid coolers paired with CPU contact frames to prevent potential CPU bending for those looking to sustain the best possible thermal and CPU performance.

With their dramatically different architecture, Intel’s Arrow Lake (Core Ultra 200) CPUs present different challenges for cooling. In addition to using more advanced manufacturing and packaging processes, the CPU’s hotspot has shifted north compared to the previous-generation Raptor Lake CPUs, which may impact the performance of some coolers.

In this article we’ll investigate just what you need to cool Intel’s Core Ultra 9 285K. But before we get into the cooler testing results, we’re going to take a quick look at MSI’s Gungnir 300R case, which we used for this testing. MSI also sent along the motherboard and power supply we used, while Intel provided the Core Ultra 9 285K CPU.

MSI MPG Gungnir 300R Airflow Computer Case specifications

Features of MSI MPG Gungnir 300R Computer Case

*️⃣ Side view and customization options, hinged door

Looking from the side, the inner design of the Gungnir 300R is fairly standard for an airflow-focused case. At the top, you have support for extra fans or up to a 360mm radiator. The bottom houses a chamber hiding the PSU and drive cage, and serves to support a pre-installed LED GPU support bar.

There are only two parts that aren’t quite traditional here: To the side of the motherboard, you can install up to two additional fans. The other, more interesting feature is the side cover.

*️⃣ Efficient design

While fishbowl-style computer cases are visually appealing to many builders, I generally prefer more traditional setups because side intake isn’t usually as efficient or effective as traditional direct airflow from front-intake fans in cases like MSI’s MPG Gungnir 300R Airflow.

*️⃣ LED GPU Support Bar

Preinstalled to the bottom cage is an GPU support bar with ARGB lighting illuminating the MSI Dragon logo.

*️⃣ Build Quality and Price

The build quality of MSI’s MPG Gungnir 300R is sturdy, with no apparent weaknesses. However, it is a bit lacking in space behind the motherboard – so you’ll have to take some care with cable management. Its price is moderate at $169.99 – you can certainly find cases that are more and less expensive.

*️⃣ Rear side view, storage support, and cable management features

The backside of the case has velcro straps in the center and along the rear for cable management. Storage enthusiasts will be pleased to see four 2.5-inch drives supported on the back and an additional two drives of either 2.5- or 3.5-inch size at the bottom drive bay – enough for a “small” file server!

*️⃣ Four pre-installed 120mm ARGB fans

The fans included with a case have a huge impact on noise levels and thermal performance. MSI includes three 120mm intake fans and one 120mm exhaust fan in the Gungnir 300R, all of which are PWM-controlled. For this cooling setup, I’ve set these fans to run at 35 dBA.

*️⃣ IO Panel

The IO panel is located at the top of the case and features a button to quickly change LED lighting settings on the fly, which is connected to a hardware ARGB hub. There’s also one USB-C and two USB-A ports here, along with separate mic and headphone jacks.

*️⃣ Dust filters

The top, side, and bottom of the case are protected with easily removable dust filters. The front is filtered with mesh, but does not feature an easily removable filter.

Image 1 of 2

*️⃣ Radiator and Fan Support

Radiators up to 360mm can be mounted at the top or front of the case. Two fans can be installed on the side of the unit, three on the top, and one on the rear exhaust.

Other components

*️⃣ MSI Ventus 3X OC RTX 4070Ti Super GPU

MSI provided the Ventus 3X OC RTX 4070 Ti GPU for use with CPU + GPU cooler testing. During stress testing, this GPU will consume up to ~290W of power, but typically consumes much less in gaming.

What I like about this GPU is that it features the same bulky heatsink and fans featured on MSI’s more power-hungry RTX 4090 Ventus. This might be a bit of overkill for a card that consumes less than 300W, but it allows for cool and quiet operation in any workload you’re likely to encounter.

*️⃣ MSI Z890 Carbon Wifi Motherboard

MSI also sent along its newly released Z890 Carbon Wifi motherboard for testing. One of the things that I especially liked about this board is that the CPU power connection has been moved from its traditional spot in the top left corner to a spot above the RAM slots. This makes installing the CPU power cables a much simpler task.

Many motherboard makers skimp on built-in SSD heatsinks, cooling only one side – this can cause throttling in some scenarios – but not MSI. In addition to supporting five NVMe SSDs, the SSD heatsinks that come with the Z890 Carbon Wifi motherboard are double-sided and so should effectively cool SSDs in common scenarios.

The motherboard also has two PCIe 5 x16 slots and one PCI-e 4 x16 expansion slot for use with GPUs and other devices.

*️⃣ MSI MAG AI1250GL PCI-E 5 PSU

To power our cooling setup, MSI provided a MAG AI1250GL PCI-5-supporting PSU. The PSU also ran cool and quiet in all scenarios I tested it under, but the components I’ve used won’t push the PSU to its limits – your experience might be different with more power-hungry parts.

Finally, Intel provided both a Core Ultra 7 265K and Core Ultra 9 285K CPU for use in our cooling testing. Today we’re covering the cooling needs for the higher-end Core Ultra 9 285K, and as you’ll see in the benchmarks below it’s much easier to tame than previous-generation Intel i9 CPUs – roughly comparable to Intel’s i7-13700K in cooling difficulty.

Testing configuration – Intel LGA1851 platform

I do a few things to ensure my results are useful to system builders and upgraders. To begin, I strictly regulate the ambient temperature to 23C during testing. I do not accept results at 22 or 24C as valid, even though, in theory, there shouldn’t be much of a difference.

I also do my best to emulate typical user conditions, with thermal typical thermal loads, tested in a real case – not on an open bench, which can decrease the difficulty of cooling.