8 vs. 12 Channel DDR5-6000 Memory Performance With AMD 5th Gen EPYC
As I wrote about last week within the Supermicro H13SSL-N EPYC Turin motherboard review, one of the factors leading me to purchasing that EPYC 9005 series motherboard was that this board offered support for full 12 channel DDR5-6000 memory performance compared to some of the other lower-cost Socket SP5 motherboards offering just 8 memory channels. For those wanting to quantify the performance difference between eight and twelve memory channels with AMD EPYC 9005 "Turin" processors, here are some benchmarks for showing the workloads that can really benefit from all 12 memory channels and other workloads where eight memory channels can be largely sufficient if looking to minimize costs.
The Supermicro H13SSL-N is a single socket SP5 motherboard for both EPYC 9004 and EPYC 9005 processors. The R2.0 revision of the motherboard supports EPYC 9005 processors up to 400 Watts and offers full twelve memory channel DDR5-6000 ECC RDIMM support. But when looking at other AMD SP5 server motherboards of similar price some of them like the Tyan Tomcat HX S8050, ASRock Rack BERGAMOD8-2L2T, and ASRock Rack GENOAD8UD-2T/X550 only offer eight memory channels. The eight memory channel limit is done to keep costs lower, some of the motherboards designed for compact servers and even micro-ATX form factor, and for some budget-minded server builds going for eight memory channels is sufficient.
For helping to quantify the impact of 8 vs. 12 memory channels for the 5th Gen AMD EPYC processors, I ran some benchmarks on the Supermicro H13SSL-N when populated for both eight and twelve channels. This is to help show the benefits for some workloads that can great make use of the additional memory channels while for other areas indeed you can keep costs lower and justifiable with eight memory channels. So take these benchmarks namely for reference purposes.
The Supermicro H13SSL-N was tested with the EPYC 9655 400 Watt processor, Micron 7450 Pro MTFDKCB3T2TFS NVMe SSD storage, and using 12 x 64GB DDR5-6000MT/s Micron MTC40F2046S1RC64BDY QSFF RDIMMs. Tests were done with the full 12 memory channel support and then removing four of the RDIMMs for comparing the resulting performance impact. Of course, depending upon the CPU core/thread count you are planning to use can also influence the benefits of going for twelve memory channels and so for this was a middle of the road 96-core EPYC 9655 comparison.
While the H13SSL-N was running within a 4U chassis with three 120mm intake fans, three 80mm exhaust fans, and liquid cooling for the processor, the DDR5-6000 DIMMs were running rather toasty -- similar to what I've seen with other DDR5-6000/DDR5-6400 AMD and Intel servers too. With this 4U server build I ended up installing two Corsair Vengeance Airflow Memory Cooling Fans (CMYAF) and they made a dramatic difference in the DDR5-6000 RDIMM temperatures. There was a ~30 degree difference and the memory modules not coming anywhere close to throttling. These Corsair Airflow Memory Cooling Fans were used on the RAM throughout all of the eight and twelve memory channel testing.
A wide variety of benchmarks were carried out while using Ubuntu 24.10 atop a Linux 6.12 kernel build.