1

I made a research while looking for a low fan-noise rackmount server.

I visited a datacenter and observed that Fujitsu Primergy RX series had much lower fan noise than HP ProLiant or Dell PowerEdge. So going with Fujitsu.

I'm wondering if putting a 3.3 GHz CPU instead of the lower class 2.4 GHz will increase excess heat and fan noise significantly?

Any experiences with that?

adamsfamily
  • 285
  • 3
  • 10
  • May I ask why you want a rackmounted low noise server? – Stuggi Feb 25 '20 at 10:07
  • Simply because we do have a rack and wanted to avoid additional pieces of hardware laying around. Do you think that a tower server will be quiter than a rackmounted one? – adamsfamily Feb 26 '20 at 12:33
  • They usually are, since a tower has more cross-sectional surface area, thus allowing lower air speeds at the same flow rates, which lowers sound levels. All the internal heat sinks can also be made bigger, further improving heat transfer and lowering sound levels. – Stuggi Feb 27 '20 at 06:05

2 Answers2

1

Many server BIOS/BMC adjust the system fan speed based on the CPU TDP. This means that using, say, a 95w CPU vs a 145w CPU is going to command a significant difference in fan speed and noise.

I saw (well, eared!) it recently, when having two identical SuperMicro servers with different CPU and memory configuration: the one sporting a Xeon Silver CPU was much quieter than the one having an higher clocked, more core Xeon Gold processor.

shodanshok
  • 47,711
  • 7
  • 111
  • 180
0

Your comparison will not work because:

  • you compare different computers with different CPUs, memory, disks
  • you compare different computers with different heat path construction inside

And when you increase the clock of processor by 50% you can expect twice of heat. Also you should check if this particular computer support the new processor (from point of clock from motherboard, cooler, heat, etc.)

Romeo Ninov
  • 5,263
  • 4
  • 20
  • 26