3 ms·

Sadly these gen-1 ThunderX cores are pretty poor in performance and not particularly power-efficient either. Cache performance is especially sucky. https://www.

by slizard 10y ago

Sadly these gen-1 ThunderX cores are pretty poor in performance and not particularly power-efficient either. Cache performance is especially sucky. https://www.servethehome.com/exclusive-first-cavium-thunderx-dual-48-core-96-core-total-arm-benchmarks/ https://www.servethehome.com/exclusive-first-cavium-thunderx... http://www.anandtech.com/show/10353/investigating-cavium-thunderx-48-arm-cores http://www.anandtech.com/show/10353/investigating-cavium-thu... I hope that this is just a rebranded system showing the redyness of System76 and raising awareness about such products so that they're better positioned for the ThunderX2 release later this year.

runeks 10y ago
ARM was never particularly power-efficient. ARM was designed for low power, not power efficiency. The only case where an ARM core is more power efficient than a Xeon core is when they're both idle -- primarily because the ARM core was designed to be idle most of the time, while the Xeon core was designed to be in use most of the time.
- gpderetta 10y ago
  ARM the architecture has very little to do with power efficiency as it is out of scope (except that the old the ARM 32bit instruction set was amenable for a very low power implementation). ARM the company does design both low power and power efficient implementations of the ARM architecture. In fact for some specific power targets (mobile for example) they are more efficient than the corresponding Intel CPUs. This is still irrelevant as Cavium design its own ARM cores from scratch (i.e. it doesn't use ARM reference designs) specifically for server use; now, they are not very good, but that's another story.
  - runeks 10y ago
    > In fact for some specific power targets (mobile for example) they are more efficient than the corresponding Intel CPUs. I agree, but only if the measurement of power efficiency includes idle periods -- eg. "browse for 5 minutes, lock phone and wait 30 minutes, repeat". I don't believe you can provide a source showing an ARM CPU executing a particular workload (i.e. not including idle periods) at greater power efficiency than a same-generation Intel x86 CPU.

by slizard 10y ago

runeks 10y ago

ARM was never particularly power-efficient. ARM was designed for low power, not power efficiency. The only case where an ARM core is more power efficient than a Xeon core is when they're both idle -- primarily because the ARM core was designed to be idle most of the time, while the Xeon core was designed to be in use most of the time.

gpderetta 10y ago
ARM the architecture has very little to do with power efficiency as it is out of scope (except that the old the ARM 32bit instruction set was amenable for a very low power implementation). ARM the company does design both low power and power efficient implementations of the ARM architecture. In fact for some specific power targets (mobile for example) they are more efficient than the corresponding Intel CPUs. This is still irrelevant as Cavium design its own ARM cores from scratch (i.e. it doesn't use ARM reference designs) specifically for server use; now, they are not very good, but that's another story.
- runeks 10y ago
  > In fact for some specific power targets (mobile for example) they are more efficient than the corresponding Intel CPUs. I agree, but only if the measurement of power efficiency includes idle periods -- eg. "browse for 5 minutes, lock phone and wait 30 minutes, repeat". I don't believe you can provide a source showing an ARM CPU executing a particular workload (i.e. not including idle periods) at greater power efficiency than a same-generation Intel x86 CPU.