1

Topic: Dual-processor systems

Whether it is possible to make dual-processor system, in which one processor high-frequency, and another -  multinuclear? It is necessary that on first to launch virtualization of ancient applications, and on the second to do all remaining. Or processors should be identical...

2

Re: Dual-processor systems

Hello, Ejnstok Fajr, you wrote: whether > It is possible to make dual-processor system, in which one processor high-frequency, and another -  multinuclear? It is possible. > or processors should be identical... Should not.

3

Re: Dual-processor systems

Hello, Ejnstok Fajr, you wrote: whether > It is possible to make dual-processor system, in which one processor high-frequency, and another -  multinuclear? It is technically possible, but it seemed to me schedulers windows/linux do not consider frequency when scatter flows on kernels. I.e. it will be necessary to set afinity mask at start. Plus will be underwater  with NUMA and interruptions at active I/O. > It is necessary that on first to launch virtualization of ancient applications, and on the second to do all remaining. Dual-processor systems in itself cost more expensive. Probably it is more expedient single-processor with hardly more  xeon th? > Or processors should be identical... Within one generation and model audit can be different. But normally difference in the price between models with pair additional kernels or the tenth gigahertz is small enough, what not  and to take identical models.

4

Re: Dual-processor systems

Whether > It is possible to make dual-processor system, in which one processor high-frequency, and another -  multinuclear? So  many (or all?) mobile phones such

5

Re: Dual-processor systems

Hello, Ejnstok Fajr, you wrote: whether > It is possible to make dual-processor system, in which one processor high-frequency, and another -  multinuclear? > It is necessary that on first to launch virtualization of ancient applications, and on the second to do all remaining. Than it can be more favourable than usage of the normal multinuclear processor with adaptive control by frequencies of kernels?

6

Re: Dual-processor systems

Hello, Ejnstok Fajr, you wrote: whether > It is possible to make dual-processor system, in which one processor high-frequency, and another -  multinuclear? It is possible, but in it it is a little sense, for the modern processors are able to disperse themselves when the part of kernels is intensively used only, and even like to drive different kernels on different frequencies.

7

Re: Dual-processor systems

Hello, ononim, you wrote: whether >> It is possible to make dual-processor system, in which one processor high-frequency, and another -  multinuclear? O> so  many (or all?) mobile phones such Like, no. I.e. there was a joke with 1 slow not guzzling and 4 fast guzzling kernels, but in my opinion went out.

8

Re: Dual-processor systems

Whether >>> It is possible to make dual-processor system, in which one processor high-frequency, and another -  multinuclear? O>> so  many (or all?) mobile phones such aik> Like, no. I.e. there was a joke with 1 slow not guzzling and 4 fast guzzling kernels, but in my opinion went out. Well here I open Sony Xperia XZ1 Compact, and at once I see the Amount of kernels 8 (4+4) Microarchitecture  Kryo 280 Performance 2450 MHz + Kryo 280 Efficiency 1900 MHz Well that is not that that a choice vysokochastnost-mnogojadernost, and bluntly efficiency/sberegajnost. All all the same one horse-radish is the same ASMP which the HARDWARE "invented".

9

Re: Dual-processor systems

Hello, ononim, you wrote: whether >>>> It is possible to make dual-processor system, in which one processor high-frequency, and another -  multinuclear? O>>> so  many (or all?) mobile phones such aik>> Like, no. I.e. there was a joke with 1 slow not guzzling and 4 fast guzzling kernels, but in my opinion went out. O> well here I open Sony Xperia XZ1 Compact, and at once I see O> O> the Amount of kernels O> 8 (4+4) O> Microarchitecture  Kryo 280 Performance 2450 MHz + Kryo 280 Efficiency 1900 MHz O> Well that is not that that a choice vysokochastnost-mnogojadernost, and bluntly efficiency/sberegajnost. All all the same one horse-radish is the same ASMP which the HARDWARE "invented". Well and far not to walk - raspberry pi, actually it not ARM. BCM2836 (and its new modifications) is actually VideoCore with built in ARM the coprocessor

10

Re: Dual-processor systems

Hello, ononim, you wrote: O>> Well that is not that that a choice vysokochastnost-mnogojadernost, and bluntly efficiency/sberegajnost. All all the same one horse-radish is the same ASMP which the HARDWARE "invented".  you, I thought hammered on it in a bolt. O> well and far not to walk - raspberry pi, actually it not ARM. BCM2836 (and its new modifications) is actually VideoCore with built in ARM the coprocessor It soc how you understand that it videocore with  arm, rather the reverse? Well, the first loads videocore, but it not an occasion

11

Re: Dual-processor systems

Hello, aik, you wrote: aik> Like, no. I.e. there was a joke with 1 slow not guzzling and 4 fast guzzling kernels, but in my opinion went out. Went out nothing. Smoke about ARM big. LITTLE.