3 ms·

The last presentation on Power at Hotchips left me with the impression that "Core" in IBM terms is 9/10ths marketing speak and whatever Power might have that th

The last presentation on Power at Hotchips left me with the impression that "Core" in IBM terms is 9/10ths marketing speak and whatever Power might have that their engineers might think of as "cores" probably has little to do with what their marketing people are telling their customers and software vendors who do their licensing "per core".

slededit 9y ago
Could you expand a bit on this? Are you saying they are inflating their core count or the reverse? Everything I've looked at indicates significantly more powerful cores.
- wmf 9y ago
  They "inflated" it from 12 to 24 but recently it looks like they deflated it back to 12 cores. However you count it, a full Power9 chip is competitive with a 24-core Xeon or 32-core Epyc.
  - olddellian 9y ago
    A full POWER9 chip should be compared with a 48-core Xeon. In regards to cores, the distinction is between the PowerVM and Linux (OpenPOWER/PowerNV) ecosystem variants. Both are made to process 96 threads, but the difference is in whether the thread processing units are grouped eight to a core (SMT8) or four to a core (SMT4). The PowerVM version, some say for licensing reasons, gets the SMT8 cores. Either way, you get 96 threads: 12 cores * SMT8 = 96 threads 24 cores * SMT4 = 96 threads
- Quequau 9y ago
  Sorry but I just noticed this reply. I don't want to claim that IBM is "inflating" their core count, nor the inverse. Rather that what passes for what we might think of as a "core" based on the ways software vendors describe things isn't directly transferable to what IBM has actually built. IBM has extremely flexible hardware which, using IBM's hypervisor, can be configured on the fly to present itself in a number of different ways. For example: few powerful compute units, many medium compute units, or very many less capable compute units. So right up front the idea (which a lot of people might assume without really thinking about) that a given chip has a fixed number of "cores" doesn't really hold true. More to the point the next question that presents itself "what makes up for a core and what determines how many a given processor has" is best answered with "it depends" (at least as far as IBM Power goes). One of the main reasons this matters is that all sorts of businesses use the idea of "cores" as a fixed entity as part of their pricing structures... and now it's all muddied.

slededit 9y ago

Could you expand a bit on this? Are you saying they are inflating their core count or the reverse? Everything I've looked at indicates significantly more powerful cores.

wmf 9y ago
They "inflated" it from 12 to 24 but recently it looks like they deflated it back to 12 cores. However you count it, a full Power9 chip is competitive with a 24-core Xeon or 32-core Epyc.
- olddellian 9y ago
  A full POWER9 chip should be compared with a 48-core Xeon. In regards to cores, the distinction is between the PowerVM and Linux (OpenPOWER/PowerNV) ecosystem variants. Both are made to process 96 threads, but the difference is in whether the thread processing units are grouped eight to a core (SMT8) or four to a core (SMT4). The PowerVM version, some say for licensing reasons, gets the SMT8 cores. Either way, you get 96 threads: 12 cores * SMT8 = 96 threads 24 cores * SMT4 = 96 threads
Quequau 9y ago
Sorry but I just noticed this reply. I don't want to claim that IBM is "inflating" their core count, nor the inverse. Rather that what passes for what we might think of as a "core" based on the ways software vendors describe things isn't directly transferable to what IBM has actually built. IBM has extremely flexible hardware which, using IBM's hypervisor, can be configured on the fly to present itself in a number of different ways. For example: few powerful compute units, many medium compute units, or very many less capable compute units. So right up front the idea (which a lot of people might assume without really thinking about) that a given chip has a fixed number of "cores" doesn't really hold true. More to the point the next question that presents itself "what makes up for a core and what determines how many a given processor has" is best answered with "it depends" (at least as far as IBM Power goes). One of the main reasons this matters is that all sorts of businesses use the idea of "cores" as a fixed entity as part of their pricing structures... and now it's all muddied.