Go read Animal Farm.
Meager Intel threads were no match for SPARC processors.
Nor were IBM POWER CPUs at the time, marketing-wise. They received a 0.75
discount.
Some processors were more equal than others.
Around that time there was an O'Reilly text from Cary Milsap that discussed
queuing theory. For Intel/AMD x86/x86_64 after 8 cores the system had an
adverse response in net throughput as the core count increased. It might
have even peaked at 4 cores.
Happy seasonal time of year to log off from the VPN. Enjoy.
Mladen, you can do better than this feeble attempt.
Pd
On Wed, Nov 27, 2019, 3:37 PM Vishnu Potukanuma <vishnupotukanuma@xxxxxxxxx>
wrote:
I am not sure about licensing,,, but if the above is correct... then it
has to be this.
probably when intel introduced with the concept of hyperthreading back in
Pentium4 days , where a CPU Core handled 2 threads at once... Even though
there was only one physical CPU Core, OS saw it as 2 CPUs, in order to
consider that that they must have factored the cost as cpu cores * .5 to
charge only for the physical cores... and slowly along the way, this was
misinterpreted and Oracle started loosing money?
Thanks,
Vishnu
On Thu, Nov 28, 2019 at 1:51 AM Mark J. Bobak <mark@xxxxxxxxx> wrote:
:-)
On Tue, Nov 26, 2019 at 8:10 PM Mladen Gogala <gogala.mladen@xxxxxxxxx>
wrote:
No, it doesn't make sense but it is correct.
On 11/26/19 11:24 AM, Jeffrey Beckstrom wrote:
This would be for IAS/Weblogic on X64 hardware. Our understanding is
that each core has a factor of .5. We are considering a 2 processor
server with 8 cores per processor for a total of 16 cores. Based on
the factor this would represent 8 processors to be licensed.
Does this make sense?
--
Mladen Gogala
Database Consultant
Tel: (347) 321-1217
--
//www.freelists.org/webpage/oracle-l
