Article: HP Wins Oracle Lawsuit
By: Michael S (already5chosen.delete@this.yahoo.com), August 6, 2012 1:07 am
Room: Moderated Discussions
Kira (kirsc.delete@this.aeterna.ru) on August 6, 2012 1:44 am wrote:
>
> And, to provide a counterpoint to the TPC-H (which is a
> pretty grungy benchmark in a lot of ways), a 512GB 128-core Montecito Superdome
> manages to be about 15% slower than a 1TB 256-core S64 VII 2.88GHz M9K on SAP
> SD-2.
That's simply not true. SAP-SD scores under old rules (2 sec latency) are not directly comparable to scores under new rules (1 sec latency).
>(8-socket Power7 is better than both, for what it's worth...)
>
> Claiming
> S64 VII is faster than Tuk just isn't supported by the data
But the opposite claim (SD32 faster) is supported with far far less data.
> - it's very much a last-gen design, with a last-gen I/O subsystem.
I/O is PCI-express, same as Superdome2. Only DRAM is "previous generation".
>
> And, to provide a counterpoint to the TPC-H (which is a
> pretty grungy benchmark in a lot of ways), a 512GB 128-core Montecito Superdome
> manages to be about 15% slower than a 1TB 256-core S64 VII 2.88GHz M9K on SAP
> SD-2.
That's simply not true. SAP-SD scores under old rules (2 sec latency) are not directly comparable to scores under new rules (1 sec latency).
>(8-socket Power7 is better than both, for what it's worth...)
>
> Claiming
> S64 VII is faster than Tuk just isn't supported by the data
But the opposite claim (SD32 faster) is supported with far far less data.
> - it's very much a last-gen design, with a last-gen I/O subsystem.
I/O is PCI-express, same as Superdome2. Only DRAM is "previous generation".