Cloud storage: Lower cost and increase uptime
Enter the slightly more portly Power 750+ and 760+ midrangers
The new Power 750+ machine comes in a 5U rack-mounted chassis, which is 1U taller than the Power 550 and Power 750 machines it replaces. Part of the reason is that this box has a lot more memory stuffed into it, which is enabled because by doubling up the processors per socket you also double up the memory controllers in the machine. This means you can double up the memory capacity, to a top of 1TB in this case.
Like its predecessors, the Power 750+ is a four-socket server, and in this case the sockets are stuffed with a Power7+ DCM with a total of eight cores running at either 3.5GHz or 4GHz. It would be cool if all 80MB of L3 cache memory on the Power7+ die was available to half the cores used in the DCM, but it looks like IBM is maintaining the same 10MB per core ratio in the DCM.
The good news is that you get two GX++ system ports coming off the pair of chips in the DCM, which allows you to hang as many as 1,302 drives off the machine for a maximum capacity of 1.17PB using 900GB disks. That is a lot of disk to be able to hang directly off a four-socket server with 32 cores.
The Power 750+ server has six PCI-Express 2.0 slots (presumably full height x8 slots) and has an integrated multifunction Ethernet card that gives customers a variant of different Gigabit and 10GE adapter options. The machine has only six drive bays, which looks to be a little skinny for a 5U box, but then again, this machine, like all other Power Systems boxes, assumes you will have remote I/O drawers and not try to have all the drives crammed into the server bay. (This is why you use InfiniBand as a storage interconnect, after all.)
The Power 760+ is essentially the same chassis but the DCM shifts from four cores running at 3.5GHz or 4GHz to six cores running at 3.1GHz or 3.4GHz. Why bother, you ask? Again, you have twice the memory controllers and twice the GX++ system buses coming out of those sockets, so giving up a little clock speed gets you a maximum of 48 cores in a four-socket machine, up to 2TB of main memory, and a little more than 20 per cent more aggregate throughput on jobs that like cores and threads than you can get with the Power 750+ system.
And, perhaps even more significantly, you have a box that overlaps nicely in terms of processing oomph with the low end of the Power 770+ system, giving customers who don't need as much peripheral expansion as the four-chassis, 16U Power 770+ system offers a machine that is more suited to their needs.
These two midrange Power7+ boxes offer the same operating system options as the baby Power7+ boxes. They will be available on March 15. ®
COMMENTS
untitled comment
> The chips support 20 logical partitions (IBM's name for what everyone else calls a virtual machine)
TPM, everyone else? Oracle calls them LDOMs.... :)
> The Power 750+ server has six PCI-Express 2.0 slots
Looks like the PCI-X is finally gone for good!
Re: IBM Power is extremely expensive so they must drive perception of "low cost"
@Phil.
If you had bothered reading the oracle presentation from hotchips, you would have seen that it uses OLTP workloads, when it put forth it's claim that the throughput of the T4 chip is more or less equal to the T3.
The really big feat was that Oracle managed to increase single threaded throughput by almost a factor of five, hence actually taking the T4 out of the niche where the T3 was stuck.
Now nobody likes using specint/fp. it's not a particular good benchmark as your rant about. Historical for example POWER servers have normally been much better at OLTP and IO heavy benchmarks, as this is what the servers have been geared at.
But it still doesn't change that the chip throughput of the T4 is roughly the same as the T3.
Now for example if you compare the T4 and the T3 you can actually use the Enterprise2010 benchmark, where you have can see that 16 T4 chips give you aprox x4 more throughput than 4 T3 chips does.
http://www.spec.org/jEnterprise2010/results/res2011q3/jEnterprise2010-20110907-00027.html
and
http://www.spec.org/jEnterprise2010/results/res2010q3/jEnterprise2010-20100825-00014.html
As for all the T4 world records.. again .. we've been over this several times. T4 is the fastest chip in the world on Oracle product benchmarks where the only competing chip is the T4. The only industry standard benchmarks that have been done on the T4 is the spec Enterprise2010, where it now gets trashed, as other vendors turn to the benchmark, and figures out how to tune for it. Just like I predicted.
And then there is the tpc-h benchmark, again a benchmark that Oracle cracked some years ago, and if you look at how the setup is on the Oracle benchmarks, it's very very different from all others. Again Great work by Oracle, but it doesn't really say that much about the superiority of the T4. Again one measuring doesn't really make a trend now does it.
And funny to see you echoing the newest Oracle marketing message. Power have bad I/O. It's so Carl Rove. Attack your opponents strongest side with FUD.
Just to counter that, then IBM did a real nice benchmark some years ago.. a SPC1 benchmark, where they rather than having a LOT of hardware with a lot of RAM for caching, and IO processors and and and.. simply used a virtualized solution with the storage attached to the Virtual IO servers, that then virtualized the storage and shared it to the virtual machine that ran the benchmark. The machine used was a partitioned POWER6 based power 595. Again an almost 5 year old machine.
Now.... the link to the benchmark is here:
www.storageperformance.org/benchmark_results_files/SPC-1/IBM/A00083_IBM_Power-595_with_PowerVM_SSDs/a00083_IBM_Power595-PowerVM-SSDs_executive-summary.pdf
As you can see the IO setup uses 14 ancient PCI-X adapters and 2 old PCI-X drawers, that are cabled for max connectivity and not performance (4 cables not 8) and it runs virtualized POWER solution, with VIO servers and all. Furthermore the setup uses Virtual SCSI. Old and slow compared to what you could do today with NPIV:
At the time of submission this was the World Record for the SPC1 benchmark. And again... this is a pretty standard setup, that isn't even extreem in any way.
Where it gets fun is to compare it to a Oracle Sun ZFS Storage 7420c Appliance benchmark, with a setup that isn't that different what Oracle could do today.
www.storageperformance.org/benchmark_results_files/SPC-1/Oracle/A00108_Oracle-Sun-ZFS-7420c/a00108_Oracle-Sun_ZFS-7420c%20_SPC1_executive-summary-r1.pdf
The SUN benchmark uses 2 storage servers with a total of 1TB of RAM and 64 X7550 Xeon processors. and a shitload of IO adapters. The host system driving the benchmark uses 6 PCI-e Gen2 dual port 8Gbit adapters. And even though the old POWER machine only uses to old PCI-X based IO with 14 pci-x SAS adapters, the drawers are even cabled for max connectivity and not performance (4 cables not 8) and it runs virtualized POWER solution, with trashes it with a factor of 2 both on throughput and on response time.
Now how you can think that a POWER7+ based server, with a IO system that is 3 generations newer than what is in the old POWER6 based p595, can have IO problems compared to a T4 based machine is IMHO a riddle. Again the T3-1 that is used in the Oracle benchmark above has the same IO system that the T4-1 has, and the T3-1/T4-1 IO system is IMHO better per chip, than the T4-4 has.
To be quite honest... you should try to read a manual. And try to understand what is going on under the covers rather than just echoing marketing material.
// Jesper
Re: IBM Power is extremely expensive so they must drive perception of "low cost"
You have failed to answer the question. Oracle of course knows the Specint for the T4 since it is a very easy benchmark to execute. It is probably only a fraction of a fraction of the work compared to building one of Oracles one-off DYI clustered solutions to beat IBM in TPC-C.
T3 scores were quite good at the time so those were published. Until proven otherwise customers will assume that the scores for T4 aren't very impressive.
Specint is important because integer work is what lots of applications do, including databases. It is one of many data points in the decision making process. Oracle should let us customers decide what is important and not.
Is bhp the only data point when buying a car? No. Would that information be of interest if you were looking for a sports car? Yes.
IT infrastructure monitoring strategies
What you need to know about cloud backup
Enabling efficient data center monitoring
Agentless Backup is Not a Myth
Top 10 SIEM implementer’s checklist
