The NumaQ standard system configurations scale from 128 cores with 1TB memory to 4,096 cores with 32TB of global shared memory
in a single system image environment.
The largest portion of Endeavour is able to meet the large shared memory
requirement with 4 terabytes of addressable memory and can apply over 1,000 cores against an Open MP application.
Large global shared memory
saves time: time-to-results, time-to-solution and time-to-innovation.
Our spatial decomposition code has proven effective in a shared memory
environment  as well, where the speedups are a factor of 29 on 32 processors of an SGI Origin 3000 system and a factor of 50 on 64 processors of the same system.
Memory is the real bottleneck," says IBM Fellow and director of server design Jim Rymarczyk, who says that distributed shared memory
is the natural extension of symmetric multiprocessing architectures in which multiple CPUs share a single pool of memory, a unified memory architecture (UMA).
He adds, "It is also worth mentioning that pilot customers have been surprised by the ease of use through statements like: 'Using Numaconnect shared memory
clusters is just like running programs on my desktop, just much more powerful
No longer do you need the budget of major government funded research facilities to afford supercomputers that because of their exceptionally large shared memory
significantly reduce the time and effort required for research, simulation and modeling.
With up to 8TB of global shared memory
with as few as 12 processor cores, SGI Altix 4700 systems are enabling applications for a number of important projects:
Model Distribution: If a model is too complex to be computed within the time step, RT-LAB allows the model to be distributed over several processors, automatically handling the inter-processor communication through TCP/IP, FireWire or Shared Memory
This is not a problem in the shared-memory multiprocessor version if the MPI implementation is smart enough to take advantage of the shared memory
Tera Computer Co is the latest supercomputer company to support the OpenMP application programming interface for shared memory
programming, as part of bundled software available with its Tera MTA multi-threaded architecture systems.
The SRC-guided research significantly extends the path for cores to communicate by reading and writing to a shared space - known as cache-coherent shared memory