AMD SeaMicro

AMD SeaMicro

SeaMicro SM15000 Fabric Compute Systems

AMD’s SeaMicro SM15000 family of Fabric Compute Systems provides the equivalent of 32 1RU dual socket servers, massive bandwidth, top of rack Ethernet switching, and high capacity shared storage, with centralized management in a small, compact 10RU form factor.

System Overview:
CPU Options
– 64 AMD Opteron™ processors with up to 64 GB DRAM and eight “Piledriver” cores per CPU
– 64 Intel® CPUs (“Haswell” or “Ivy Bridge” microarchitecture) with up to 32 GB DRAM and four cores per CPU
Up to 4 Terabytes of DRAM (AMD Opteron processors only)
Freedom™ Supercompute Fabric providing bandwidth of 1.28 Terabits/sec
Up to 16 x 10GbE uplinks or up to 64 1 GbE uplinks
Integrated Layer 2 switching
Fabric Storage up to five petabytes
– Up to 64 internal drives SAS/SATA solid-state drives or hard-disk drives
– Up to 1,408 SFF or LFF drives with external Freedom™ Fabric Storage arrays
Optional Hardware RAID 0, 1, 5, 6, and 10
Massively fault tolerant with in-service serviceability software upgrades
Runs off the shelf OS, Hypervisors, and applications
3.2 KW Average power consumption

Big Red II Cray Supercomputer at Indiana University

Big Red II Cray Supercomputer at Indiana University
With a peak performance of 1 petaflop, Big Red II offers users a hybrid compute environment consisting of: CPU compute nodes: Featuring two AMD Abu Dhabi X86-64 processors with 16 cores each, these nodes provide 32 cores and 64 GB of memory per node. GPU-enabled compute nodes: Containing one AMD Interlagos X86-64 processor and one NVIDIA Kepler K20 GPU, these nodes provide 32 GB of memory. All compute nodes are connected through the Cray Gemini interconnect. – Main Page – Main Page
This is the wiki page for the perf performance counters subsystem in Linux. Performance counters are CPU hardware registers that count hardware events such as instructions executed, cache-misses suffered, or branches mispredicted. They form a basis for profiling applications to trace dynamic control flow and identify hotspots. perf provides rich generalized abstractions over hardware specific capabilities. Among others, it provides per task, per CPU and per-workload counters, sampling on top of these and source code event annotation. The userspace perf tools present a simple to use interface with commands like perf stat: obtain event counts perf record: record events for later reporting perf report: break down events by process, function, etc. perf annotate: annotate assembly or source code with event counts perf top: see live event count To learn more, see the examples in the Tutorial.

Supermicro H8QGi+-F Four AMD Opteron™ 6000 series Motherboard

Supermicro H8QGi+-F Four AMD Opteron™ 6000 series Motherboard
1. Four AMD Opteron™ 6000 series (6300 ready) processors (Socket G34)

16/12/8/4-Core ready; HT3.0 Link


2. AMD SR5690 / SP5100 Chipset

3. Up to 1TB DDR3 1600MHz ECC

Registered DIMM; 32x DIMM sockets

4. Intel© 82576 Dual-Port Gigabit Ethernet


5. 6x SATA2 (3.0 Gbps) ports via AMD

SP5100 Controller; RAID 0, 1, 10

6. Expansion slot:

1 PCI-E 2.0 x16

7. Integrated Matrox G200eW Graphics

8. Integrated IPMI 2.0 with Dedicated LAN


pipelining (superscalar, OoO, VLIW (very long instruction word), branch prediction, predication)multi-core & simultaneous multithreading (SMT, hyper-threading)

SIMD vector instructions (MMX/SSE/AVX, AltiVec)

caches and the memory hierarchy