asteroza + cpu   79

custom CPU's for 2000 euros, but on a 350 nanometer fab process. But hey, no NDA's!
private  custom  chip  CPU  small  lot  DIY  hardware  electronics  devices  semcionductor  fab  manufacturing 
10 weeks ago by asteroza
Thermal Grizzly High Performance Cooling Solutions - Conductonaut
Apparently a VERY fancy pants thermal grease for CPU's. Has some limitations but is VERY high performance...
CPU  cooler  heatsink  thermal  grease  compound  DIY 
august 2019 by asteroza
Solved: Windows 2000 VM on ESXi 6.7 showing 100... |VMware Communities
So, it looks like the consensus is, where possible, use the uniprocessor HAL and a single vCPU with a single core, but if you really need it, switch to the multiprocessor HAL and stay unisocket with multicore. Multisocket seems to be a world of hurt if you have a problematic situation. The problem here is a lot of documentation is spread out and recommending uniprocessor, but what they really mean is single socket.
VMware  ESXi  windows  2000  high  CPU  usage  sysadmin  tips  tricks 
august 2019 by asteroza
Home - SiFive
Intel/Qualcomm/Samsung are getting serious, feeding money into this fabless RISC-V shop, as they are desperately trying to get out from under the thumb of ARM. Probably good for RISC-V ecosystem, but how much will they give back to the opensource RISC-V community? Plus, there's the China angle to consider...
RISC-V  fabless  SoC  design  CPU  custom  hardware  electronics  devices  alternative  ARM  opensource 
june 2019 by asteroza
What is VMX Unrestricted Guest? |VMware Communities
legacy flag is helpful, but doesn't completely work sometimes though.
VMware  workstation  player  CPU  version  bypass  sysadmin  tips  tricks 
may 2019 by asteroza
decapped chips are kinda soothing visual material
CPU  microcontroller  chip  semiconductor  die  hardware  electronics  devices  reverse  engineering  gallery 
april 2019 by asteroza
$0.01 Flexible Plastic ARM Processor by PragmatIC - YouTube
PlasticARM is a FlexIC based ARM Cortex M0 based SoC on a flexible plastic (substrate?)
flexible  PCB  plastic  film  hardware  electronics  devices  CPU  ARM  SoC  circuit 
december 2018 by asteroza
MIPS Goes Open Source | EE Times
I wonder if this is fallout from rampant MIPS pirating in china, such that it was effectively unstoppable, so might as well control the defacto standard via stewardship of the ISA in the open...
MIPS  ISA  opensource  hardware  electronics  devices  CPU  core 
december 2018 by asteroza
Trying to reinvent the CPU/GPU
CPU  GPU  research  technology 
june 2018 by asteroza
T2 Unlimited – Going Beyond the Burst with High Performance | AWS News Blog
New burst CPU type T2 instance, so you don't cap yourself at the knees when you actually need to burst
AWS  EC2  burst  CPU  VM 
december 2017 by asteroza
NVIDIA Deep Learning Accelerator
Wow, Nvidia just opensourced a deep learning chip. Likely trying to kneecap all the AI hardware/chip startups. Notably, this is not for training, but is a deep learning inference accelerator...
Nvidia  opensource  chip  CPU  design  deep  learning  machine  AI 
october 2017 by asteroza
scylladb/scylla: NoSQL data store using the seastar framework, compatible with A...
The shared nothing, thread per CPU, CPU sharding, distributed DB using Cassandra interfaces
CPU  sharding  noSQL  database  cassandra  Delicious 
june 2017 by asteroza
Annapurna labs
Amazon has its own CPU company to power it's IoT gadgets, rather than selling assorted whitebox stuff from asia...
amazon  ARM  CPU  IoT  internet-of-things  hardware  electronics  devices  Delicious 
june 2016 by asteroza
Yahoo built a frankenstein framework to support frankenstein hadoop clusters, with a custom MPI over RDMA protocol, infiniband interconnects,and Yarn scheduling SPark and other Deep Learning workloads. The smart approach is an integrated single cluster, so deep learning workloads stay in cluster, but can migrate between GPU and CPU stages.
Caffe  Spark  hadoop  deep  learning  cluster  framework  GPU  GPGPU  CPU  yarn  scheduling  MPI  RDMA  infiniband  yahoo  opensource  software  Delicious 
february 2016 by asteroza
Scylla DB
Interesting, sharding a DB down to the CPU level, such that inter-CPU comms is functionally the same as inter-server, and is thus massively parallelizable, at the expense of per thread scheduling complexity.
ScyllaDB  Cassandra  seastar  C++  rewrite  NoSQL  database  software  Delicious  CPU  sharding  high  performance 
september 2015 by asteroza
Google Will Do Anything To Beat Moore's Law
Interesting... Power8 is little endian, not big endian, so porting stuff got a lot easier...
google  PowerPC  Power8  power  IBM  OpenPower  CPU  hardware  electronics  devices  Delicious 
may 2015 by asteroza
OpenCL™ Drivers and Runtimes for Intel® Architecture
CPU drivers for OpenCL to run directly on intel CPU's, which is sometimes faster than native C++ code...
intel  CPU  OpenCL  runtime  SDK  software  download  windows  linux  Delicious 
january 2015 by asteroza
CoolChip Technologies Inc. | The Kinetic Cooling Company
The fan IS the heatsink, I think, though how heat gets across the moving interface is an interesting problem
CPU  cooling  heatsink  thermal  transfer  fan  hardware  electronisc  devices  Delicious 
december 2014 by asteroza
Computer Laboratory: BERI open-source hardware downloads
CHERI is built on the MIPS ISA, but concepts are portable to RISC-V and ARMv8.
CHERI  BERI  secure  processor  FPGA  softcore  opensource  hardware  CPU  design  security  memory  protection  Delicious 
july 2014 by asteroza
1U Rackmount Server
Short depth Atom CPU based 1U rackmount servers, for when you are length constrained.
atom  CPU  1U  rackmount  server  hardware  electronics  devices  small  form  factor  depth  length  short  Delicious 
september 2011 by asteroza
VMware KB: Setting the number of cores per CPU in a virtual machine
Apparently a secret trick for enabling multicore style vCPU's. Basically, this would be a brute force way of providing a desktop OS with more CPU (since they a license limited to two CPU sockets)
VMware  ESXi  vCPU  CPU  multicore  VMX  setting  configuration  VM  sysadmin  tips  tricks  Delicious 
january 2011 by asteroza
Intel launches FPGA-equipped Atom |
Cute, but the FPGA is linked by PCIexpress x1 which is a narrow pipe, so it kinda defeats the purpose
intel  CPU  atom  FPGA  altera  hardware  electronics  devices  integrated  SoC  embedded  Delicious 
november 2010 by asteroza
A compact pseudo-blade cluster enclosure with 512 atom processor powered servers inside. Needless to say, a SAN is required. But, with Atom CPU's , this has very low power draw.
SeaMicro  SM10000  small  form  factor  compact  intel  atom  CPU  cluster  blade  server  hardware  electronics  devices  green  IT  Delicious 
august 2010 by asteroza
[CPU性能]smallなのに3GHz相当のCPU - 検証!ニフティクラウド:ITpro
So, the baseline 3GHz VCPU helps out the singlethreaders, but EC2 still rules the multithreading roost for now? That, and the smallest instance is gimped hard. In a way it makes sense, since Nifty will be involved with moving various business/enterprise customer internal apps, which were likely written single threaded ages ago, to their cloud as new installs (or P2V?) since japanese IT is forever wed to the idea of heavily customized software (which keeps the SI'ers in the gravy train). That, and their cloud is a brand new datacenter with fresh hardware, compared to Amazon which has servers stretching back 3 years in vintage (the singapore zone is pretty new though, but the article makes no mention of the zone used in testing, which could make a big difference). Though once Amazon opens a japanese AWS zone, it's game over.
Nifty  Amazon  faceoff  benchmark  CPU  VCPU  virtualization  VM  cloud  service  performance  japan  Delicious 
june 2010 by asteroza

related tags

1U  10U  11.3  100%  acceleration  ACME  adobe  AES  AI  altera  alternative  amazon  AMD  analysis  analyzer  anonymity  apple  application  AR1688  Archipelago  Area  ARM  ASIC  assembly  atom  AWS  benchmark  benchwork  BERI  BlackFin  blade  blocks  board  brick  bug  building  Bull  burst  bypass  bzip  BZIP2  C++  Caffe  Calao  cap  cassandra  CHERI  chinese  chip  chipset  circuit  circuits  clock  cloud  cluster  collaborative  comaprison  community  compact  comparison  compatibility  compiler  compound  compression  computer  computing  concept  configuration  control  coolbook  cooler  cooling  core  count  cpu  cpushare  credential  cryptography  custom  cycle  daemons  database  datacenter  DB  debugging  decentralized  decompression  deep  Delicious  demand  density  depth  design  detection  development  device  devices  die  disk  distributed  DIY  domain  download  DSP  EC2  efficiency  electronics  electronisc  embedded  encrypted  encryption  engine  engineering  english  ESXi  exchanges  execution  exploit  fab  fabless  faceoff  factor  FAIL  fan  film  firefox  flash  flexible  form  Fox  FPGA  framework  free  frequency  futures  gallery  GCC  geospatial  GIS  google  GP  GPGPU  GPL  GPU  grease  green  grid  guard  hack  hacking  hackinig  hadoop  hardware  heatpipe  heatsink  hidden  high  history  hoftix  host  hosting  HPC  HVCI  Hyper-V  hypervisor  i7  IAX  IBM  ICC  idle  infiniband  information  Ingres  instruction  integrated  intel  intelligence  internet  internet-of-things  IOPS  IoT  ISA  IT  japan  java  JOP  JVM  kb975530  Laboratory  laptop  lasso  learning  length  library  limit  limiter  linux  liquid  Logical  lot  low  lowRISC  mac  machine  management  manufacturing  market  marketplace  Marvell  McCaslin  memory  menumeters  metal  metapad  Micro  microcontroller  microprocessor  microsoft  MID  MID2007  MIPS  mobile  mode  modular  monitoring  motherboard  MPI  multi  multicore  multiprocessor  multithread  multithreaded  MXNet  nehalem  network  neural  New  Nifty  NoSQL  noSQL  notebook  NovaScale  NUMA  Nvidia  on  onboard  open  OpenCL  opencores  OpenPower  OpenRD  OpenRD-client  opensource  opteron  optimization  OS  OSX  overclock  p2p  Palm  parallel  parasite  patch  PBZIP2  PC  PCB  pentesting  performance  personal  plastic  player  portable  power  Power8  PowerPC  printed  priority  privacy  private  process  processing  processor  profiler  profiling  program  programming  protected  protection  pumpless  quad  quota  R422  rack  rackmount  RDMA  RedFlag  reference  research  resource  resources  result  reverse  rewrite  RISC  RISC-V  rowhammer  RTP  runtime  SaaS  SafeEnd  SanDisk  SBC  scheduling  ScyllaDB  SDK  SeaMicro  seastar  secure  security  semcionductor  semiconductor  server  service  services  set  setting  sharding  shared  sharing  short  SiCortex  SilentFLux  SIP  sizing  skew  sleep  SM10000  small  SoC  socket  softcore  software  soulpad  SPARC  Spark  speculative  SSD  startup  state  Stealey  STL  Stock  storage  Sun  supercomputer  sysadmin  system  technology  temperature  terminal  thermal  thread  threadmaster  time  timestamp  tips  tools  topological  tor  trading  transfer  tricks  TrustFlash  TS  UDP  UMPC  undervolt  undervolting  usage  USB  user  utilities  VCPU  vCPU  VDI  vector  VectorWise  version  VHDL  virtual  virtualization  visualization  visualstudio  VM  VMware  VMX  vNUMA  VoIP  VT-x  WALLS  Wide  windows  workstation  x64  x86  x86-64  Xen  Xeon  yahoo  yarn  York 

Copy this bookmark: