boinc - enhancing research workloads for the benefit of mankind & humanity - Computer Optimization - CPU & GPU
HPC - High Performance Computation for beneficial goals and obvious worth.
(Guide, experimentation, developer kit's and manuals)
By Rupert S
何百万のコアで何をするのですか?
混乱した毛穴から血が流出するまで、混乱の罠から惑星を救いなさい。
永遠の海のイルカのような時間の川で踊りましょう。
夕方の海岸まで科学の蝋燭をちらつかせる。
what would we do with a million cores!?
save a planet from the grip of chaos death till the blood runs from shattered pores..
dance in the rivers of time like the dolphins in the seas of evermore..
flicker the candle of science till evening shores.
*
Observing the workloads of many beneficial projects we find that commonly the workload data set is small,
In addition to the memory set being smaller or larger than a machine can compute optimally; we find that feature sets such as fae and avx have commonly not been implemented,
Some projects like asteroids at home and the seti project are using enhanced computation instruction sets ... like avx and memory loads that benefit from the 4gb or more ram that is available on decent gaming and home laptops.
Not all modern machines have loads of ram; However research and or university establishments use sufficiently powerful machines that can glow on the boinc record in full glory with a 256mb to 768mb workload,
In addition the machines are operand,xen ... commonly and servers may have such as Sparc or power pc specific hardware and instruction sets,
In order to examine examples .. below we can see workloads include small data arrays; in the 40mb to 79mb range..
In line with servers and gaming rigs .. we have 1gb of ram per core, of course not all issues require a larger array in the workload and some machines have 256mb per core !
However much Ram you allocate to the projected workload; small memory loads can and will be sufficient for data swapping and or paging (like DNA Replicators)...
Some task can sufficiently benefit from larger thread and data models, to my mind DNA and mapping data are fine examples of specific workloads; Where memory counts,
In addition thread count can be 4 or other numbers and i suggest that a single task can use more than one core and instruction set (neon for example or Symmetric threading FPU, SMT)
Specific workload optimisation, or rather generic with SSE and AVX and FPU threading and precision optimisation would be very cool while we deal with the workload running app.
In particular the Ryzen multi-core is a new and exciting product,
So take care to read the guides in the lower half of the document, AVX2, RDSEED, ADX and additional encryption formats are some of the most exciting changes to the AMD Ryzen Arch.
Showing the problems that properly optimising code for Chemical/Biological examination can face.
AVX similarities to GPU core, Function of AVX can be thought of as CPU extension function of the same usage as GPU!
In short combined with FPU very much in the same performance category as the GPU cores and of much worth to scientific research and development of game dynamics, sound, video and spaces in N-Dimension space.
CPU extensions can prepare vector space for GPU to enhance the speed and optimize vector tables before GPU rendering and sound space in 3D for surround sound...
Interpolate texture, sound and other data with bit swapping.. In SIMD instructions.
RND Function can be used to explore additional data spaces.
Encryption function to enhance unpredictable behavior or to save space.
Further thought ... Efficiency :
add a MHz/Dhrystone's/MIP'S performance per watt to each system ...
then projects will further optimise workloads to improve upon workload energy & environmental efficiency versus work carried out.
Work Hours x Mhz / (efficiency per watt)
-------
Hours / % of projects finished with work completed
Also bear in mind that GPU's need watt efficiency and task management to optimise power used versus work done....
worker priority should always be :
efficiency + merit of the work
--------
time / % necessity
Please examine the issue further.
Rupert S
https://www.worldcommunitygrid.org
https://boinc.berkeley.edu/
http://www.charityengine.com/
https://lhcathome.cern.ch/https://cern.n-helix.com/lhcathome/cpu_list.php
CERNVM-FS-Both : Run & Install Commands : RS
https://is.gd/CERN_SH_Scripts
https://cvmfs.readthedocs.io/en/stable/cpt-quickstart.html
https://cvmfs.readthedocs.io/en/stable/cpt-configure.html
HPC Computing work load Photos -
HPCSet 2 photos -
HPC Set 3 Photos
Conducting Research Photo set 1 -
Photo set 2 -
photo set 3
http://esa-space.blogspot.ru/2017/04/rng-and-random-web.html - we need Chaos Seeds : Random seeds for our work
https://community.amd.com/thread/213045 - particular instruction differences for microcode optimisation
http://32ipi028l5q82yhj72224m8j.wpengine.netdna-cdn.com/wp-content/uploads/2017/03/GDC2017-Optimizing-For-AMD-Ryzen.pdf - code optimisation a few very important lessons... may seem simple to some but obviously is not to be taken for granted.
CPU Optimisation - utility and function.
http://www.noamross.net/blog/2013/4/25/faster-talk.html - speeding up code a guide - profiling and bench-marking.
http://www.pgroup.com/doc/pgi17ug-x64.pdf - PGI Compiler guide
http://www.agner.org/optimize/ - code optimisation for all programmers on X86,X86-64bit and some others.. this is a terrific resource !
http://www.agner.org
https://github.com/ctuning/ck - data & program - testing and tuning
for example : Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 htt pni ssse3 fma cx16 sse4_1 sse4_2 popcnt aes f16c syscall nx lm avx sse4a osvw xop wdt fma4 topx page1gb rdtscp bmi1
or for example : Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 htt pni ssse3 fma cx16 sse4_1 sse4_2 popcnt aes f16c syscall nx lm avx svm sse4a osvw ibs xop skinit wdt lwp fma4 tce tbm topx page1gb rdtscp bmi1
or for example : Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 htt pni ssse3 fma cx16 sse4_1 sse4_2 popcnt aes f16c syscall nx lm avx sse4a osvw xop wdt fma4 topx page1gb rdtscp bmi1
for an improved upon instruction list in the newer boinc application.. (with appropriate configuration)
11000 Mips & 2700 FPU Mips - Per Core
**
an article that took some deep learning... itself ôo, anyway very interesting....
hip c++ will we think be simpler than open CL then as a higher level code port...
and machine converted CUDA-code to 99.6%
http://llvm.org/
http://llvm.org/docs/FAQ.html
https://gcc.gnu.org/
*not free obviously .. intel*
https://software.intel.com/en-us/articles/intel-advisor-roofline
**
*compilers with FORTRAN specifics and preferably C/C++ and HPC (compatibility C++/C compatible with FORTRAN preferably)
https://gcc.gnu.org/wiki/HomePage
https://gcc.gnu.org/wiki/GFortranBinaries
https://software.intel.com/en-us/intel-parallel-studio-xe/try-buy/#parallelstudioxe
http://www.pgroup.com/products/pgiworkstationg.htm (limitations nVidia compatable GPU Cuda code & no obvious statment of OpenCL Support)
http://llvm.org/ - llvg it seems has fortran compatibility.. (needs research)
http://llvm.org/docs/FAQ.html
http://www.pathscale.com/ - check it out
Fortrans Speacialists (no c++ etcetera)
https://www.absoft.com/products/windows-fortran-compiler-suite/
http://www.fortran.com/products-page/compilers/fortrantools-for-windows/
https://www.cs.sfu.ca/~fedorova/Teaching/CMPT886/Spring2007/papers/adaptive-execution.pdf
*ibm guidance*
http://www.prace-ri.eu/best-practice-guide-ibm-power-775-html/
https://www.redbooks.ibm.com/redbooks/pdfs/sg248280.pdf
Release code to use Power chips and emulation code embeded in boinc mainframe 800 Core Multiplex
https://access.redhat.com/articles/3158511 - Power9 Edition RedHat
https://www.hpcwire.com/off-the-wire/ibm-releases-new-compilers-exploit-power9-technology/
**
PC/Mac/Windows/Linux/Android - high performance computation - the method and the means
http://science.n-helix.com/2018/09/hpc.html
http://science.n-helix.com/2018/09/hpc-pack-install-guide.html
https://www.khronos.org/news/events/2016-isc-high-performance
https://www.khronos.org/assets/uploads/developers/library/2008_siggraph_bof_opengl/OpenCL%20and%20OpenGL%20SIGGRAPH%20BOF%20Aug08.pdf HPC Report
*
http://www.ziti.uni-heidelberg.de/ziti/uploads/ce_group/2017-ISC.pdf - Overview of MPI message characteristics of HPC Server proxy applications.
*Interesting statistics from which one can conclude that 64 to 256 core units is the space within which,
The maximum increase in message noise/entropic noise; Related to inter process communication is observed.*
https://www.microsoft.com/en-us/download/details.aspx?id=54507 Microsoft HPC Pack 2016 including linux
https://technet.microsoft.com/en-us/library/cc514029(v=ws.11).aspx all HPC Packs 2016,2012 to 2008 info and download
https://msdn.microsoft.com/en-us/library/ff976568.aspx Microsoft High Performance Computing for Developers - info and downloads
https://docs.microsoft.com/en-us/azure/virtual-machines/windows/hpcpack-cluster-active-directory - information and virtualisation
OpenVX for high performance Computing : Multi platform spec
"OpenVX for HPC Neural Nets and processing .... a new way to deliver on research, gaming & processing of data and images"
https://www.khronos.org/news/tags/tag/OpenVX
https://www.khronos.org/news/press/openvx-1.2-specification-cross-platform-acceleration-power-efficient-vision
https://www.ibm.com/blogs/research/2017/12/pruning-ai-networks/
https://arxiv.org/abs/1611.05162v4 - net-trim
Somewhat over complex formula..
Considering that the objective is to trim the network to as few as necessary nodal's to..
Reduce complexity and improve performance.
(May want to net-trim verbose complexity out of science and code generation.)
**
Open CL "GPU Development" links
https://www.khronos.org/blog/iwocl-where-you-learn-the-latest-on-opencl
https://www.khronos.org/opencl/
https://www.khronos.org/opencl/resources for SDK, learning & optimisation resources.
http://developer.amd.com/tools-and-sdks/opencl-zone/amd-accelerated-parallel-processing-app-sdk/opencl-optimization-guide/
https://github.com/RadeonOpenCompute - ROCm: Platform for GPU Enabled HPC and UltraScale Computing
http://gpuopen.com/professional-compute/
http://gpuopen.com/compute-product/hcrng/
https://bitbucket.org/multicoreware/hcrng
http://gpuopen.com/compute-product/clrng/
installing the AMD SDK improves compute performance, Optimise your code !
https://streamhpc.com/blog/2017-05-21/amd-open-sourced-rocms-opencl-driver-stack/
https://github.com/RadeonOpenCompute/ROCm-OpenCL-Runtime/blob/amd-master/README.md
http://developer.amd.com/tools-and-sdks/opencl-zone/
http://developer.amd.com/tools-and-sdks/opencl-zone/amd-accelerated-parallel-processing-app-sdk/
http://gpuopen.com/games-cgi/
http://developer.amd.com/tools-and-sdks/graphics-development/
http://hgpu.org information and interesting learning & source
http://dspace.princeton.edu/jspui/bitstream/88435/dsp01wm117r22g/1/Jia_princeton_0181D_11168.pdf Optimisation for parallel computing information.
https://arxiv.org/pdf/1705.05249 - CLBlast: A Tuned OpenCL BLAS Library demonstration.
https://arxiv.org/pdf/1710.08616
https://arxiv.org/pdf/1710.08616.pdf - FORTRAN for GPU and multiprocessor usage in Scientific research,
Also of interest in the generation of coding Format, style, implementation & Structure.
"The new implementation performs up to 4.9x faster when comparing one GPU to one
multi-core CPU socket. On a full-scale production run with 1581 x 1301 x 58
grid size and 2km resolution, 24 Tesla P100 GPUs are shown to replace more
than 50 18-core Broadwell Xeon sockets."
"GPUs are an attractive target architecture, with a memory bandwidth that is
typically 5 to 7 times higher than Intel Xeon architectures of a similar generation."
"Compared to CPUs, GPUs support a very high number of parallel threads while
having a very low thread switching overhead - however with the cost of small
caches available per thread and a low single-threaded performance."
LHC Cern 6 Track GPU Study < help needed...
https://lhcathome.cern.ch/lhcathome/index.php - coders desired.
RS
**
HIP - HSA - the CUDA Compatible C++ for Heterogeneous Computing
http://developer.amd.com/wordpress/media/2012/09/7637-HIP-Datasheet-V1_4-US-Letter.pdf
http://developer.amd.com/wordpress/media/2012/10/hsa10.pdf - a full guide
http://www.hsafoundation.com/
http://www.hsafoundation.com/hsa-developer-tools/
https://github.com/HSAFoundation/HSA-docs-AMD/wiki#initial-implementation
https://github.com/HSAFoundation/HSAIL-Tools
https://github.com/RadeonOpenCompute/ROCK-Kernel-Driver - Driver for kernel
http://www.amd.com/Documents/SDN-Whitepaper.pdf - Smart Software Defined Networks
http://support.amd.com/TechDocs/55766_SEV-KM%20API_Spec.pdf - Secure Encrypted Virtualisation Key Management
http://support.amd.com/TechDocs/Protecting%20VM%20Register%20State%20with%20SEV-ES.pdf - PROTECTING VM REGISTER STATE WITH SEV-ES
http://support.amd.com/TechDocs/50742_15h_Models_60h-6Fh_BKDG.pdf - bios and kernel drivers
**
Machine Intelligence code optimization platforms
https://www.tensorflow.org/ - machine intelligence
https://github.com/tensorflow/tensorflow
https://github.com/hughperkins/tf-coriander - openCL Tensor flow
PyTorch - Machine learning with graphs, Tesor philosophie and python -
https://github.com/pytorch/pytorch -
http://pytorch.org
Hyperdash python SDK - PyTorch
https://github.com/hyperdashio/hyperdash-sdk-py
Richard Herbert real time learning with PyTorch - Real-time Machine Learning with PyTorch and Filestack
https://blog.filestack.com/tutorials/realtime-machine-learning-pytorch/
"Kirill DubovikovFollow - Knowledge distiller, Data Scientist and Software Architect"
https://medium.com/towards-data-science/pytorch-vs-tensorflow-spotting-the-difference-25c75777377b
speed and data comparison
https://medium.com/@yaroslavvb/tensorflow-meets-pytorch-with-eager-mode-714cce161e6c
**
ARM Development software/SDK's & tools
https://developer.arm.com/products/software-development-tools
https://developer.arm.com/products/software-development-tools/hpc for high performance computing (ideal for Boinc)
https://developer.arm.com/products/software-development-tools/compilers for both HPC and APP development.
https://developer.arm.com/products/system-design/fixed-virtual-platforms
https://www.synopsys.com/verification/virtual-prototyping/vdk/vdk-for-arm.html
https://www.synopsys.com/designware-ip/technical-bulletin/designware-hybrid-ip.html
**
**
IOT links - (internet of things)
https://www.infoq.com/articles/thread-protocol-for-home-automation
http://wso2.com/wso2_resources/wso2_whitepaper_a-reference-architecture-for-the-internet-of-things.pdf
**
compiler optimisation - process
https://crd.lbl.gov/departments/computer-science/PAR/research/roofline/
https://www.nextplatform.com/2017/05/25/nersc-supercomputing-site-eases-path-optimization-scale/
https://www-ssl.intel.com/content/www/us/en/events/hpcdevcon/parallel-programming-track.html#utilizing
**
Linux arch reference material
https://www.ibm.com/developerworks/library/l-linuxuniversal/
**
Agency GPL
https://code.nasa.gov/
Workers :
https://www.upwork.com/hire/driver-development-freelancers/
Update 2:
for a comparison of Gflops/Mips throughput of various Boinc Tasks ..
here we show the relevance of the code or function used ... AVX for example is multi threaded ! and so is the FPU pipeline of the AMD FX & Ryzen processor.....
http://bit.ly/HPCImpact (original non edited photos ...)
and set 2 (newer)
http://bit.ly/2HPCImpact ....
set 3
http://bit.ly/HPCImpact2 to examine of the improvement code streamlining brings.
Some of our work with the updated graphics
http://bit.ly/ReserchPhotos
see the work throughput GFlops compared to code efficiency per task !
sometimes entropy is needed to for-fill the task one would imagine (for example on android)
http://bit.ly/tRNG-Dev
the improvement of the
boinc and
worldcommunitygrid projects has been observed, noted and one feels improved upon, ..
further improvement should be implemented as soon as possible; To improve work versus output efficiency.
thank you kindly programmers/Workers & scientists for your perseverance & effort.
RS
http://bit.ly/BoincStudies - Result Studies
Update 3 Q & A:
"In reference to the use of virtual box there is a new product by berkley >
http://singularity.lbl.gov/ called singularity that handles repeatable condition containers... and has low overhead for virtualisation data-set.
As to the particle spread one should possibly consider the multiple core and threaded core model specific to the Ryzen and intel sets...
One could imagine that the multi-threaded nature of arm server cores combined with the nature of multi-threaded and headed arm CPU's and GPU Run-script environments is a new and uncompromising land of opportunity and challenge.
Many of the instructions on the FMV4 and Vector instruction sets have multi-threaded en-action at lower precision..."
http://fife.fnal.gov/singularity-on-the-osg/
RS
----
Eric Mcintosh accredited scientist Cern
Project administrator
Project developer
Project tester
Project scientist
"Well we are far from trying to optimise GPU code.
First let me explain that we have a tracking loop over turns
(up to 1,000,000 hoping for 10,000,000 soon) which contains
a large number of inner loops over particles, currently up to 64.
Luckily these loops over particles can be paralleled as each
particle is totally independent. In addition the original author F. Schmidt
pre-calculated everything possible before entering the tracking loop.
Each turn involves some 10,000 steps over a varying number of inner loops,
e.g. straight section, quadruple, beam-beam interaction, power supply ripple, etc etc
Of which there are about 50 different possibilities. A straight section is really just
a multiply and add, whereas beam beam involves hundreds or more FLOP's.
The first idea would be to use a much larger number of particles to best
utilise the GPU. This however would produce a large amount of I/O and
use a lot of disk space, but maybe not insurmountable,
However all the code is FORTRAN, the outer loop calls subroutines (could inline), and has many tests/branches.
It would be great if the main loop fitted entirely into the GPU and we would have
rare Host access for I/O or BOINC checkpoint and progress calls or when
one or more particles are lost.
My colleague Ricardo is actively looking at redoing in C which would also allow
much more portability and also allow to be parallel on multi-core systems.
For the moment we just run tasks in parallel, which works rather well (apart
from some current infrastructure problems). I hope to come up with
some numbers next week on GPU testing.
The code itself has been regularly measured and optimised; for example we
re-ordered array indices to optimise memory access and rewrote the Error Function
of a Complex Number to be faster but with adequate precision.
Portability does come at a price but ensures accuracy of results. I shall publish
measurements in an upcoming paper. I am sure we gain much more from being portable
and being able to use almost any IEEE 754 compliant processor.
On the issue of SixTrack and/or experiments this will shortly be under discussion at
CERN I am sure. Currently SixTrack has many more Hosts/volunteers, is simple to install,
and has been around for 13 years. Not everyone loves VMbox. Not a big deal at
present as we rarely have enough SixTrack work to keep all volunteers busy.
I hope to re-address all this in some weeks after current BOINC infrastructure issues
are resolved and we have the new "super" sixtrack with much broader application
e.g.collimation studies and we support a much wider range of platforms MacOS ARM
and use features such as AVX.
Eric.
____________"
Update 4 : Virtualisation
QEMU is obviously be of use on many projects because of machine emulation and virtualisation..
Comes in flavours including Windows, Mac and Linux.
*
Docker Sever & Docker CE (community edition) and this comes with sever edition!
So what do the projects & system.. feel and sense around the subject of using Docker CE ?
Obviously the professional version could be used for support of the main project and the CE edition or pro for the user..
https://store.docker.com/editions/community/docker-ce-desktop-windows
https://store.docker.com/search?offering=community&q=&type=edition
https://www.ctl.io/developers/blog/post/what-is-docker-and-when-to-use-it/
https://www.digitalocean.com/community/tutorials/how-to-install-and-use-docker-getting-started
https://www.howtoforge.com/tutorial/how-to-use-docker-introduction/
VM and Microprocessor bug fixes incoming..
Hopefully microcode quickly also.
Creating a better virtualization header that is:
More efficient at isolating the contained OS with attributes in the OS's to contain secured data?
We find answers to improve efficiency and protect against VM>VM data transfer or to use this for a creative purpose!
We need answers! and science. : Microcode update
"Thank you for googles firm responses to the bug, faith in google is high..
The micro code be updated to flush & or contain the the speculative data in a data-cycle secure storage,
Within the framework of cache and ram/virtual-ram?
Cycle efficiency would be at most two cycles and a flush Xor bit data overlay,
Bit Masking before and after pre-fetch presents & also uses data - this method would be fast! (c)Rupert S"
Google systems have been updated for Meltdown bug
https://security.googleblog.com/2018/01/todays-cpu-vulnerability-what-you-need.html
Attack mitigation -
https://support.google.com/faqs/answer/7622138#android
"Microsoft issued an emergency update today,
Amazon said it protected AWS customers running Amazon's tailored Linux version and will roll out the MSFT patch,
for other customers to day"
We need answers! and science. : Microcode : update
(c)RS
specter & meltdown information
**
how to convert VM's and use hyper V and Docker