Add hpcc benchmark linkies
Jeff Moe
parent
6f06b0371d
commit
32f30f4cfb
|
|
@ -8,11 +8,63 @@ top500
|
|||
======
|
||||
In what year would this be the world's fastest computer?
|
||||
|
||||
https://top500.org
|
||||
`<https://top500.org>`_
|
||||
|
||||
HPC Challenge benchmarks.
|
||||
|
||||
`<https://hpcchallenge.org/hpcc>`_
|
||||
|
||||
Note, the HPC Challenge benchmarks are in the Debian ``hpcc`` package.
|
||||
|
||||
Linpack
|
||||
-------
|
||||
The Linpack TPP benchmark "measures the floating point rate of execution for solving a linear
|
||||
system of equations."
|
||||
|
||||
`<https://www.netlib.org/benchmark/hpl>`_
|
||||
|
||||
DGEMM
|
||||
-----
|
||||
DGEMM "measures the floating point rate of execution of double precision real matrix-matrix multiplication."
|
||||
|
||||
STREAM
|
||||
------
|
||||
STREAM is "a simple synthetic benchmark program that measures sustainable memory bandwidth (in GB/s)
|
||||
and the corresponding computation rate for simple vector kernel."
|
||||
|
||||
`<https://www.cs.virginia.edu/stream>`_
|
||||
|
||||
PTRANS
|
||||
------
|
||||
PTRANS (parallel matrix transpose) "exercises the communications where pairs of processors
|
||||
communicate with each other simultaneously. It is a useful test of the total communications
|
||||
capacity of the network."
|
||||
|
||||
`<https://www.netlib.org/parkbench/html/matrix-kernels.html>`_
|
||||
|
||||
RandomAccess
|
||||
------------
|
||||
"RandomAccess measures the rate of integer random updates of memory (GUPS)."
|
||||
|
||||
`<https://hpcchallenge.org/projectsfiles/hpcc/RandomAccess.html>`_
|
||||
|
||||
FFT
|
||||
---
|
||||
"FFT measures the floating point rate of execution of double precision complex
|
||||
one-dimensional Discrete Fourier Transform (DFT)."
|
||||
|
||||
`<http://www.ffte.jp>`_
|
||||
|
||||
Communication Bandwidth and Latency
|
||||
-----------------------------------
|
||||
Communication bandwidth and latency is "a set of tests to measure latency and bandwidth of a
|
||||
number of simultaneous communication patterns; based on b_eff (effective bandwidth benchmark)."
|
||||
|
||||
`<https://fs.hlrs.de/projects/par/mpi/b_eff>`_
|
||||
|
||||
|
||||
tinygrad Examples
|
||||
=================
|
||||
tinygrad
|
||||
========
|
||||
Benchmarks in tinygrad.
|
||||
|
||||
mlnotcommons
|
||||
|
|
@ -21,25 +73,22 @@ Proprietary with a few libre datasets and benchmarks available.
|
|||
|
||||
Don't let "Commons" in the name lead you to think this is available to the mere public.
|
||||
Lots of proprietary bits involved, closed lists, corporate signups and signatures, etc.
|
||||
|
||||
This isn't like Wikipedia Commons at all.
|
||||
|
||||
Their use of "Commons" in their name perhaps causes confusion in the marketplace
|
||||
with Wikipedia Commons (and other groups that serve the public).
|
||||
This isn't like Wikipedia Commons at all.
|
||||
|
||||
The upstream tinycorp is working on implementing some of their benchmarks using
|
||||
``tinygrad`` and AMD GPUs.
|
||||
|
||||
`<https://mlcommons.org/datasets>`_
|
||||
|
||||
https://mlcommons.org/datasets/
|
||||
`<https://mlcommons.org/benchmarks>`_
|
||||
|
||||
https://mlcommons.org/benchmarks/
|
||||
|
||||
https://github.com/mlcommons
|
||||
`<https://github.com/mlcommons>`_
|
||||
|
||||
|
||||
``examples/mlperf/model_spec.py``
|
||||
---------------------------------
|
||||
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
||||
Test run, just hitting one GPU.
|
||||
|
||||
.. literalinclude:: _static/_output/tinygrad/mlperf-model-spec.py.txt
|
||||
|
|
|
|||
|
|
@ -9,7 +9,7 @@ msgid ""
|
|||
msgstr ""
|
||||
"Project-Id-Version: tinyrocs: Direct to Chip Liquid Cooled GPU AI Cluster 0\n"
|
||||
"Report-Msgid-Bugs-To: \n"
|
||||
"POT-Creation-Date: 2024-02-07 10:26-0700\n"
|
||||
"POT-Creation-Date: 2024-02-07 11:05-0700\n"
|
||||
"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
|
||||
"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
|
||||
"Language: en\n"
|
||||
|
|
@ -37,64 +37,165 @@ msgid "In what year would this be the world's fastest computer?"
|
|||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:11
|
||||
msgid "https://top500.org"
|
||||
msgid "`<https://top500.org>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:13
|
||||
msgid "HPC Challenge benchmarks."
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:15
|
||||
msgid "tinygrad Examples"
|
||||
msgid "`<https://hpcchallenge.org/hpcc>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:16
|
||||
msgid "Benchmarks in tinygrad."
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:19
|
||||
msgid "mlnotcommons"
|
||||
#: ../../../_source/benchmarks.rst:17
|
||||
msgid "Note, the HPC Challenge benchmarks are in the Debian ``hpcc`` package."
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:20
|
||||
msgid "Proprietary with a few libre datasets and benchmarks available."
|
||||
msgid "Linpack"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:22
|
||||
#: ../../../_source/benchmarks.rst:21
|
||||
msgid ""
|
||||
"Don't let \"Commons\" in the name lead you to think this is available to the "
|
||||
"mere public. Lots of proprietary bits involved, closed lists, corporate "
|
||||
"signups and signatures, etc."
|
||||
"The Linpack TPP benchmark \"measures the floating point rate of execution "
|
||||
"for solving a linear system of equations.\""
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:25
|
||||
msgid "This isn't like Wikipedia Commons at all."
|
||||
#: ../../../_source/benchmarks.rst:24
|
||||
msgid "`<https://www.netlib.org/benchmark/hpl>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:27
|
||||
msgid ""
|
||||
"Their use of \"Commons\" in their name perhaps causes confusion in the "
|
||||
"marketplace with Wikipedia Commons (and other groups that serve the public)."
|
||||
msgid "DGEMM"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:30
|
||||
#: ../../../_source/benchmarks.rst:28
|
||||
msgid ""
|
||||
"DGEMM \"measures the floating point rate of execution of double precision "
|
||||
"real matrix-matrix multiplication.\""
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:31
|
||||
msgid "STREAM"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:32
|
||||
msgid ""
|
||||
"STREAM is \"a simple synthetic benchmark program that measures sustainable "
|
||||
"memory bandwidth (in GB/s) and the corresponding computation rate for simple "
|
||||
"vector kernel.\""
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:35
|
||||
msgid "`<https://www.cs.virginia.edu/stream>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:38
|
||||
msgid "PTRANS"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:39
|
||||
msgid ""
|
||||
"PTRANS (parallel matrix transpose) \"exercises the communications where "
|
||||
"pairs of processors communicate with each other simultaneously. It is a "
|
||||
"useful test of the total communications capacity of the network.\""
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:43
|
||||
msgid "`<https://www.netlib.org/parkbench/html/matrix-kernels.html>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:46
|
||||
msgid "RandomAccess"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:47
|
||||
msgid ""
|
||||
"\"RandomAccess measures the rate of integer random updates of memory (GUPS)."
|
||||
"\""
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:49
|
||||
msgid "`<https://hpcchallenge.org/projectsfiles/hpcc/RandomAccess.html>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:52
|
||||
msgid "FFT"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:53
|
||||
msgid ""
|
||||
"\"FFT measures the floating point rate of execution of double precision "
|
||||
"complex one-dimensional Discrete Fourier Transform (DFT).\""
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:56
|
||||
msgid "`<http://www.ffte.jp>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:59
|
||||
msgid "Communication Bandwidth and Latency"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:60
|
||||
msgid ""
|
||||
"Communication bandwidth and latency is \"a set of tests to measure latency "
|
||||
"and bandwidth of a number of simultaneous communication patterns; based on "
|
||||
"b_eff (effective bandwidth benchmark).\""
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:63
|
||||
msgid "`<https://fs.hlrs.de/projects/par/mpi/b_eff>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:67
|
||||
msgid "tinygrad"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:68
|
||||
msgid "Benchmarks in tinygrad."
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:71
|
||||
msgid "mlnotcommons"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:72
|
||||
msgid "Proprietary with a few libre datasets and benchmarks available."
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:74
|
||||
msgid ""
|
||||
"Don't let \"Commons\" in the name lead you to think this is available to the "
|
||||
"mere public. Lots of proprietary bits involved, closed lists, corporate "
|
||||
"signups and signatures, etc. Their use of \"Commons\" in their name perhaps "
|
||||
"causes confusion in the marketplace with Wikipedia Commons (and other groups "
|
||||
"that serve the public). This isn't like Wikipedia Commons at all."
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:80
|
||||
msgid ""
|
||||
"The upstream tinycorp is working on implementing some of their benchmarks "
|
||||
"using ``tinygrad`` and AMD GPUs."
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:34
|
||||
msgid "https://mlcommons.org/datasets/"
|
||||
#: ../../../_source/benchmarks.rst:83
|
||||
msgid "`<https://mlcommons.org/datasets>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:36
|
||||
msgid "https://mlcommons.org/benchmarks/"
|
||||
#: ../../../_source/benchmarks.rst:85
|
||||
msgid "`<https://mlcommons.org/benchmarks>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:38
|
||||
msgid "https://github.com/mlcommons"
|
||||
#: ../../../_source/benchmarks.rst:87
|
||||
msgid "`<https://github.com/mlcommons>`_"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:42
|
||||
#: ../../../_source/benchmarks.rst:91
|
||||
msgid "``examples/mlperf/model_spec.py``"
|
||||
msgstr ""
|
||||
|
||||
#: ../../../_source/benchmarks.rst:43
|
||||
#: ../../../_source/benchmarks.rst:92
|
||||
msgid "Test run, just hitting one GPU."
|
||||
msgstr ""
|
||||
|
|
|
|||
|
|
@ -44,7 +44,8 @@ Install dependencies from Debian repositories.
|
|||
python3-venv python3-virtualenv python3-yaml quilt rsync rsyslog sshfs \
|
||||
sudo swig traceroute vim xxd python3-sphinx git-lfs hwdata \
|
||||
lua5.3 liblua5.3-dev libmpfr-dev libmsgpack-dev libfmt-dev \
|
||||
environment-modules python3-numpy pybind11-dev libopengl-dev zip zsh
|
||||
environment-modules python3-numpy pybind11-dev libopengl-dev zip zsh \
|
||||
hpcc
|
||||
|
||||
# Packages like this aren't used from Debian's repository.
|
||||
# Make sure they are gone.
|
||||
|
|
|
|||
Loading…
Reference in New Issue