Skip to content

Instantly share code, notes, and snippets.

View js947's full-sized avatar
💭
🐝

Jeffrey Salmond js947

💭
🐝
View GitHub Profile
@js947
js947 / CSD3-early_access.md
Created August 7, 2017 21:46 — forked from fspiga/CSD3-early_access.md
EPSRC Tier-2 "Cambridge Service for Data Driven Discovery" (CSD3) Pump-priming / Proof of Concept access

CSD3 Pump-priming / Proof-of-concept access

The Tier-2 "Cambridge Service for Data Driven Discovery" (CSD3) national service will start with the first Tier-2 Open Call. Check the EPSRC website for more information about the Open Calls, based on what I know it is going to be 12nd October and the technical assessment due on 21st September (a process very much like ARCHER applications).

At the current stage there are two clusters available for access:

  • a GPU-accelerated cluster made of 360 P100 called "Wilkes -2" (~1.2 PFlop/s)
  • a many-core cluster made of 340 Intel Xeon Phi called "Peta4-KNL" (~0.5 TFlop/s)

During this Pump-priming / Proof-of-concept access (which will always be open) we accept requests for access the systems to help people to prepare their applications. It is very important to point out that the Pump-priming / Proof-of-concept access is not for running science. For this reason, the allowance will be limited (~20k GPU hours and ~20k KNL hours). Expansions of this allowance are compl

Keybase proof

I hereby claim:

  • I am js947 on github.
  • I am js947 (https://keybase.io/js947) on keybase.
  • I have a public key whose fingerprint is 435C 0427 9D43 A37F 6A55 3EF3 8429 9AC1 65EC 2941

To claim this, I am signing this object:

@js947
js947 / clinfo.py
Last active October 27, 2016 09:32
Print opencl platform and device info
#!/usr/bin/env python
from __future__ import print_function
import pyopencl as cl
for p in cl.get_platforms():
print(p)
for d in p.get_devices():
print(d)
print(d.get_info(cl.device_info.NATIVE_VECTOR_WIDTH_FLOAT))
@js947
js947 / frk.js
Last active August 29, 2015 14:02
function frk(a, r, b) {
return function(x) {
return r(a(x), b(x))
}
}
function sum(a) {
return a.reduce(function(acc, val) {
return acc + val
}, 0)
@js947
js947 / frk.cpp
Last active August 29, 2015 14:02
#include <vector>
#include <functional>
#include <numeric>
#include <iostream>
using namespace std;
template<typename T>
function<T(vector<T>)> frk(
function<T(vector<T>)> a,
function<T(T,T)> r,
@js947
js947 / frk.ru
Last active August 29, 2015 14:02
def frk a, r, b
lambda do |x|
r.call(a.call(x), b.call(x))
end
end
def sum a
a.reduce(0) do |sum, value|
sum + value
end
@js947
js947 / frk.py
Last active August 29, 2015 14:02
import operator
def frk(a, r, b):
return lambda x: r(a(x), b(x))
avg = frk(sum, operator.div, len)
ans = avg([1.0, 2.0, 3.0, 4.0])
@js947
js947 / frk.jl
Last active August 29, 2015 14:02
function frk(a, r, b)
x -> r(a(x),b(x))
end
avg = frk(sum, /, length)
ans = avg([1.0, 2.0, 3.0, 4.0])