dmikushin

// $ cat makefile 
// CUPTI = /opt/cuda/extras/CUPTI
// 
// all: cupti
//
// cupti: cupti.cu
//	nvcc -I$(CUPTI)/include -arch=sm_30 $< -o $@ -L$(CUPTI)/lib64 -lcupti -Xlinker -rpath=$(CUPTI)/lib64
//
// clean:
//	rm -rf cupti

dmikushin

template<class PriorityEvaluation, class TProcInfo>
class ProcInfoWithPriority : public TProcInfo
{
};
 
template<class TProcInfo, class TCustom = void>
class ProcInfoVisitor
{
};
 

dmikushin

// This sketch essentially checks if all elements of AVX vector are greater than zero
// In infinite loop we ensure equality of two implementations
//
// gcc -mavx cmp_check.c -o cmp_check -O3 -ffast-math

#include <assert.h>
#include <stdlib.h>
#include <stdio.h>
#include <x86intrin.h>

dmikushin

// AVX-512 horizontal multiply for k1om (Intel Xeon Phi Knights Corner)
//
// (c) 2016 Dmitry Mikushin [email protected]
//
// $ icc -mmic -std=c99 -O3 reduce_mul.c -o reduce_mul
// $ micnativeloadex ./reduce_mul
// -0.004276 vs -0.004276 

#include <immintrin.h>
#include <stdio.h>

dmikushin

// AVX-512 swizzle in native assembly for k1om (Intel Xeon Phi Knights Corner)
//
// (c) 2016 Dmitry Mikushin [email protected]
//
// $ icc -no-gcc -mmic -O3 -std=c99 test_swizzle.c -o test_swizzle
// $ micnativeloadex ./test_swizzle
// 1.000000 2.000000 3.000000 4.000000 5.000000 6.000000 7.000000 8.000000 
// 8.000000 8.000000 8.000000 8.000000 8.000000 8.000000 8.000000 8.000000

#include <immintrin.h>

dmikushin

Build LLVM-based C/C++ compiler with OpenMP4 for CUDA GPUs

Create working directory:

mkdir -p ~/forge/openmp4
cd ~/forge/openmp4

Clone LLVM & Compiler Runtime & Clang sources:

dmikushin

JNIEXPORT jobject JNICALL Java_..._getValue(
	JNIEnv *env, jobject ...)
{
  ...

	jobject resultObj = result.toJava().getObject();
	return resultObj; // Not NULL!
}

// Log shows:

dmikushin

From b4a79a41f995b27038011f92f0da3c6852a8596c Mon Sep 17 00:00:00 2001
From: Dmitry Mikushin <[email protected]>
Date: Wed, 13 Sep 2017 03:28:26 +0300
Subject: [PATCH] Make Alpine 2.21 to bounce all selected messages

---
 alpine/mailcmd.c |  5 -----
 alpine/reply.c   | 12 ++++++------
 2 files changed, 6 insertions(+), 11 deletions(-)

dmikushin

#ifndef TIMING_H
#define TIMING_H

#if defined(_WIN32)

#define CLOCK_REALTIME_WIN32 0
#include <windows.h>

struct timespec_win32
{

dmikushin

Welcome to Ubuntu 19.10!

[    7.111555] systemd[1]: Failed to bump fs.file-max, ignoring: Invalid argument
[    7.149252] systemd[1]: File /lib/systemd/system/systemd-journald.service:12 configures an IP firewall (IPAddressDeny=any), but the local system does not support BPF/cgroup.
[    7.150764] systemd[1]: Proceeding WITHOUT firewalling in effect! (This warning is only shown for the first loaded unit using IP firewalling.)
[    7.165346] systemd[1]: /lib/systemd/system/dbus.socket:4: ListenStream= references a path below legacy directory /var/run/, updating /var/run/dbus/system_bus_socket ��→ /r.
[    7.194813] random: systemd: uninitialized urandom read (16 bytes read)
[    7.195455] systemd[1]: Reached target Remote File Systems.
[  OK  ] Reached target Remote File Systems.
[    7.210891] random: systemd: uninitialized urandom read (16 bytes read)