LTE turbo N_ir

 

Please refer to TS 136 212 - V14.2.0 - LTE; Evolved Universal Terrestrial Radio Access (E-UTRA); Multiplexing and channel coding (3GPP TS 36.212 version 14.2.0 Release 14) (etsi.org)

AVX - ternary operation

This site is very useful for me.

 AVX512: ternary functions evaluation (0x80.pl)

SIMD - transpose

 Intel SIMD (AVX) transpose intrinsics

Intel® Intrinsics Guide

- _shuffle_* (mask)

- permute (mask)

- _unpacklo_, _unpackhi_

- _MM_TRANSPOSE4_PS & c++ - Fastest way to transpose 4x4 byte matrix - Stack Overflow

Example

https://godbolt.org/z/xzfPqnrdo

Intel vs. AMD architecture ...

Agner 

Software optimization resources. C++ and assembly. Windows, Linux, BSD, Mac OS X (agner.org)

microarchitecture.pdf (agner.org)

Chips and Cheese

AMD’s Zen 4 Part 1: Frontend and Execution Engine – Chips and Cheese

CCD / CCX

CCX: core complex = 4 cores

CCD: core complex die = CCX + CCX, i.e., max 8 cores

refer to: AMD CCD and CCX in Ryzen Processors Explained | Hardware Times

SIMD

lec17_x86SIMD.pdf (ntu.edu.tw)

What is the difference between rint() and round() in C or C++?

 From the ChatGPT

In C++, rint is a function from the cmath library that returns the nearest integer value to a given floating-point number, using rounding towards the nearest even number in case of tie.

On the other hand, round is a function that also rounds a floating-point number to the nearest integer value, but it is part of the cmath library only starting from C++11. In earlier versions of C++, round is not available.

So, if you are using C++11 or later, you can use round instead of rint for rounding to the nearest integer value.

Please refer to banker's round

Sample Code: https://onlinegdb.com/q36smlQMz

Beamforming - Zero Forcing

1. Zero Forcing Model

https://www.sharetechnote.com/html/Communication_ChannelModel_ZF.html 

2. Zero Forcing precoding

https://en.wikipedia.org/wiki/Zero-forcing_precoding

https://www.koreascience.or.kr/article/JAKO201334559957395.pdf

Intel SIMD tutorials

By Intel

1. Intel® Intrinsics Guide

https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html

2. Mirror of Intel® Intrinsics Guide

https://www.laruence.com/sse/

3. Guide to vectorization with Intel C++ compilers

 https://www.intel.com/content/dam/www/public/us/en/documents/guides/compiler-auto-vectorization-guide.pdf

By others

1. Faster Set Intersection with SIMD instructions by Reducing ...

http://www.vldb.org/pvldb/vol8/p293-inoue.pdf

2. CS3330: A quick guide to SSE/SIMD

https://www.cs.virginia.edu/~cr4bd/3330/F2018/simdref.html

3. Improving performance with SIMD intrinsics in three use cases

https://stackoverflow.blog/2020/07/08/improving-performance-with-simd-intrinsics-in-three-use-cases/

* 4. Matrix Multiplication Revisited | Richard Startin's Blog

https://richardstartin.github.io/posts/mmm-revisited

5. A Fast Matrix-Matrix Multiply Using Intel's SSE Instructions

https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.23.6754&rep=rep1&type=pdf

6. https://shura.shu.ac.uk/18355/1/Kelefouras-Matrix-MatrixMultiplicationMethodologyforSIngleMulti-Core%28AM%29.pdf

7. https://www.uio.no/studier/emner/matnat/ifi/IN3200/v19/teaching-material/avx512.pdf

8. https://compilers.cs.uni-saarland.de/papers/leissa_vecimp_tr.pdf