StackOverflow Questions for Tag: fma

user24200147

Reputation: 1

What is considered as "2 FMAs"?

performancecpuavx512cpu-speedfma

Score: 0

Answers: 0

Z boson

Reputation: 33659

Optimize for fast multiplication but slow addition: FMA and doubledouble

assemblyx86floating-pointfmadouble-double-arithmetic

Score: 12

Answers: 3

einpoklum

Reputation: 131405

Is floating point expression contraction allowed in C++?

c++floating-pointlanguage-lawyerfma

Score: 10

Answers: 2

Y-Jiechao

Reputation: 5

Why does '_mm256_fmadd_ps' cause precision loss?

cprecisionavxavx2fma

Score: 0

Answers: 1

Anili

Reputation: 140

Achieving More FMA3 Performance Than The Theoretical Maximum

cbenchmarkingavx2flopsfma

Score: 1

Answers: 0

Anili

Reputation: 140

High Variance In Manual Vectorization Performance

cperformancevectorizationavx2fma

Score: 2

Answers: 1

cyrusbehr

Reputation: 1301

AVX2: Computing dot product of 512 float arrays

c++simdavx2dot-productfma

Score: 23

Answers: 3

Serge Rogatch

Reputation: 15020

Can we replace XOR with multiply-add?

c++cudabit-manipulationxorfma

Score: 0

Answers: 0

Martin Brown

Reputation: 2649

Why this AVX2 slowdown with FMA x86 MS C Compiler?

cx86avx2microbenchmarkfma

Score: 0

Answers: 0

MSalters

Reputation: 179779

How to get data out of AVX registers?

c++visual-c++avxfma

Score: 16

Answers: 5

Fedor

Reputation: 20997

Clang fused multiply-add depends on constancy of expression arguments

coptimizationfloating-pointclangfma

Score: 5

Answers: 0

Christian Sarofeen

Reputation: 2250

Accurate method to calculate double FMA and Shared memory latency

cudalatencymicrobenchmarkgpu-shared-memoryfma

Score: 3

Answers: 1

anna

Reputation: 59

v4fmaddps instructions for packed 32-bit integers

floating-pointavx512fma

Score: 1

Answers: 0

BadHellie

Reputation: 307

GCC 12 (minGW 64): how to enable fused multiply add code generation

assemblygccoptimizationavxfma

Score: 0

Answers: 0

xiaohuihui

Reputation: 45

How should I implement a generic FMA/FMAF instruction in software?

mathfloating-pointfma

Score: 2

Answers: 1

A.nechi

Reputation: 541

FMA intrinsics not working: is it Hardware or Compiler?

cx86simdintrinsicsfma

Score: 4

Answers: 2

pem

Reputation: 465

Fast fixed-size polynomial evaluation: MSVC vs GCC

c++gccvisual-c++compiler-optimizationfma

Score: 3

Answers: 0

Peter

Reputation: 38455

Fastest way to multiply and sum/add two arrays (dot product) - unaligned surprisingly faster than FMA

c#.net-6.0intrinsicsavx2fma

Score: 5

Answers: 1

tigertang

Reputation: 457

Deleteing initialization leads to avx2 fma performance drop. Why?

c++x86cpuavx2fma

Score: 1

Answers: 0

Mirco Mannino

Reputation: 35

Latency and number of FMA units

c++intelsimdmicro-optimizationfma

Score: 1

Answers: 0

PreviousPage 1Next

StackOverflow Questions for Tag: fma

What is considered as &quot;2 FMAs&quot;?

Optimize for fast multiplication but slow addition: FMA and doubledouble

Is floating point expression contraction allowed in C++?

Why does &#39;_mm256_fmadd_ps&#39; cause precision loss?

Achieving More FMA3 Performance Than The Theoretical Maximum

High Variance In Manual Vectorization Performance

AVX2: Computing dot product of 512 float arrays

Can we replace XOR with multiply-add?

Why this AVX2 slowdown with FMA x86 MS C Compiler?

How to get data out of AVX registers?

Clang fused multiply-add depends on constancy of expression arguments

Accurate method to calculate double FMA and Shared memory latency

v4fmaddps instructions for packed 32-bit integers

GCC 12 (minGW 64): how to enable fused multiply add code generation

How should I implement a generic FMA/FMAF instruction in software?

FMA intrinsics not working: is it Hardware or Compiler?

Fast fixed-size polynomial evaluation: MSVC vs GCC

Fastest way to multiply and sum/add two arrays (dot product) - unaligned surprisingly faster than FMA

Deleteing initialization leads to avx2 fma performance drop. Why?

Latency and number of FMA units

What is considered as "2 FMAs"?

Why does '_mm256_fmadd_ps' cause precision loss?