New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Factorize SIMD code #222

Closed

slaperche-scality opened this issue Jul 27, 2018 · 1 comment

Assignees

Labels

bug enhancement

Contributor

slaperche-scality commented Jul 27, 2018

There is way too much code duplication/copy paste between all the SIMD implementation (simd_{128,256}_u{16,32}.h).

We must refactor this before adding more SIMD-feature, it's already barely maintainable…

The text was updated successfully, but these errors were encountered:

slaperche-scality added bug enhancement labels

slaperche-scality assigned slaperche-scality and lamphamsy

vrancurel unassigned lamphamsy

slaperche-scality mentioned this issue

FFT::Radix2 uses bit-reversed permutation algorithm #217

Merged

slaperche-scality mentioned this issue

simplify the compile time option #242

Merged

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          use _mm_set_epi16

f642316

By using `_mm_set_epi16` we can set the value using a single instruction
instead of three.
With this change, we can get rid of the MMX instructions
`_mm_setzero_si64` and `_mm_set_pi16`.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          simd: bootstrap the SIMD abstraction layer

59874bc

First, add some useful instruction set-specific definitions.

Two types:
- RegisterType: underlying representation of the future Register class.
- MaskType: same as RegisterType, but for the Mask class.

Three useful constants:
- INSTRUCTION_SET: useful for discrimination.
- ALIGNMENT: useful for memory allocation.
- REG_BITSZ: useful for iteration.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add tests for SIMD

eaf7713

This first test is pretty dumb, but it will be useful to bootstrap and
setup the CircleCI pipeline.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          run the unit tests in all SIMD mode

9c8ead4

That way, we are sure that every codepath is covered and we won't
silently break one of them.

Refs: scality#222, scality#251

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          run the unit tests in all SIMD mode

de57f81

That way, we are sure that every codepath is covered and we won't
silently break one of them.

Refs: scality#222, scality#251

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          run the unit tests in all SIMD mode

67dd8f2

That way, we are sure that every codepath is covered and we won't
silently break one of them.

Refs: scality#222, scality#251

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          simd: bootstrap the SIMD abstraction layer

840fb26

First, add some useful instruction set-specific definitions.

Two types:
- RegisterType: underlying representation of the future Register class.
- MaskType: same as RegisterType, but for the Mask class.

Three useful constants:
- INSTRUCTION_SET: useful for discrimination.
- ALIGNMENT: useful for memory allocation.
- REG_BITSZ: useful for iteration.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add tests for SIMD

7a7e117

This first test is pretty dumb, but it will be useful to bootstrap and
setup the CircleCI pipeline.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          run the unit tests in all SIMD mode

04f0e83

That way, we are sure that every codepath is covered and we won't
silently break one of them.

Refs: scality#222, scality#251

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          run the unit tests in all SIMD mode

7cedfcb

That way, we are sure that every codepath is covered and we won't
silently break one of them.

Refs: scality#222, scality#251

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          simd: bootstrap the SIMD abstraction layer

1f62dd5

First, add some useful instruction set-specific definitions.

Two types:
- RegisterType: underlying representation of the future Register class.
- MaskType: same as RegisterType, but for the Mask class.

Three useful constants:
- INSTRUCTION_SET: useful for discrimination.
- ALIGNMENT: useful for memory allocation.
- REG_BITSZ: useful for iteration.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add tests for SIMD

047bcfb

This first test is pretty dumb, but it will be useful to bootstrap and
setup the CircleCI pipeline.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          run the unit tests in all SIMD mode

2b4754b

That way, we are sure that every codepath is covered and we won't
silently break one of them.

Refs: scality#222, scality#251

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add a custom allocator for aligned memory

4599d1f

With this allocator  we can use the STL container, such as std::vector,
and be sure that we can use aligned load instruction when necessary.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          simd: bootstrap the SIMD abstraction layer

7c121f0

First, add some useful instruction set-specific definitions.

Two types:
- RegisterType: underlying representation of the future Register class.
- MaskType: same as RegisterType, but for the Mask class.

Three useful constants:
- INSTRUCTION_SET: useful for discrimination.
- ALIGNMENT: useful for memory allocation.
- REG_BITSZ: useful for iteration.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add tests for SIMD

This first test is pretty dumb, but it will be useful to bootstrap and
setup the CircleCI pipeline.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          run the unit tests in all SIMD mode

f83a7a4

Now, we are sure that every codepath is covered and we won't silently
break one of them.

Refs: scality#222, scality#251

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add a custom allocator for aligned memory

bdbb5ae

With this allocator we can use the STL container, such as std::vector,
and be sure that we can use aligned load instruction when necessary.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add test for the custom allocator

5d0ef99

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          use the new allocator for aligned allocations

7ce4d38

The new implementation is more C++-like (can be used with STL container)
and more robust (overflow handling, strict aliasing rule, …).

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          run the unit tests in all SIMD mode

e112256

Now, we are sure that every codepath is covered and we won't silently
break one of them.

Refs: scality#222, scality#251

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add a custom allocator for aligned memory

03514dd

With this allocator we can use the STL container, such as std::vector,
and be sure that we can use aligned load instruction when necessary.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add test for the custom allocator

e7a412e

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          use the new allocator for aligned allocations

83e2bec

The new implementation is more C++-like (can be used with STL container)
and more robust (overflow handling, strict aliasing rule, …).

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          run the unit tests in all SIMD mode

65e34d8

Now, we are sure that every codepath is covered and we won't silently
break one of them.

Refs: scality#222, scality#251

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add a custom allocator for aligned memory

88be0ed

With this allocator we can use the STL container, such as std::vector,
and be sure that we can use aligned load instruction when necessary.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add test for the custom allocator

7ad3fde

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          use the new allocator for aligned allocations

fdcf475

The new implementation is more C++-like (can be used with STL container)
and more robust (overflow handling, strict aliasing rule, …).

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add a custom allocator for aligned memory

e262ef0

With this allocator we can use the STL container, such as std::vector,
and be sure that we can use aligned load instruction when necessary.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add test for the custom allocator

422a907

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          use the new allocator for aligned allocations

d141979

The new implementation is more C++-like (can be used with STL container)
and more robust (overflow handling, strict aliasing rule, …).

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add a custom allocator for aligned memory

14735a6

With this allocator we can use the STL container, such as std::vector,
and be sure that we can use aligned load instruction when necessary.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add test for the custom allocator

12bccc2

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          use the new allocator for aligned allocations

35ad3a8

The new implementation is more C++-like (can be used with STL container)
and more robust (overflow handling, strict aliasing rule, …).

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add a custom allocator for aligned memory

ff4eef9

With this allocator we can use the STL container, such as std::vector,
and be sure that we can use aligned load instruction when necessary.

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          add test for the custom allocator

635f83f

Refs: scality#222

slaperche-scality added a commit to slaperche-scality/quadiron that referenced this issue


          use the new allocator for aligned allocations

ba0ec36

The new implementation is more C++-like (can be used with STL container)
and more robust (overflow handling, strict aliasing rule, …).

Refs: scality#222

Contributor

lamphamsy commented Jun 18, 2019

It's fixed by #253 #254

lamphamsy closed this as completed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment