Skip to content

[SVE] Add support for the SVE target.#180

Merged
fpetrogalli merged 1 commit intomasterfrom
enable-sve
Mar 14, 2018
Merged

[SVE] Add support for the SVE target.#180
fpetrogalli merged 1 commit intomasterfrom
enable-sve

Conversation

@fpetrogalli
Copy link
Copy Markdown

This commit adds support for the AArch64 Scalable Vector Extension
(SVE) [1]. The vector functions are provided to target Vector Length
Agnostic (VLA) execution [2].

To build SLEEF with SVE support, a compiler that support the SVE Arm C
Language Extensions (ACLE) [2] must be used.

At the time of publishing this patch, the only compiler with SVE ACLE
support is Arm Compiler for HPC [3].

The Cmake configuration expectes Arm Instruction Emulator (ArmIE) [4]
to execute the tests on native AArch64 hardware without SVE support.

The SVE target is build without taking advantage of the native masking
capabilities of SVE. This will be targeted in a upcoming release of
SLEEF, together with the AVX512F native masking capabilities [5].

Additional changes introduced in this patch are:

  1. The mkrename* script have been modified to support VLA names in the
    functions. In particular, 'x' is used to represent the vector
    length of the SVE symbols.

  2. '__sizeless_struct' is a prototype language extension only
    implemented by Arm Compiler For HPC [3] to allow the declaration of
    SVE tuple types as described in section 3.4 of Arm C Language
    Extensions for SVE [2].

  3. A new 'iutsve' executable is generated to test the SVE functions.

[1] https://developer.arm.com/products/software-development-tools/hpc/sve
[2] https://developer.arm.com/docs/100987/0000
[3] https://developer.arm.com/products/software-development-tools/hpc/arm-compiler-for-hpc
[4] https://developer.arm.com/products/software-development-tools/hpc/arm-instruction-emulator
[5] #142

This commit adds support for the AArch64 Scalable Vector Extension
(SVE) [1]. The vector functions are provided to target Vector Length
Agnostic (VLA) execution [2].

To build SLEEF with SVE support, a compiler that support the SVE Arm C
Language Extensions (ACLE) [2] must be used.

At the time of publishing this patch, the only compiler with SVE ACLE
support is Arm Compiler for HPC [3].

The Cmake configuration expectes Arm Instruction Emulator (ArmIE) [4]
to execute the tests on native AArch64 hardware without SVE support.

The SVE target is build without taking advantage of the native masking
capabilities of SVE. This will be targeted in a upcoming release of
SLEEF, together with the AVX512F native masking capabilities [5].

Additional changes introduced in this patch are:

1. The mkrename* script have been modified to support VLA names in the
   functions. In particular, 'x' is used to represent the vector
   length of the SVE symbols.

2. '__sizeless_struct' is a prototype language extension only
   implemented by Arm Compiler For HPC [3] to allow the declaration of
   SVE tuple types as described in section 3.4 of Arm C Language
   Extensions for SVE [2].

3. A new 'iutsve' executable is generated to test the SVE functions.

[1] https://developer.arm.com/products/software-development-tools/hpc/sve
[2] https://developer.arm.com/docs/100987/0000
[3] https://developer.arm.com/products/software-development-tools/hpc/arm-compiler-for-hpc
[4] https://developer.arm.com/products/software-development-tools/hpc/arm-instruction-emulator
[5] #142
@fpetrogalli fpetrogalli requested a review from shibatch March 13, 2018 16:44
Copy link
Copy Markdown
Owner

@shibatch shibatch left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good to me.
I confirmed that the test passes with SVE on my environment.
We need to think of a way to enable testing on a CI environment, though.

@fpetrogalli
Copy link
Copy Markdown
Author

Did you use armclang and armie?

@shibatch
Copy link
Copy Markdown
Owner

yes

@fpetrogalli
Copy link
Copy Markdown
Author

Might be a good idea to report the ctest output here (just the run test/test passed).

@shibatch
Copy link
Copy Markdown
Owner

Here you are.

[uduki]~/work/sleef3/sleef-enable-sve/build-aarch64$ ctest -j 4
Test project /import/namihei.naist.jp/home/n-sibata/work/sleef3/sleef-enable-sve/build-aarch64
      Start  1: iut
      Start  2: iutadvsimd
      Start  3: iutsve
      Start  4: gnuabi_compatibility_ADVSIMD
 1/18 Test  #4: gnuabi_compatibility_ADVSIMD ......   Passed    0.04 sec
      Start  5: gnuabi_compatibility_SVE
 2/18 Test  #5: gnuabi_compatibility_SVE ..........   Passed    0.01 sec
      Start  6: gnuabi_compatibility_SVE_masked
 3/18 Test  #6: gnuabi_compatibility_SVE_masked ...   Passed    0.01 sec
      Start  7: naivetestdp_1
 4/18 Test  #7: naivetestdp_1 .....................   Passed    0.04 sec
      Start  8: naivetestdp_2
 5/18 Test  #8: naivetestdp_2 .....................   Passed    0.02 sec
      Start  9: naivetestdp_3
 6/18 Test  #9: naivetestdp_3 .....................   Passed    0.03 sec
      Start 10: naivetestdp_4
 7/18 Test #10: naivetestdp_4 .....................   Passed    0.03 sec
      Start 11: naivetestdp_5
 8/18 Test #11: naivetestdp_5 .....................   Passed    0.04 sec
      Start 12: naivetestdp_10
 9/18 Test #12: naivetestdp_10 ....................   Passed    2.81 sec
      Start 13: naivetestsp_1
10/18 Test #13: naivetestsp_1 .....................   Passed    0.02 sec
      Start 14: naivetestsp_2
11/18 Test #14: naivetestsp_2 .....................   Passed    0.41 sec
      Start 15: naivetestsp_3
12/18 Test #15: naivetestsp_3 .....................   Passed    0.03 sec
      Start 16: naivetestsp_4
13/18 Test #16: naivetestsp_4 .....................   Passed    0.02 sec
      Start 17: naivetestsp_5
14/18 Test #17: naivetestsp_5 .....................   Passed    0.04 sec
      Start 18: naivetestsp_10
15/18 Test #18: naivetestsp_10 ....................   Passed    2.85 sec
16/18 Test  #1: iut ...............................   Passed  334.39 sec
17/18 Test  #2: iutadvsimd ........................   Passed  352.21 sec
18/18 Test  #3: iutsve ............................   Passed  672.23 sec

100% tests passed, 0 tests failed out of 18

Total Test time (real) = 672.43 sec
[uduki]~/work/sleef3/sleef-enable-sve/build-aarch64$

@fpetrogalli
Copy link
Copy Markdown
Author

Thanks. What system are you using?

@fpetrogalli fpetrogalli merged commit e4217b4 into master Mar 14, 2018
@shibatch
Copy link
Copy Markdown
Owner

This one is Firefly-RK3399.

@shibatch shibatch deleted the enable-sve branch August 28, 2020 11:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants