Skip to content

stdsimd fails to build with target_cpu=skylake-avx512 #157

Closed
rust-lang/rust
#47828
Closed
stdsimd fails to build with target_cpu=skylake-avx512#157
rust-lang/rust
#47828
Labels
A-llvmA-x86

Description

@hdevalence
hdevalence
opened on Oct 27, 2017

When compiling stdsimd with rustc 1.22.0-nightly (b7960878b 2017-10-18) and -C target_cpu=skylake-avx512, the build fails:

$ RUSTFLAGS="-C target_cpu=skylake-avx512" cargo build --release
   Compiling stdsimd v0.0.2 (file:///<stripped>/stdsimd)
LLVM ERROR: Cannot select: t83: v32i16 = X86ISD::VINSERT t82, t78, Constant:i64<0>
  t82: v32i16 = bitcast t81
    t81: v16i32 = BUILD_VECTOR Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
      t80: i32 = Constant<0>
  t78: i16 = truncate t5
    t5: i64,ch = CopyFromReg t0, Register:i64 %vreg2
      t4: i64 = Register %vreg2
  t79: i64 = Constant<0>
In function: _ZN61_$LT$stdsimd..v512..u16x32$u20$as$u20$core..fmt..LowerHex$GT$3fmt17h59e1149cebeaa07fE
error: Could not compile `stdsimd`.

To learn more, run the command again with --verbose.

Activity

  1. hdevalence commented on Oct 27, 2017

    @hdevalence
    hdevalence
    on Oct 27, 2017
    ContributorAuthor

    Discussion moved from comments here: #146 (comment) and here #146 (comment)

  2. hdevalence commented on Oct 27, 2017

    @hdevalence
    hdevalence
    on Oct 27, 2017
    ContributorAuthor

    @alexcrichton re #146 (comment) , one motivation is to have SSE2/AVX2 code written using intrinsics be able to use the extra registers xmm16..xmm31, ymm16..ymm31 that AVX512VL provides. Is this possible using feature detection?

  3. alexcrichton commented on Oct 30, 2017

    @alexcrichton
    alexcrichton
    on Oct 30, 2017
    Member

    Oh yeah definitely makes sense to fix this! I tested a snippet which I think may have been a reduction of this? It looks like this is fixed in LLVM 5 so this may be blocked on rust-lang/rust#43370

  4. hdevalence commented on Oct 30, 2017

    @hdevalence
    hdevalence
    on Oct 30, 2017
    ContributorAuthor

    Building the branch mentioned in that issue and using that as the local Rust toolchain fixed the issue for me.

  5. hdevalence commented on Dec 12, 2017

    @hdevalence
    hdevalence
    on Dec 12, 2017
    ContributorAuthor

    Update: using the current nightly Rust and the current stdsimd has stopped working on skylake-avx512.

  6. gnzlbg commented on Dec 12, 2017

    @gnzlbg
    gnzlbg
    on Dec 12, 2017
    Contributor

    Does it work with a previous stdsimd version but the current nightly Rust?

    Do you have a reproducer? Does it fail to compile, at run-time, etc?

    EDIT: I get the following error:

    RUSTFLAGS="-C target_cpu=skylake-avx512" cargo build --release
   Compiling coresimd v0.0.3 (file:///stdsimd/coresimd)
LLVM ERROR: Cannot select: 0x1194d1680: v32i16 = X86ISD::VINSERT 0x120351680, 0x12df5d5b0, Constant:i64<0>
  0x120351680: v32i16 = bitcast 0x1203eeb60
    0x1203eeb60: v16i32 = BUILD_VECTOR Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>, Constant:i32<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
      0x1194d3680: i32 = Constant<0>
  0x12df5d5b0: i16 = truncate 0x1194d3f70
    0x1194d3f70: i64,ch = CopyFromReg 0x12df0c730, Register:i64 %vreg2
      0x1194d56e8: i64 = Register %vreg2
  0x1194d5478: i64 = Constant<0>
In function: _ZN62_$LT$coresimd..v512..u16x32$u20$as$u20$core..fmt..LowerHex$GT$3fmt17h0a46f3602aaf0609E
error: Could not compile `coresimd`.

To learn more, run the command again with --verbose.

  7. hdevalence commented on Dec 15, 2017

    @hdevalence
    hdevalence
    on Dec 15, 2017
    ContributorAuthor

    Oops, sorry, I missed this. I guess you reproduced it; my test case is this AVX2 code: https://github.com/isislovecruft/curve25519-dalek/pull/94

    It doesn't use any 512-bit vectors, but there's a mild performance increase when using AVX512VL since the increase in registers avoids spills.

  8. gnzlbg commented on Dec 15, 2017

    @gnzlbg
    gnzlbg
    on Dec 15, 2017 via email
    Contributor
  9. alexcrichton
    added a commit that references this issue on Jan 28, 2018

    rustc: Upgrade to LLVM 6

    d9031ae
  10. bors
    added a commit that references this issue on Jan 28, 2018

    Auto merge of #47828 - alexcrichton:llvm-6, r=<try>

    38bd381
  11. alexcrichton
    added
    A-llvm
    A-x86
    on Jan 29, 2018
  12. bors
    added a commit that references this issue on Jan 30, 2018

    Auto merge of #47828 - alexcrichton:llvm-6, r=nikomatsakis

    22d405f

  13. 16 remaining items

  14. alexcrichton
    added 3 commits that reference this issue on Feb 7, 2018

    rustc: Upgrade to LLVM 6

    ced1f3d
  15. bors
    added a commit that references this issue on Feb 9, 2018

    Auto merge of #47828 - alexcrichton:llvm-6, r=<try>

    cef95b0
  16. alexcrichton
    added a commit that references this issue on Feb 10, 2018

    rustc: Upgrade to LLVM 6

    6b7b6b6
  17. bors
    added a commit that references this issue on Feb 10, 2018

    Auto merge of #47828 - alexcrichton:llvm-6, r=nikomatsakis

    39abcc0

  18. alexcrichton commented on Feb 11, 2018

    @alexcrichton
    alexcrichton
    on Feb 11, 2018
    Member

    I believe this was fixed by rust-lang/rust#47828

  19. alexcrichton
    closed this as completedon Feb 11, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    A-llvmA-x86

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions