[llvm-dev] Problem with clang optimizer?

240 views
Skip to first unread message

Blumenthal, Uri - 0553 - MITLL via llvm-dev

unread,
Sep 23, 2021, 10:14:10 PM9/23/21
to LLVM-DEV LIST

I’m not sure if this is the correct list, so please direct me to the right one if this bug report shouldn’t go here.

 

The problem is: invoking clang (v12) with -O2 or better optimization flags generates wrong object code for the following C++. Compiling it with -O1 generates working binary.

 

 

=================

 

#include <cstdint>

#include <cassert>

 

template<size_t ROT, typename T>

inline constexpr T rotl(T input)

   {

   static_assert(ROT > 0 && ROT < 8*sizeof(T), "Invalid rotation constant");

   return static_cast<T>((input << ROT) | (input >> (8*sizeof(T) - ROT)));

   }

 

inline void SHA3_round(uint64_t T[25], const uint64_t A[25], uint64_t RC)

   {

   const uint64_t C0 = A[0] ^ A[5] ^ A[10] ^ A[15] ^ A[20];

   const uint64_t C1 = A[1] ^ A[6] ^ A[11] ^ A[16] ^ A[21];

 

   // the calculation of C2 fails for -O3 or -O2 with clang 12

   // FWIW: it would produce a value that doesn't fit into a _signed_ 64-bit int

   const uint64_t C2 = A[2] ^ A[7] ^ A[12] ^ A[17] ^ A[22];

 

   const uint64_t C3 = A[3] ^ A[8] ^ A[13] ^ A[18] ^ A[23];

   const uint64_t C4 = A[4] ^ A[9] ^ A[14] ^ A[19] ^ A[24];

 

   const uint64_t D0 = rotl<1>(C0) ^ C3;

   const uint64_t D1 = rotl<1>(C1) ^ C4;

   const uint64_t D2 = rotl<1>(C2) ^ C0;

   const uint64_t D3 = rotl<1>(C3) ^ C1;

   const uint64_t D4 = rotl<1>(C4) ^ C2;

 

   const uint64_t B00 =          A[ 0] ^ D1;

   const uint64_t B01 = rotl<44>(A[ 6] ^ D2);

   const uint64_t B02 = rotl<43>(A[12] ^ D3);

   const uint64_t B03 = rotl<21>(A[18] ^ D4);

   const uint64_t B04 = rotl<14>(A[24] ^ D0);

   T[ 0] = B00 ^ (~B01 & B02) ^ RC;

   T[ 1] = B01 ^ (~B02 & B03);

   T[ 2] = B02 ^ (~B03 & B04);

   T[ 3] = B03 ^ (~B04 & B00);

   T[ 4] = B04 ^ (~B00 & B01);

 

   const uint64_t B05 = rotl<28>(A[ 3] ^ D4);

   const uint64_t B06 = rotl<20>(A[ 9] ^ D0);

   const uint64_t B07 = rotl< 3>(A[10] ^ D1);

   const uint64_t B08 = rotl<45>(A[16] ^ D2);

   const uint64_t B09 = rotl<61>(A[22] ^ D3);

   T[ 5] = B05 ^ (~B06 & B07);

   T[ 6] = B06 ^ (~B07 & B08);

   T[ 7] = B07 ^ (~B08 & B09);

   T[ 8] = B08 ^ (~B09 & B05);

   T[ 9] = B09 ^ (~B05 & B06);

 

   // --- instructions starting from here can be removed

   //     and the -O3 dicrepancy is still triggered

 

   const uint64_t B10 = rotl< 1>(A[ 1] ^ D2);

   const uint64_t B11 = rotl< 6>(A[ 7] ^ D3);

   const uint64_t B12 = rotl<25>(A[13] ^ D4);

   const uint64_t B13 = rotl< 8>(A[19] ^ D0);

   const uint64_t B14 = rotl<18>(A[20] ^ D1);

   T[10] = B10 ^ (~B11 & B12);

   T[11] = B11 ^ (~B12 & B13);

   T[12] = B12 ^ (~B13 & B14);

   T[13] = B13 ^ (~B14 & B10);

   T[14] = B14 ^ (~B10 & B11);

 

   const uint64_t B15 = rotl<27>(A[ 4] ^ D0);

   const uint64_t B16 = rotl<36>(A[ 5] ^ D1);

   const uint64_t B17 = rotl<10>(A[11] ^ D2);

   const uint64_t B18 = rotl<15>(A[17] ^ D3);

   const uint64_t B19 = rotl<56>(A[23] ^ D4);

   T[15] = B15 ^ (~B16 & B17);

   T[16] = B16 ^ (~B17 & B18);

   T[17] = B17 ^ (~B18 & B19);

   T[18] = B18 ^ (~B19 & B15);

   T[19] = B19 ^ (~B15 & B16);

 

   const uint64_t B20 = rotl<62>(A[ 2] ^ D3);

   const uint64_t B21 = rotl<55>(A[ 8] ^ D4);

   const uint64_t B22 = rotl<39>(A[14] ^ D0);

   const uint64_t B23 = rotl<41>(A[15] ^ D1);

   const uint64_t B24 = rotl< 2>(A[21] ^ D2);

   T[20] = B20 ^ (~B21 & B22);

   T[21] = B21 ^ (~B22 & B23);

   T[22] = B22 ^ (~B23 & B24);

   T[23] = B23 ^ (~B24 & B20);

   T[24] = B24 ^ (~B20 & B21);

   }

 

int main()

{

    uint64_t T[25];

 

    uint64_t A[25] = {

        15515230172486u, 9751542238472685244u, 220181482233372672u,

        2303197730119u, 9537012007446913720u, 0u, 14782389640143539577u,

        2305843009213693952u, 1056340403235818873u, 16396894922196123648u,

        13438274300558u, 3440198220943040u, 0u, 3435902021559310u, 64u,

        14313837075027532897u, 32768u, 6880396441885696u, 14320469711924527201u,

        0u, 9814829303127743595u, 18014398509481984u, 14444556046857390455u,

        4611686018427387904u, 18041275058083100u };

 

    SHA3_round(T, A, 0x0000000000008082);

 

    assert(T[0]  == 16394434931424703552u);

    assert(T[1]  == 10202638136074191489u);

    assert(T[2]  == 6432602484395933614u);

    assert(T[3]  == 10616058301262943899u);

    assert(T[4]  == 14391824303596635982u);

    assert(T[5]  == 5673590995284149638u);

    assert(T[6]  == 15681872423764765508u);

    assert(T[7]  == 11470206704342013341u);

    assert(T[8]  == 8508807405493883168u);

    assert(T[9]  == 9461805213344568570u);

    assert(T[10] == 8792313850970105187u);

    assert(T[11] == 13508586629627657374u);

    assert(T[12] == 5157283382205130943u);

    assert(T[13] == 375019647457809685u);

    assert(T[14] == 9294608398083155963u);

    assert(T[15] == 16923121173371064314u);

    assert(T[16] == 4737739424553008030u);

    assert(T[17] == 5823987023293412593u);

    assert(T[18] == 13908063749137376267u);

    assert(T[19] == 13781177305593198238u);

    assert(T[20] == 9673833001659673401u);

    assert(T[21] == 17282395057630454440u);

    assert(T[22] == 12906624984756985556u);

    assert(T[23] == 3081478361927354234u);

    assert(T[24] == 93297594635310132u);

 

    return 0;

}

=================

 

Your help debugging and fixing this problem is appreciated!

--

Regards,

Uri Blumenthal                              Voice: (781) 981-1638 

Secure Resilient Systems and Technologies   Cell:  (339) 223-5363

MIT Lincoln Laboratory                     

244 Wood Street, Lexington, MA  02420-9108      

 

Web:     https://www.ll.mit.edu/biographies/uri-blumenthal

Root CA: https://www.ll.mit.edu/llrca2.pem

 

There are two ways to design a system. One is to make is so simple there are obviously no deficiencies.

The other is to make it so complex there are no obvious deficiencies.

                                                                                                                                     -  C. A. R. Hoare

 

Jakub (Kuba) Kuderski via llvm-dev

unread,
Sep 23, 2021, 10:37:49 PM9/23/21
to Blumenthal, Uri - 0553 - MITLL, LLVM-DEV LIST
Hi Uri,

I tried to reproduce it on goldbolt with clang 12.0.0 and 12.0.1 but things seem fine when I run it there: https://godbolt.org/z/vrq8j6Kj7.
Can you share your exact clang invocation? Does it only reproduce in some specific environment?

Also, it generally helps to reduce code bug reports as much as possible; creduce can help with that: https://embed.cs.utah.edu/creduce/using/.

-Jakub

_______________________________________________
LLVM Developers mailing list
llvm...@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev


--
Jakub Kuderski

Michael Kruse via llvm-dev

unread,
Sep 23, 2021, 10:39:31 PM9/23/21
to Blumenthal, Uri - 0553 - MITLL, LLVM-DEV LIST
Reproducer: https://godbolt.org/z/f6nxnG3sY
(cstddef and -march must be added as well)

Michael

Craig Topper via llvm-dev

unread,
Sep 23, 2021, 11:27:38 PM9/23/21
to Michael Kruse, LLVM-DEV LIST, Blumenthal, Uri - 0553 - MITLL
opt-bisect points to SLP vectorizer. And it looks like it doesn't fail on trunk.

~Craig

Blumenthal, Uri - 0553 - MITLL via llvm-dev

unread,
Sep 24, 2021, 9:19:59 AM9/24/21
to Craig Topper, Michael Kruse, LLVM-DEV LIST

Also, it seems to impact only MacOS? One colleague wasn’t able to reproduce with Clang-12 on Linux.

This is where this problem was discovered and is tracked, with all the details:

https://github.com/randombit/botan/issues/2802

 

If you nail this one, it would be great.

 

Thanks!

--

Regards,

Uri

 

There are two ways to design a system. One is to make is so simple there are obviously no deficiencies.

The other is to make it so complex there are no obvious deficiencies.

                                                                                                                                     -  C. A. R. Hoare

Blumenthal, Uri - 0553 - MITLL via llvm-dev

unread,
Sep 24, 2021, 9:31:11 AM9/24/21
to Jakub (Kuba) Kuderski, LLVM-DEV LIST

I tried to reproduce it on goldbolt with clang 12.0.0 and 12.0.1 but things seem fine when I run it there: https://godbolt.org/z/vrq8j6Kj7.
Can you share your exact clang invocation? Does it only reproduce in some specific environment?

 

Save the source I posted before into “sha3-reproducer.cxx” file. Let me know if you want it re-posted here.

 

$ clang++-mp-12 -v

clang version 12.0.1

Target: x86_64-apple-darwin20.6.0

Thread model: posix

InstalledDir: /opt/local/libexec/llvm-12/bin

$ clang++-mp-12 -o s -O3 sha3-reproducer.cxx

$ ./s

Assertion failed: (T[0] == 16394434931424703552u), function main, file sha3-reproducer.cxx, line 103.

Abort trap: 6

$ clang++-mp-12 -o s -O2 sha3-reproducer.cxx

$ ./s

Assertion failed: (T[0] == 16394434931424703552u), function main, file sha3-reproducer.cxx, line 103.

Abort trap: 6

$ clang++-mp-12 -o s -O1 sha3-reproducer.cxx

$ ./s

$ 

 

Clang-12 is installed via Macports, which is why we invoke the executable as clang++-mp-12.

 

The same problem manifests in exactly the same way in the Xcode-13 version of Clang (presumably based on LLVM Clang-12).

 

I’ll be happy to provide more of specific details, if you let me know what you need.



Also, it generally helps to reduce code bug reports as much as possible; creduce can help with that: https://embed.cs.utah.edu/creduce/using/.

 

Understood. Unfortunately, the above reproducer is the best we could come up with. An alternative is trying to build the Botan package itself https://github.com/randombit/botan.git.




Dimitry Andric via llvm-dev

unread,
Sep 25, 2021, 5:51:43 AM9/25/21
to Blumenthal, Uri - 0553 - MITLL, LLVM-DEV LIST
It is only occurring (as far as I can see now) on x86_64, with -mavx enabled. Or with a target CPU that supports AVX. And it is not Apple clang specific.

-Dimitry
signature.asc

Craig Topper via llvm-dev

unread,
Sep 25, 2021, 12:06:53 PM9/25/21
to Dimitry Andric, LLVM-DEV LIST, Blumenthal, Uri - 0553 - MITLL
It reproduced for me with -march=nehalem which does not have AVX.
~Craig

Blumenthal, Uri - 0553 - MITLL via llvm-dev

unread,
Sep 25, 2021, 8:48:34 PM9/25/21
to Craig Topper, Dimitry Andric, LLVM-DEV LIST

I found that

  • The problem disappears with -mno-sse4.1
  • The problem manifests with both Apple Clang from Xcode-13, and LLVM Clang-12 (and not with Xcode-12 or LLVM Clang-11)
  • I could experiment only on Apple platform, as that’s the only one I have that runs LLVM Clang-12.

 

--

Regards,

Uri

 

There are two ways to design a system. One is to make is so simple there are obviously no deficiencies.

The other is to make it so complex there are no obvious deficiencies.

                                                                                                                                     -  C. A. R. Hoare

 

 

From: Craig Topper <craig....@gmail.com>
Date: Saturday, September 25, 2021 at 12:07
To: Dimitry Andric <dim...@andric.com>
Cc: Uri Blumenthal <u...@ll.mit.edu>, LLVM-DEV LIST <llvm...@lists.llvm.org>
Subject: Re: [llvm-dev] Problem with clang optimizer?

 

It reproduced for me with -march=nehalem which does not have AVX.

Craig Topper via llvm-dev

unread,
Sep 26, 2021, 1:12:43 AM9/26/21
to Blumenthal, Uri - 0553 - MITLL, LLVM-DEV LIST
Looking at the IR here https://godbolt.org/z/zaMW1renW I believe the issue is on this instruction on line 361

%30 = extractelement <2 x <2 x i64>*> %bc438, i32 0

It should be extracting from index 1 instead of index 0.

~Craig

Jameson Nash via llvm-dev

unread,
Sep 29, 2021, 7:40:34 PM9/29/21
to Craig Topper, LLVM-DEV LIST, Blumenthal, Uri - 0553 - MITLL
This may be fixed now (https://reviews.llvm.org/D106613), but it remains to be confirmed for https://bugs.llvm.org/show_bug.cgi?id=51957

Blumenthal, Uri - 0553 - MITLL via llvm-dev

unread,
Oct 25, 2021, 5:35:37 PM10/25/21
to Jameson Nash, Craig Topper, LLVM-DEV LIST

I just tried Clang-13 (with LLVM-13), and the problem is still there. Vectorizer still broken wrt. SSE-4.1 instruction extensions:

 

$ echo $CXXFLAGS

-std=gnu++17 -O3 -march=native -isysroot /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk

$ clang++-mp-13 $CXXFLAGS -o t sha3-reproducer.cxx 

$ ./t

Assertion failed: (T[0] == 16394434931424703552u), function main, file sha3-reproducer.cxx, line 103.

Abort trap: 6

$ clang++-mp-13 $CXXFLAGS -mno-sse4.1 -o t sha3-reproducer.cxx 

$ ./t

$ 

Dimitry Andric via llvm-dev

unread,
Oct 25, 2021, 11:25:53 PM10/25/21
to Blumenthal, Uri - 0553 - MITLL, LLVM-DEV LIST
Hi Uri,

Unfortunately the fix for this didn't make into 13.0.0, and will hopefully be part of 13.0.1 (when that comes out I can't say though).

-Dimitry
signature.asc

Anton Afanasyev via llvm-dev

unread,
Oct 26, 2021, 4:14:47 AM10/26/21
to Blumenthal, Uri - 0553 - MITLL, LLVM-DEV LIST
Hi Uri,
could you please reproduce this at godbolt.org? AFAIK, this issue is veiled at clang-13, though real fix isn't backported (see Dimitry's comment: https://bugs.llvm.org/show_bug.cgi?id=51957#c7).
I can't reproduce it on clang-13: https://godbolt.org/z/4Mdrd5388
  Thanks,
    Anton


вт, 26 окт. 2021 г. в 00:35, Blumenthal, Uri - 0553 - MITLL via llvm-dev <llvm...@lists.llvm.org>:

Blumenthal, Uri - 0553 - MITLL via llvm-dev

unread,
Oct 26, 2021, 9:40:40 AM10/26/21
to Anton Afanasyev, LLVM-DEV LIST

Ok, I found that confusion comes from the different binaries supplied as "version 13", see below.

 

Now I am confused.

 

Does it mean that distributions (Ubuntu on Linux, Macports on MacOS, etc.) took pre-release (aka, still-beta) LLVM/Clang and released it as Clang-13?

 

Or that LLVM or Clang were patched after being released, without updating the version numbers?

 

> So, requested Clang, but compiled with GCC???

Godbold actually uses Clang here, "-gcc-toolchain" is just its option to provide GCC installation.

 

I do not understand the above – that “checkbox-button” seems to be named “All compilation options” for the currently selected compiler/toolchain…?

On the other hand, I’m probably not the main customer of that tool, so my understanding is not crucial. ;-)

 

Yes, I've managed to reproduce it too on my ubuntu box with "clang-13" package installed:  .  .  .  .  .

But that's not the clang-13 used by me and godbold, tagged as llvmorg-13.0.0 and announced on 4 October 2021. The ubuntu-packaged clang-13 uses e5f2898bc751 tip commit pushed on 27 March 2021, so it's an older version.

 

I hear you – but what about Mac? Macports Clang-13.0.0 was released only a couple of days ago. I don’t know how to check its commit level, but I strongly doubt it’s as dated as, e.g., Ubuntu release…

 

If you want to get subject issue fixed, you can use clang-13 from here: https://github.com/llvm/llvm-project/releases/tag/llvmorg-13.0.0, for instance, here is apple-darwin version: https://github.com/llvm/llvm-project/releases/download/llvmorg-13.0.0/clang+llvm-13.0.0-x86_64-apple-darwin.tar.xz

 

I do, but I much prefer that Macports maintains “big” packages on my machines, and LLVM-Clang definitely qualify. Thus, I’d rather not track llvm-project on GitHub myself.

 

Can I hope that the fix Dimitry mentioned, will be in 13.0.1 that will (hopefully, eventually) filter downstream?

 

Anton Afanasyev via llvm-dev

unread,
Oct 26, 2021, 12:43:23 PM10/26/21
to Blumenthal, Uri - 0553 - MITLL, LLVM-DEV LIST
Hi Uri,
sorry for confusing you! I didn't tackle this enough (and had an old strange clang-13 installation on a weird box).
Your issue is actually reproduced on godbolt with "-march=skylake" option: https://godbolt.org/z/jaMP3W1cn .
That means this bug is veiled only in particular. As Dimitry said, the next clang-13.0.1 release should grasp the whole fix.
  Thanks,
    Anton

вт, 26 окт. 2021 г. в 16:40, Blumenthal, Uri - 0553 - MITLL <u...@ll.mit.edu>:

Blumenthal, Uri - 0553 - MITLL via llvm-dev

unread,
Oct 26, 2021, 12:47:03 PM10/26/21
to Anton Afanasyev, LLVM-DEV LIST

could you please reproduce this at godbolt.org?

 

Shows up with 12.0.0 and 12.0.1, but not on 13.0.0 on that site. But there seem to be some issues with that site – see below.

 

 

AFAIK, this issue is veiled at clang-13,

 

I’m not sure. First, on my machines – it shows dependence on CPU (fails on Skylake, passes on Skylake-avx512).

 

$ clang++-mp-13 -v -O3 -march=native -o t sha3-reproducer.cxx

clang version 13.0.0

Target: x86_64-apple-darwin20.6.0

Thread model: posix

InstalledDir: /opt/local/libexec/llvm-13/bin

"/opt/local/libexec/llvm-13/bin/clang" -cc1 -triple x86_64-apple-macosx11.0.0 -Wundef-prefix=TARGET_OS_ -Werror=undef-prefix -Wdeprecated-objc-isa-usage -Werror=deprecated-objc-isa-usage -emit-obj --mrelax-relocations -disable-free -disable-llvm-verifier -discard-value-names -main-file-name sha3-reproducer.cxx -mrelocation-model pic -pic-level 2 -mframe-pointer=all -fno-rounding-math -munwind-tables -target-sdk-version=11.3 -fcompatibility-qualified-id-block-type-checking -fvisibility-inlines-hidden-static-local-var -target-cpu skylake -target-feature +sse2 -target-feature -tsxldtrk -target-feature +cx16 -target-feature +sahf -target-feature -tbm -target-feature -avx512ifma -target-feature -sha -target-feature -gfni -target-feature -fma4 -target-feature -vpclmulqdq -target-feature +prfchw -target-feature +bmi2 -target-feature -cldemote -target-feature +fsgsbase -target-feature -ptwrite -target-feature -amx-tile -target-feature -uintr -target-feature +popcnt -target-feature -widekl -target-feature +aes -target-feature -avx512bitalg -target-feature -movdiri -target-feature +xsaves -target-feature -avx512er -target-feature -avxvnni -target-feature -avx512vnni -target-feature -amx-bf16 -target-feature -avx512vpopcntdq -target-feature -pconfig -target-feature -clwb -target-feature -avx512f -target-feature +xsavec -target-feature -clzero -target-feature -pku -target-feature +mmx -target-feature -lwp -target-feature -rdpid -target-feature -xop -target-feature +rdseed -target-feature -waitpkg -target-feature -kl -target-feature -movdir64b -target-feature -sse4a -target-feature -avx512bw -target-feature +clflushopt -target-feature +xsave -target-feature -avx512vbmi2 -target-feature +64bit -target-feature -avx512vl -target-feature -serialize -target-feature -hreset -target-feature +invpcid -target-feature -avx512cd -target-feature +avx -target-feature -vaes -target-feature -avx512bf16 -target-feature +cx8 -target-feature +fma -target-feature -rtm -target-feature +bmi -target-feature -enqcmd -target-feature +rdrnd -target-feature -mwaitx -target-feature +sse4.1 -target-feature +sse4.2 -target-feature +avx2 -target-feature +fxsr -target-feature -wbnoinvd -target-feature +sse -target-feature +lzcnt -target-feature +pclmul -target-feature -prefetchwt1 -target-feature +f16c -target-feature +ssse3 -target-feature +sgx -target-feature -shstk -target-feature +cmov -target-feature -avx512vbmi -target-feature -amx-int8 -target-feature +movbe -target-feature -avx512vp2intersect -target-feature +xsaveopt -target-feature -avx512dq -target-feature +adx -target-feature -avx512pf -target-feature +sse3 -debugger-tuning=lldb -target-linker-version 650.9 -v -fcoverage-compilation-dir=/Users/ur20980/src -resource-dir /opt/local/libexec/llvm-13/lib/clang/13.0.0 -isysroot /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk -I/usr/local/include -stdlib=libc++ -internal-isystem /opt/local/libexec/llvm-13/bin/../include/c++/v1 -internal-isystem /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/local/include -internal-isystem /opt/local/libexec/llvm-13/lib/clang/13.0.0/include -internal-externc-isystem /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include -O3 -fdeprecated-macro -fdebug-compilation-dir=/Users/ur20980/src -ferror-limit 19 -stack-protector 1 -fblocks -fencode-extended-block-signature -fregister-global-dtors-with-atexit -fgnuc-version=4.2.1 -fcxx-exceptions -fexceptions -fmax-type-align=16 -fcolor-diagnostics -vectorize-loops -vectorize-slp -D__GCC_HAVE_DWARF2_CFI_ASM=1 -o /var/folders/_l/4q83bg9j5ysb7qd1n9xpnb4h0000gn/T/sha3-reproducer-beadcf.o -x c++ sha3-reproducer.cxx

clang -cc1 version 13.0.0 based upon LLVM 13.0.0 default target x86_64-apple-darwin20.6.0

ignoring nonexistent directory "/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/local/include"

ignoring nonexistent directory "/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/Library/Frameworks"

#include "..." search starts here:

#include <...> search starts here:

/usr/local/include

/opt/local/libexec/llvm-13/bin/../include/c++/v1

/opt/local/libexec/llvm-13/lib/clang/13.0.0/include

/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include

/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/System/Library/Frameworks (framework directory)

End of search list.

"/opt/local/libexec/llvm-13/bin/ld" -demangle -lto_library /opt/local/libexec/llvm-13/lib/libLTO.dylib -dynamic -arch x86_64 -platform_version macos 11.0.0 11.3 -syslibroot /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk -o t -L/usr/local/lib /var/folders/_l/4q83bg9j5ysb7qd1n9xpnb4h0000gn/T/sha3-reproducer-beadcf.o -lc++ -lSystem /opt/local/libexec/llvm-13/lib/clang/13.0.0/lib/darwin/libclang_rt.osx.a

$ ./t

Assertion failed: (T[0] == 16394434931424703552u), function main, file sha3-reproducer.cxx, line 104.

Abort trap: 6

$

 

though real fix isn't backported (see Dimitry's comment: https://bugs.llvm.org/show_bug.cgi?id=51957#c7).

 

See above – fails on Skylake.

 

And here, on Skylake-avx512 it seems to pass:

 

$ clang++-mp-13 -v $CXXFLAGS -O3 -march=native -o t sha3-reproducer.cxx 

clang version 13.0.0

Target: x86_64-apple-darwin20.6.0

Thread model: posix

InstalledDir: /opt/local/libexec/llvm-13/bin

 "/opt/local/libexec/llvm-13/bin/clang" -cc1 -triple x86_64-apple-macosx11.0.0 -Wundef-prefix=TARGET_OS_ -Werror=undef-prefix -Wdeprecated-objc-isa-usage -Werror=deprecated-objc-isa-usage -emit-obj --mrelax-relocations -disable-free -disable-llvm-verifier -discard-value-names -main-file-name sha3-reproducer.cxx -mrelocation-model pic -pic-level 2 -mframe-pointer=all -fno-rounding-math -munwind-tables -target-sdk-version=12.0 -fcompatibility-qualified-id-block-type-checking -fvisibility-inlines-hidden-static-local-var -target-cpu skylake-avx512 -target-feature +sse2 -target-feature -tsxldtrk -target-feature +cx16 -target-feature +sahf -target-feature -tbm -target-feature -avx512ifma -target-feature -sha -target-feature -gfni -target-feature -fma4 -target-feature -vpclmulqdq -target-feature +prfchw -target-feature +bmi2 -target-feature -cldemote -target-feature +fsgsbase -target-feature -ptwrite -target-feature -amx-tile -target-feature -uintr -target-feature +popcnt -target-feature -widekl -target-feature +aes -target-feature -avx512bitalg -target-feature -movdiri -target-feature +xsaves -target-feature -avx512er -target-feature -avxvnni -target-feature -avx512vnni -target-feature -amx-bf16 -target-feature -avx512vpopcntdq -target-feature -pconfig -target-feature +clwb -target-feature +avx512f -target-feature +xsavec -target-feature -clzero -target-feature -pku -target-feature +mmx -target-feature -lwp -target-feature -rdpid -target-feature -xop -target-feature +rdseed -target-feature -waitpkg -target-feature -kl -target-feature -movdir64b -target-feature -sse4a -target-feature +avx512bw -target-feature +clflushopt -target-feature +xsave -target-feature -avx512vbmi2 -target-feature +64bit -target-feature +avx512vl -target-feature -serialize -target-feature -hreset -target-feature +invpcid -target-feature +avx512cd -target-feature +avx -target-feature -vaes -target-feature -avx512bf16 -target-feature +cx8 -target-feature +fma -target-feature +rtm -target-feature +bmi -target-feature -enqcmd -target-feature +rdrnd -target-feature -mwaitx -target-feature +sse4.1 -target-feature +sse4.2 -target-feature +avx2 -target-feature +fxsr -target-feature -wbnoinvd -target-feature +sse -target-feature +lzcnt -target-feature +pclmul -target-feature -prefetchwt1 -target-feature +f16c -target-feature +ssse3 -target-feature -sgx -target-feature -shstk -target-feature +cmov -target-feature -avx512vbmi -target-feature -amx-int8 -target-feature +movbe -target-feature -avx512vp2intersect -target-feature +xsaveopt -target-feature +avx512dq -target-feature +adx -target-feature -avx512pf -target-feature +sse3 -debugger-tuning=lldb -target-linker-version 650.9 -v -fcoverage-compilation-dir=/Users/ur20980/src -resource-dir /opt/local/libexec/llvm-13/lib/clang/13.0.0 -isysroot /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk -stdlib=libc++ -internal-isystem /opt/local/libexec/llvm-13/bin/../include/c++/v1 -internal-isystem /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/local/include -internal-isystem /opt/local/libexec/llvm-13/lib/clang/13.0.0/include -internal-externc-isystem /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include -O3 -std=gnu++17 -fdeprecated-macro -fdebug-compilation-dir=/Users/ur20980/src -ferror-limit 19 -stack-protector 1 -fblocks -fencode-extended-block-signature -fregister-global-dtors-with-atexit -fgnuc-version=4.2.1 -fcxx-exceptions -fexceptions -fmax-type-align=16 -fcolor-diagnostics -vectorize-loops -vectorize-slp -D__GCC_HAVE_DWARF2_CFI_ASM=1 -o /var/folders/c6/lnc_0m093ys8w16md_fm1mnxhtfnj8/T/sha3-reproducer-5022a4.o -x c++ sha3-reproducer.cxx

clang -cc1 version 13.0.0 based upon LLVM 13.0.0 default target x86_64-apple-darwin20.6.0

ignoring nonexistent directory "/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/local/include"

ignoring nonexistent directory "/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/Library/Frameworks"

#include "..." search starts here:

#include <...> search starts here:

 /opt/local/libexec/llvm-13/bin/../include/c++/v1

 /opt/local/libexec/llvm-13/lib/clang/13.0.0/include

 /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include

 /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/System/Library/Frameworks (framework directory)

End of search list.

 "/opt/local/libexec/llvm-13/bin/ld" -demangle -lto_library /opt/local/libexec/llvm-13/lib/libLTO.dylib -dynamic -arch x86_64 -platform_version macos 11.0.0 12.0 -syslibroot /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk -o t /var/folders/c6/lnc_0m093ys8w16md_fm1mnxhtfnj8/T/sha3-reproducer-5022a4.o -lc++ -lSystem /opt/local/libexec/llvm-13/lib/clang/13.0.0/lib/darwin/libclang_rt.osx.a

$ ./t

$ 

 

 

I can't reproduce it on clang-13: https://godbolt.org/z/4Mdrd5388

 

I can’t reproduce it on clang-13 on Godbolt, but unfortunately, it’s 100% consistent on my machines. Also, I’m not certain the Godbolt site uses correct compiler. Here’s what it shows me:

 

 

So, requested Clang, but compiled with GCC???

 

 

вт, 26 окт. 2021 г. в 00:35, Blumenthal, Uri - 0553 - MITLL via llvm-dev <llvm...@lists.llvm.org>:

Reply all
Reply to author
Forward
0 new messages