Artem-B

#! /bin/sh

MODEL=${}

apt update
apt install -y nvtop htop btop tmux
tmux new-session 'nvtop' \; \
     split-window 'python3 download-model.py --threads 15 ${1:-}'

Artem-B

CUDA devices found: 2
Device 0: "NVIDIA RTX 5000 Ada Generation", Selected, SM89, 33937293312 [bytes]
Device 1: "Quadro GP100", Unused, SM60, 17064263680 [bytes]
Testing void test_edges() [T = double] on host
Testing void test_edges() [T = __half] on host
Testing void test_edges() [T = __nv_bfloat16] on host
Testing void test_edges() [T = double] on GPU
imag 0:1 i=1.000000e-02:1.000000e-02 j=-1.000000e-02:1.000000e-02 r=1.034025e+00:-5.220046e-02 z=1.034025e+00:-5.220046e-02 diff=0.000000e+00:6.938894e-18
imag 0:3 i=1.000000e-02:1.000000e-02 j=1.000000e-02:-1.000000e-02 r=9.646358e-01:4.869749e-02 z=9.646358e-01:4.869749e-02 diff=0.000000e+00:-6.938894e-18
real 0:5 i=1.000000e-02:1.000000e-02 j=-1.000000e+02:1.000000e-02 r=-8.804310e+184:3.751669e+183 z=-8.804310e+184:3.751669e+183 diff=-6.038340e+169:-1.251069e+171

Artem-B

export HF_HUB_ENABLE_HF_TRANSFER=1 &&
apt update &&
apt install -y nvtop &&
pip install hf_transfer &&
export HF_HUB_ENABLE_HF_TRANSFER=1 

MODEL=MikeRoz/MarsupialAI_Dumbstral-169B-5.0bpw-h6-exl2
MODEL=BigHuggyD/MarsupialAI_Monstral-123B-v2_exl2_8.0bpw_h8
MODEL BigHuggyD/TheDrummer_Behemoth-123B-v1.2_exl2_7.0bpw_h8

Artem-B

{
  "t1730931331339": {
    "title": "CCCL nvcc+clang_libc++",
    "data": "z:OYLghAFBqd5QCxAYwPYBMCmBRdBLAF1QCcAaPECAMzwBtMA7AQwFtMQByARg9KtQYEAysib0QXACx8BBAKoBnTAAUAHpwAMvAFYTStJg1DIAruiakl9ZATwDKjdAGFUtEywZ7HAGTwNMAHLuAEaYxCAA7ACspAAOqAqEdgwubh568Ym2Ar7%2BQSyh4dGWmNbZDEIETMQEqe6eXCVlyZXVBLmBIWGRMQpVNXXpjX1tHfmFPQCUlqgmxMjsHACkAEwAzH7IblgA1EtrTqbmAPR96McA%2BhdsLCQAnhf72EsaAIIv7%2BtYNP47ACrYIR/C5OADyASB2AAGsoAEogqFQlYaD6rNbfPyYf6A4FgiEAmGw1FvAiYFixAykvYHLZMBQKf5PD59YgmGw7AKoAAimBZqDuqIiACEPjsdgQ7rFMN9/jsAG5iEyYC4SqX7EVvUU7K4IBIEK7ai5YOV4BYGgFAkHgyGEhFIjT/ABUOzEtFQolJECOTBAIDOvsSAC9lQQdgxJlqlsKtWLiJgCHMGDtvQGCOhfa73UwiMR9k4/k8IJMAHSZj2YCDh9WRiJc4mvMU6vUGq7G03Ki7Yy14m1wu3I%2BWoPDoHZYMvZit/Z2xUjJsw%2Bv1pgN4YMqsMRt5iqMahtinZxhPEJMpxfpkDjnN5gtrbBF4tj2hu8sQGfr6ubva1wV1tY7j6k8lKSxPNaXpRkb2ZAhWXZFwTEEHk%2BQFTVow/Js%2BhbI1MBNM1OwtXFrQJOEdkwVQKVNQgdlg%2BDeSg/kID8AhnTQODSWIDddz3HYQGTWZBDCC4vV41j2K3aMv01VCLl1dDO1bLD23NHErXxaE%2BycREBwAPyoggENou4i0/EVxPeD9VWlTAqFlBU3BDSVMHfUzdzQ/VZMw7COy7fCVNtdT7SdF1HyzT0T39P0VxDN8P23GN

Artem-B

; ModuleID = 'LLVMDialectModule'
source_filename = "LLVMDialectModule"
target datalayout = "e-p6:32:32-i64:64-i128:128-v16:16-v32:32-n16:32:64"

@global_smem = external addrspace(3) global [0 x i8], align 16

define ptx_kernel void @kernel(ptr addrspace(1) %0, ptr addrspace(1) %1, ptr addrspace(1) readnone captures(none) %2) local_unnamed_addr #0 !dbg !6 {
  %4 = tail call i32 @llvm.nvvm.read.ptx.sreg.tid.x(), !dbg !9
  %5 = and i32 %4, 31, !dbg !9
  %6 = lshr i32 %4, 5, !dbg !9

Artem-B

[  1] ; RUN: not llc < %s -march=nvptx64 -mcpu=sm_100 -mattr=+ptx88 2>&1 | FileCheck %s [FAIL]
not llc < /build/work/c6bdb884076ec4b5dd33a50ce69e302b9272/google3/runfiles/google3/third_party/llvm/llvm-project/llvm/test/CodeGen/NVPTX/load-store-atomic.err.ll -march=nvptx64 -mcpu=sm_100 -mattr=+ptx88 2>&1 | /build/work/c6bdb884076ec4b5dd33a50ce69e302b9272/google3/runfiles/google3/third_party/llvm/llvm-project/llvm/FileCheck --allow-unused-prefixes /build/work/c6bdb884076ec4b5dd33a50ce69e302b9272/google3/runfiles/google3/third_party/llvm/llvm-project/llvm/test/CodeGen/NVPTX/load-store-atomic.err.ll
/build/work/c6bdb884076ec4b5dd33a50ce69e302b9272/google3/runfiles/google3/third_party/llvm/llvm-project/llvm/test/CodeGen/NVPTX/load-store-atomic.err.ll:5:10: error: CHECK: expected string not found in input
; CHECK: error: unsupported atomic store
         ^
<stdin>:2:31: note: scanning from here
error: unsupported atomic load
                              ^
<stdin>:11:69: note: possible intended match here
 // .glob

Artem-B

File: k101a.o
ELF Header:
  Magic:   7f 45 4c 46 02 01 01 41 08 00 00 00 00 00 00 00
  Class:                             ELF64
  Data:                              2's complement, little endian
  Version:                           1 (current)
  OS/ABI:                            41
  ABI Version:                       8
  Type:                              EXEC (Executable file)