Reports

My test shows: yes.

test before & after running (c:\usp\jobhp\ml_service_demo1\.conda) C:\usp\jobhp\ml_service_demo1>conda install pytorch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 pytorch-cuda=12.1 -c pytorch -c nvidia

The increment of space taken:
24.9GB-16.2GB=8.7GB

The folder that takes the actual space:
C:\Users\nshln\miniconda3\pkgs - 8.87GB

The folder that appears to take up space, but actually not, it uses hardlink:
C:\usp\jobhp\ml_service_demo1 - 6.27GB

terminal output:


(c:\usp\jobhp\ml_service_demo1\.conda) C:\usp\jobhp\ml_service_demo1>conda install pytorch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 pytorch-cuda=12.1 -c pytorch -c nvidia
Channels:
- pytorch
- nvidia
- defaults
Platform: win-64
Collecting package metadata (repodata.json): - 


                                            done
Solving environment: done

## Package Plan ##

  environment location: c:\usp\jobhp\ml_service_demo1\.conda

  added / updated specs:
    - pytorch-cuda=12.1
    - pytorch==2.5.1
    - torchaudio==2.5.1
    - torchvision==0.20.1


The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    blas-1.0                   |              mkl           6 KB
    brotli-python-1.0.9        |  py311h5da7b33_9         346 KB
    certifi-2025.4.26          |  py311haa95532_0         158 KB
    charset-normalizer-3.3.2   |     pyhd3eb1b0_0          44 KB
    cuda-cccl-12.9.27          |                0          16 KB  nvidia
    cuda-cccl_win-64-12.9.27   |                0         1.1 MB  nvidia
    cuda-cudart-12.1.105       |                0         964 KB  nvidia
    cuda-cudart-dev-12.1.105   |                0         549 KB  nvidia
    cuda-cupti-12.1.105        |                0        11.6 MB  nvidia
    cuda-libraries-12.1.0      |                0           1 KB  nvidia
    cuda-libraries-dev-12.1.0  |                0           1 KB  nvidia
    cuda-nvrtc-12.1.105        |                0        73.2 MB  nvidia
    cuda-nvrtc-dev-12.1.105    |                0        16.5 MB  nvidia
    cuda-nvtx-12.1.105         |                0          41 KB  nvidia
    cuda-opencl-12.9.19        |                0          17 KB  nvidia
    cuda-opencl-dev-12.9.19    |                0          62 KB  nvidia
    cuda-profiler-api-12.9.19  |                0          19 KB  nvidia
    cuda-runtime-12.1.0        |                0           1 KB  nvidia
    cuda-version-12.9          |                3          17 KB  nvidia
    filelock-3.17.0            |  py311haa95532_0          38 KB
    freeglut-3.4.0             |       hd77b12b_0         133 KB
    freetype-2.13.3            |       h0620614_0         554 KB
    giflib-5.2.2               |       h7edc060_0         105 KB
    gmp-6.3.0                  |       h537511b_0         330 KB
    gmpy2-2.2.1                |  py311h827c3e9_0         205 KB
    intel-openmp-2023.1.0      |   h59b6b97_46320         2.7 MB
    jinja2-3.1.6               |  py311haa95532_0         359 KB
    jpeg-9e                    |       h827c3e9_3         334 KB
    khronos-opencl-icd-loader-2024.05.08|       h8cc25b3_0         100 KB
    lcms2-2.16                 |       h62be587_1         568 KB
    lerc-4.0.0                 |       h5da7b33_0         185 KB
    libcublas-12.1.0.26        |                0          39 KB  nvidia
    libcublas-dev-12.1.0.26    |                0       348.3 MB  nvidia
    libcufft-11.0.2.4          |                0           6 KB  nvidia
    libcufft-dev-11.0.2.4      |                0       102.6 MB  nvidia
    libcurand-10.3.10.19       |                0        46.7 MB  nvidia
    libcurand-dev-10.3.10.19   |                0         242 KB  nvidia
    libcusolver-11.4.4.55      |                0          30 KB  nvidia
    libcusolver-dev-11.4.4.55  |                0        95.7 MB  nvidia
    libcusparse-12.0.2.55      |                0          12 KB  nvidia
    libcusparse-dev-12.0.2.55  |                0       162.5 MB  nvidia
    libdeflate-1.22            |       h5bf469e_0         180 KB
    libjpeg-turbo-2.0.0        |       h196d8e1_0         618 KB
    libnpp-12.0.2.50           |                0         305 KB  nvidia
    libnpp-dev-12.0.2.50       |                0       135.6 MB  nvidia
    libnvjitlink-12.1.105      |                0        67.3 MB  nvidia
    libnvjitlink-dev-12.1.105  |                0        13.8 MB  nvidia
    libnvjpeg-12.1.1.14        |                0           5 KB  nvidia
    libnvjpeg-dev-12.1.1.14    |                0         2.4 MB  nvidia
    libpng-1.6.39              |       h8cc25b3_0         369 KB
    libtiff-4.7.0              |       h404307b_0         1.1 MB
    libuv-1.48.0               |       h827c3e9_0         322 KB
    libwebp-1.3.2              |       h18467be_1          83 KB
    libwebp-base-1.3.2         |       h3d04722_1         303 KB
    lz4-c-1.9.4                |       h2bbff1b_1         152 KB
    markupsafe-3.0.2           |  py311h827c3e9_0          39 KB
    mkl-2023.1.0               |   h6b88ed4_46358       155.9 MB
    mkl-service-2.4.0          |  py311h827c3e9_2          67 KB
    mkl_fft-1.3.11             |  py311h827c3e9_0         176 KB
    mkl_random-1.2.8           |  py311hea22821_0         266 KB
    mpc-1.3.1                  |       h827c3e9_0          85 KB
    mpfr-4.2.1                 |       h56c3642_0         266 KB
    mpmath-1.3.0               |  py311haa95532_0         1.0 MB
    networkx-3.4.2             |  py311haa95532_0         3.1 MB
    numpy-2.0.1                |  py311hdab7c0b_1          11 KB
    numpy-base-2.0.1           |  py311hd01c5d8_1         9.6 MB
    openjpeg-2.5.2             |       h9b5d1b5_1         268 KB
    pillow-11.1.0              |  py311hea0d53e_1         916 KB
    pysocks-1.7.1              |  py311haa95532_0          36 KB
    pytorch-2.5.1              |py3.11_cuda12.1_cudnn9_0        1.21 GB  pytorch
    pytorch-cuda-12.1          |       hde6ce7c_6           7 KB  pytorch
    pytorch-mutex-1.0          |             cuda           3 KB  pytorch
    pyyaml-6.0.2               |  py311h827c3e9_0         205 KB
    requests-2.32.3            |  py311haa95532_1         128 KB
    sympy-1.13.3               |  py311haa95532_1        15.5 MB
    tbb-2021.8.0               |       h59b6b97_0         149 KB
    torchaudio-2.5.1           |      py311_cu121         7.2 MB  pytorch
    torchvision-0.20.1         |      py311_cu121         7.9 MB  pytorch
    urllib3-2.3.0              |  py311haa95532_0         242 KB
    win_inet_pton-1.1.0        |  py311haa95532_0          10 KB
    yaml-0.2.5                 |       he774522_0          62 KB
    zstd-1.5.6                 |       h8880b57_0         708 KB
    ------------------------------------------------------------
                                          Total:        2.47 GB

The following NEW packages will be INSTALLED:

  blas               pkgs/main/win-64::blas-1.0-mkl
  brotli-python      pkgs/main/win-64::brotli-python-1.0.9-py311h5da7b33_9
  certifi            pkgs/main/win-64::certifi-2025.4.26-py311haa95532_0
  charset-normalizer pkgs/main/noarch::charset-normalizer-3.3.2-pyhd3eb1b0_0
  cuda-cccl          nvidia/win-64::cuda-cccl-12.9.27-0
  cuda-cccl_win-64   nvidia/win-64::cuda-cccl_win-64-12.9.27-0
  cuda-cudart        nvidia/win-64::cuda-cudart-12.1.105-0
  cuda-cudart-dev    nvidia/win-64::cuda-cudart-dev-12.1.105-0
  cuda-cupti         nvidia/win-64::cuda-cupti-12.1.105-0
  cuda-libraries     nvidia/win-64::cuda-libraries-12.1.0-0
  cuda-libraries-dev nvidia/win-64::cuda-libraries-dev-12.1.0-0
  cuda-nvrtc         nvidia/win-64::cuda-nvrtc-12.1.105-0
  cuda-nvrtc-dev     nvidia/win-64::cuda-nvrtc-dev-12.1.105-0
  cuda-nvtx          nvidia/win-64::cuda-nvtx-12.1.105-0
  cuda-opencl        nvidia/win-64::cuda-opencl-12.9.19-0
  cuda-opencl-dev    nvidia/win-64::cuda-opencl-dev-12.9.19-0
  cuda-profiler-api  nvidia/win-64::cuda-profiler-api-12.9.19-0
  cuda-runtime       nvidia/win-64::cuda-runtime-12.1.0-0
  cuda-version       nvidia/noarch::cuda-version-12.9-3
  filelock           pkgs/main/win-64::filelock-3.17.0-py311haa95532_0
  freeglut           pkgs/main/win-64::freeglut-3.4.0-hd77b12b_0
  freetype           pkgs/main/win-64::freetype-2.13.3-h0620614_0
  giflib             pkgs/main/win-64::giflib-5.2.2-h7edc060_0
  gmp                pkgs/main/win-64::gmp-6.3.0-h537511b_0
  gmpy2              pkgs/main/win-64::gmpy2-2.2.1-py311h827c3e9_0
  intel-openmp       pkgs/main/win-64::intel-openmp-2023.1.0-h59b6b97_46320
  jinja2             pkgs/main/win-64::jinja2-3.1.6-py311haa95532_0
  jpeg               pkgs/main/win-64::jpeg-9e-h827c3e9_3
  khronos-opencl-ic~ pkgs/main/win-64::khronos-opencl-icd-loader-2024.05.08-h8cc25b3_0
  lcms2              pkgs/main/win-64::lcms2-2.16-h62be587_1
  lerc               pkgs/main/win-64::lerc-4.0.0-h5da7b33_0
  libcublas          nvidia/win-64::libcublas-12.1.0.26-0
  libcublas-dev      nvidia/win-64::libcublas-dev-12.1.0.26-0
  libcufft           nvidia/win-64::libcufft-11.0.2.4-0
  libcufft-dev       nvidia/win-64::libcufft-dev-11.0.2.4-0
  libcurand          nvidia/win-64::libcurand-10.3.10.19-0
  libcurand-dev      nvidia/win-64::libcurand-dev-10.3.10.19-0
  libcusolver        nvidia/win-64::libcusolver-11.4.4.55-0
  libcusolver-dev    nvidia/win-64::libcusolver-dev-11.4.4.55-0
  libcusparse        nvidia/win-64::libcusparse-12.0.2.55-0
  libcusparse-dev    nvidia/win-64::libcusparse-dev-12.0.2.55-0
  libdeflate         pkgs/main/win-64::libdeflate-1.22-h5bf469e_0
  libjpeg-turbo      pkgs/main/win-64::libjpeg-turbo-2.0.0-h196d8e1_0
  libnpp             nvidia/win-64::libnpp-12.0.2.50-0
  libnpp-dev         nvidia/win-64::libnpp-dev-12.0.2.50-0
  libnvjitlink       nvidia/win-64::libnvjitlink-12.1.105-0
  libnvjitlink-dev   nvidia/win-64::libnvjitlink-dev-12.1.105-0
  libnvjpeg          nvidia/win-64::libnvjpeg-12.1.1.14-0
  libnvjpeg-dev      nvidia/win-64::libnvjpeg-dev-12.1.1.14-0
  libpng             pkgs/main/win-64::libpng-1.6.39-h8cc25b3_0
  libtiff            pkgs/main/win-64::libtiff-4.7.0-h404307b_0
  libuv              pkgs/main/win-64::libuv-1.48.0-h827c3e9_0
  libwebp            pkgs/main/win-64::libwebp-1.3.2-h18467be_1
  libwebp-base       pkgs/main/win-64::libwebp-base-1.3.2-h3d04722_1
  lz4-c              pkgs/main/win-64::lz4-c-1.9.4-h2bbff1b_1
  markupsafe         pkgs/main/win-64::markupsafe-3.0.2-py311h827c3e9_0
  mkl                pkgs/main/win-64::mkl-2023.1.0-h6b88ed4_46358
  mkl-service        pkgs/main/win-64::mkl-service-2.4.0-py311h827c3e9_2
  mkl_fft            pkgs/main/win-64::mkl_fft-1.3.11-py311h827c3e9_0
  mkl_random         pkgs/main/win-64::mkl_random-1.2.8-py311hea22821_0
  mpc                pkgs/main/win-64::mpc-1.3.1-h827c3e9_0
  mpfr               pkgs/main/win-64::mpfr-4.2.1-h56c3642_0
  mpmath             pkgs/main/win-64::mpmath-1.3.0-py311haa95532_0
  networkx           pkgs/main/win-64::networkx-3.4.2-py311haa95532_0
  numpy              pkgs/main/win-64::numpy-2.0.1-py311hdab7c0b_1
  numpy-base         pkgs/main/win-64::numpy-base-2.0.1-py311hd01c5d8_1
  openjpeg           pkgs/main/win-64::openjpeg-2.5.2-h9b5d1b5_1
  pillow             pkgs/main/win-64::pillow-11.1.0-py311hea0d53e_1
  pysocks            pkgs/main/win-64::pysocks-1.7.1-py311haa95532_0
  pytorch            pytorch/win-64::pytorch-2.5.1-py3.11_cuda12.1_cudnn9_0
  pytorch-cuda       pytorch/win-64::pytorch-cuda-12.1-hde6ce7c_6
  pytorch-mutex      pytorch/noarch::pytorch-mutex-1.0-cuda
  pyyaml             pkgs/main/win-64::pyyaml-6.0.2-py311h827c3e9_0
  requests           pkgs/main/win-64::requests-2.32.3-py311haa95532_1
  sympy              pkgs/main/win-64::sympy-1.13.3-py311haa95532_1
  tbb                pkgs/main/win-64::tbb-2021.8.0-h59b6b97_0
  torchaudio         pytorch/win-64::torchaudio-2.5.1-py311_cu121
  torchvision        pytorch/win-64::torchvision-0.20.1-py311_cu121
  urllib3            pkgs/main/win-64::urllib3-2.3.0-py311haa95532_0
  win_inet_pton      pkgs/main/win-64::win_inet_pton-1.1.0-py311haa95532_0
  yaml               pkgs/main/win-64::yaml-0.2.5-he774522_0
  zstd               pkgs/main/win-64::zstd-1.5.6-h8880b57_0


Proceed ([y]/n)? y

before - 24.9GB:
after - 16.2GB:

find the hard link of one file by fsutil hardlink list <your_path>

nsht@AZIH C:\Users\nshln
$ fsutil hardlink list "C:/usp/jobhp/ml_service_demo1/.conda/Lib/site-packages/torch/lib/torch_cuda.dll"
\Users\nshln\miniconda3\pkgs\pytorch-2.5.1-py3.11_cuda12.1_cudnn9_0\Lib\site-packages\torch\lib\torch_cuda.dll
\usp\jobhp\ml_service_demo1\.conda\Lib\site-packages\torch\lib\torch_cuda.dll

this shows the file does have a hard link
(note: this file takes 850MB)

not important:

There is a related topic for a javascript node_modules package manager pnpm:
- You may think of:
  - javascript ~= python
  - pnpm ~= conda
  - npm ~= pip
  - node_modules ~= .venv / .conda
Why does my node_modules folder use disk space if packages are stored in a global store?

pnpm creates hard links from the global store to the project's node_modules folders. Hard links point to the same place on the disk where the original files are. So, for example, if you have foo in your project as a dependency and it occupies 1MB of space, then it will look like it occupies 1MB of space in the project's node_modules folder and the same amount of space in the global store. However, that 1MB is the same space on the disk addressed from two different locations. So in total foo occupies 1MB, not 2MB.

For more on this subject:
Does it work on Windows?

Short answer: Yes. Long answer: Using symbolic linking on Windows is problematic to say the least, however, pnpm has a workaround. For Windows, we use junctions instead.
- Frequently Asked Questions | pnpm https://pnpm.io/faq#does-it-work-on-windows

more related:

pnpm like installation mode to save space · Issue #3487 · python-poetry/poetry
https://github.com/python-poetry/poetry/issues/3487

python - Does anaconda support hardlink in Windows? - Stack Overflow
https://stackoverflow.com/questions/75657317/does-anaconda-support-hardlink-in-windows

pnpm for Python · pnpm · Discussion #3164
https://github.com/orgs/pnpm/discussions/3164

- $>"
  - **Use conda environments for isolation**
  - - Create a conda environment to isolate any changes pip makes.
    - Environments take up little space thanks to hard links.
  <$
  https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html

- $>"
  There’s not a whole lot to this. [Conda tries to use hard links whenever possible](https://github.com/conda/conda/blob/4.6.7/conda/gateways/disk/create.py#L297-L337). Hard links increase the reference count on disk to a specific file, without actually copying its contents.
  <$
  https://www.anaconda.com/blog/understanding-and-improving-condas-performance

pnpm, but for Python? - DEV Community
https://dev.to/ghost/pnpm-but-for-python-3hb3

79615702

Why does my `node_modules` folder use disk space if packages are stored in a global store?

Does it work on Windows?

79615702

Why does my node_modules folder use disk space if packages are stored in a global store?​

Does it work on Windows?​

Why does my `node_modules` folder use disk space if packages are stored in a global store?

Does it work on Windows?