[REVIEW] Feat: PR #2 decoding with nvImageCodec v0.6.0 #978

cdinea · 2025-11-21T02:50:16Z

cuslide2: GPU-Accelerated decoding via nvImageCodec v0.6.0

Overview

cucim.kit.cuslide2 plugin implementing GPU-accelerated whole-slide imaging (WSI) decoding using nvImageCodec v0.6.0. Replaces CPU-based decoders (libjpeg-turbo, OpenJPEG, libtiff) with GPU equivalents

Vendor Support:

✅ Aperio SVS (JPEG, JPEG2000)
✅ Philips TIFF (JPEG, multi-resolution pyramids)
✅ Generic TIFF (all compression types)

Key Features

Direct ROI Decoding: nvImageCodec decodes regions without loading full tiles
GPU Memory Path: Direct CUDA allocation, no CPU→GPU copy overhead
Metadata Parsing: Enhanced Aperio/Philips XML metadata extraction with nvImageCodec v0.6.0 workarounds
Backward Compatible: Maintains same API as cucim.kit.cuslide

Build Instructions

Create Conda Environment:

# Create conda environment with all dependencies based on your CUDA system version - use all_cuda-130 for CUDA 13 and all_cuda-129 for CUDA 12
conda env create -n cucim -f ./conda/environments/all_cuda-130_arch-x86_64.yaml

# Activate environment
conda activate cucim

Build cuslide2 Plugin:

# Set CUDA compiler (uses conda's CUDA toolkit)
export CUDACXX=$CONDA_PREFIX/pkgs/cuda-toolkit/bin/nvcc

# Build libcucim + cuslide2 plugin
./run build_local all release $CONDA_PREFIX

Install Python Package:

# Install in editable mode for development
python -m pip install --editable python/cucim

Testing

Unit Tests (26 tests)

cd python/cucim
python -m pytest tests/unit/clara/test_tiff_read_region.py -v

Expected Output:

================================== test session starts ================================== platform linux -- Python 3.13.9, pytest-8.4.2, pluggy-1.6.0 -- /home/cdinea/miniconda3/envs/cucimcuda/bin/python cachedir: .pytest_cache rootdir: /home/cdinea/Downloads/cucim_pr2/branchremote/cucim/python/cucim configfile: pyproject.toml plugins: cov-7.0.0, xdist-3.8.0, lazy-fixtures-1.4.0 collected 26 items tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_inner[testimg_tiff_stripe_32x24_16_jpeg] PASSED [ 3%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_boundary[testimg_tiff_stripe_32x24_16_jpeg] PASSED [ 7%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_outside[testimg_tiff_stripe_32x24_16_jpeg] PASSED [ 11%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_inner[testimg_tiff_stripe_32x24_16_deflate] PASSED [ 15%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_boundary[testimg_tiff_stripe_32x24_16_deflate] PASSED [ 19%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_outside[testimg_tiff_stripe_32x24_16_deflate] PASSED [ 23%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_inner[testimg_tiff_stripe_32x24_16_raw] PASSED [ 26%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_boundary[testimg_tiff_stripe_32x24_16_raw] PASSED [ 30%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_outside[testimg_tiff_stripe_32x24_16_raw] PASSED [ 34%] tests/unit/clara/test_tiff_read_region.py::test_tiff_outside_of_resolution_level[testimg_tiff_stripe_4096x4096_256_jpeg] PASSED [ 38%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_multiresolution[testimg_tiff_stripe_4096x4096_256_jpeg] PASSED [ 42%] tests/unit/clara/test_tiff_read_region.py::test_region_image_level_data[testimg_tiff_stripe_4096x4096_256_jpeg] PASSED [ 46%] tests/unit/clara/test_tiff_read_region.py::test_region_image_dtype[testimg_tiff_stripe_4096x4096_256_jpeg] PASSED [ 50%] tests/unit/clara/test_tiff_read_region.py::test_tiff_iterator[testimg_tiff_stripe_4096x4096_256_jpeg] PASSED [ 53%] tests/unit/clara/test_tiff_read_region.py::test_tiff_outside_of_resolution_level[testimg_tiff_stripe_4096x4096_256_deflate] PASSED [ 57%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_multiresolution[testimg_tiff_stripe_4096x4096_256_deflate] PASSED [ 61%] tests/unit/clara/test_tiff_read_region.py::test_region_image_level_data[testimg_tiff_stripe_4096x4096_256_deflate] PASSED [ 65%] tests/unit/clara/test_tiff_read_region.py::test_region_image_dtype[testimg_tiff_stripe_4096x4096_256_deflate] PASSED [ 69%] tests/unit/clara/test_tiff_read_region.py::test_tiff_iterator[testimg_tiff_stripe_4096x4096_256_deflate] PASSED [ 73%] tests/unit/clara/test_tiff_read_region.py::test_tiff_outside_of_resolution_level[testimg_tiff_stripe_4096x4096_256_raw] PASSED [ 76%] tests/unit/clara/test_tiff_read_region.py::test_tiff_stripe_multiresolution[testimg_tiff_stripe_4096x4096_256_raw] PASSED [ 80%] tests/unit/clara/test_tiff_read_region.py::test_region_image_level_data[testimg_tiff_stripe_4096x4096_256_raw] PASSED [ 84%] tests/unit/clara/test_tiff_read_region.py::test_region_image_dtype[testimg_tiff_stripe_4096x4096_256_raw] PASSED [ 88%] tests/unit/clara/test_tiff_read_region.py::test_tiff_iterator[testimg_tiff_stripe_4096x4096_256_raw] PASSED [ 92%] tests/unit/clara/test_tiff_read_region.py::test_array_interface_support PASSED [ 96%] tests/unit/clara/test_tiff_read_region.py::test_cuda_array_interface_support PASSED [100%] ================================== 26 passed in 4.04s ===================================

Aperio SVS Validation

python scripts/test_aperio_svs.py --download

Tests:

Multi-level pyramid (3 levels expected)
GPU decode correctness (1.90× speedup on 2048×2048 tiles)
CPU/GPU output comparison
ImageDescription XML metadata parsing
Associated images (label, macro, thumbnail)

Key Results:

✅ File loads in 0.380s
✅ GPU decode: 0.0092s (2048×2048 tile)
✅ CPU decode: 0.0174s (GPU→CPU copy)
✅ 1.90× GPU speedup demonstrated

`📥 Downloading Aperio SVS test file...
✅ Test file already exists: /tmp/CMU-1-Small-Region.svs
✅ Plugin configuration: /tmp/.cucim_aperio_test.json
🔬 Testing cuslide2 plugin with Aperio SVS

📁 File: /tmp/CMU-1-Small-Region.svs
📂 Loading SVS file...
🔧 Creating IFD[0] from nvImageCodec metadata
Dimensions: 2220x2967, 3 channels, 8 bits/sample
Codec: jpeg (compression=7)
✅ IFD[0] initialization complete
🔧 Creating IFD[1] from nvImageCodec metadata
Dimensions: 574x768, 3 channels, 8 bits/sample
✅ IFD[1] initialization complete
🔧 Creating IFD[2] from nvImageCodec metadata
Dimensions: 387x463, 3 channels, 8 bits/sample
✅ IFD[2] initialization complete
🔧 Creating IFD[3] from nvImageCodec metadata
Dimensions: 1280x431, 3 channels, 8 bits/sample
✅ IFD[3] initialization complete
✅ Loaded in 0.380s
📊 Image Information:
Dimensions: [2967, 2220, 3]
Levels: 3
Device: cpu
🔍 Resolution Levels:
Level 0: 2220x2967 (downsample: 1.0x)
Level 1: 1280x431 (downsample: 4.3x)
Level 2: 387x463 (downsample: 6.1x)
🚀 Testing GPU decode (nvImageCodec)...
✅ GPU decode successful!
Time: 0.5288s
Shape: [512, 512, 3]
🖥️ Testing CPU decode (baseline)...
✅ CPU decode successful!
Time: 0.0029s
📏 Testing larger tile (2048x2048)...
GPU: 0.0092s
CPU: 0.0174s
🎯 Speedup: 1.90x
✅ Test completed successfully!
`

Philips TIFF Validation

python scripts/test_philips_tiff.py /tmp/Philips-1.tiff

Tests:

Multi-level pyramid (8 levels expected)
GPU decode correctness
Philips XML metadata (DPUfsImport format)
Associated images (label, macro, thumbnail)
DICOM_PIXEL_SPACING extraction

Key Results:

✅ 8 resolution levels detected
✅ 22 metadata entries extracted
✅ DICOM pixel spacing: 0.2269 × 0.2269 μm/pixel
✅ Multi-level reads: <0.001s per level

`✅ Plugin configuration: /tmp/.cucim_philips_test.json

🔬 Testing Philips TIFF with cuslide2

📁 File: /tmp/Philips-1.tiff
📂 Loading Philips TIFF file...
🔧 Creating IFD[0] from nvImageCodec metadata
Dimensions: 45056x35840, 3 channels, 8 bits/sample
✅ IFD[0] initialization complete
[... 7 more IFDs ...]
✅ Loaded in 0.379s
📊 Image Information:
Format: Philips TIFF
Dimensions: [35840, 45056, 3]
Levels: 8
🔍 Resolution Levels:
Level 0: 45056x35840 (downsample: 1.0x)
Level 1: 22528x17920 (downsample: 2.0x)
Level 2: 11264x8960 (downsample: 4.0x)
Level 3: 5632x4480 (downsample: 8.0x)
Level 4: 2816x2240 (downsample: 16.0x)
Level 5: 1408x1120 (downsample: 32.0x)
Level 6: 704x560 (downsample: 64.0x)
Level 7: 352x280 (downsample: 128.0x)
📋 Philips Metadata:
✅ Found 22 Philips metadata entries
DICOM_PIXEL_SPACING: [0.000226891, 0.000226907]
DICOM_MANUFACTURER: Hamamatsu
PIM_DP_IMAGE_TYPE: WSI
... and 16 more entries
📏 Pixel Spacing:
DICOM Pixel Spacing: 0.2269 x 0.2269 μm/pixel
🚀 Testing GPU decode (nvImageCodec)...
✅ GPU decode successful!
Time: 0.5250s
Shape: [512, 512, 3]
🖥️ Testing CPU decode...
✅ CPU decode successful:
Time: 0.0014s
Pixel sum: 189181125, mean: 240.56
📏 Testing larger tile (2048x2048)...
✅ GPU: 0.0168s
🔀 Testing multi-level reads...
✅ Level 0: 0.0010s ([512, 512, 3])
✅ Level 1: 0.0009s ([512, 512, 3])
✅ Level 2: 0.0007s ([512, 512, 3])
✅ Philips TIFF test completed!
`

Technical Implementation

Decoding Pipeline:

Parse TIFF: NvImageCodecTiffParser extracts IFD metadata using nvimgcodecCodeStreamCreateFromFile
Create Sub-Streams: Per-IFD code streams via nvimgcodecCodeStreamCreateFromCodeStreamByIndex
ROI Decode: nvimgcodecDecoderDecode with region parameters (x, y, width, height)
Memory Management: Direct cudaMalloc for GPU or cucim_malloc for CPU

Metadata Workarounds (nvImageCodec v0.6.0):

Aperio: Detect from ImageDescription starting with "Aperio " OR metadata blob kind=1 (MED_APERIO)
Philips: Detect from ImageDescription XML (<DataObject ObjectType="DPUfsImport">) OR metadata blob kind=2 (MED_PHILIPS)
Known limitation: v0.6.0 misclassifies Aperio as Leica (kind=3) and Philips as Ventana (kind=4) - workarounds implemented

Known Limitations

nvImageCodec v0.6.0 doesn't expose individual TIFF tags (SOFTWARE, etc.) - requires metadata blob inspection
Associated images with strip-based layout require full decode (no ROI support for strips)
Base64-encoded associated images (macro/label in XML metadata) not yet supported

Migration Guide

No code changes required for existing cuCIM users:

# Same API works with cuslide2
from cucim import CuImage
img = CuImage("slide.svs")
region = img.read_region(location=(0, 0), size=(1024, 1024), device="cuda")

- Fix libjpeg-turbo cmake configuration for both cuslide and cuslide2 - Update nvimgcodec cmake dependency configuration - Update examples CMakeLists - Update build scripts and documentation

…MakeLists

…e-only PR

…vironment files

…ipt location

Signed-off-by: cdinea <[email protected]>

Co-authored-by: jakirkham <[email protected]>

Signed-off-by: cdinea <[email protected]>

…a kinds

mkepa-nv · 2025-11-26T11:19:04Z

cpp/plugins/cucim.kit.cuslide2/src/cuslide/nvimgcodec/nvimgcodec_decoder.cpp

+            fmt::print("  ℹ️  Using CPU buffer for ROI decoding\n");
+            #endif
+        }
+        else if (gpu_available)


Can gpu not be available? Is this a valid case if target is gpu? Maybe you should error instead? User is probably expecting to receive device buffer in this case, and if we return host one it can break things

thank you for catching this @mkepa-nv - i addressed this bug in this commit

mkepa-nv · 2025-11-26T11:25:45Z

cpp/plugins/cucim.kit.cuslide2/src/cuslide/nvimgcodec/nvimgcodec_decoder.cpp

+        #endif
+        return true;  // roi_stream, image, decode_future cleaned up by RAII
+    }
+    catch (const std::exception& e)


This catch and others in nvimgodec/ directory are probably not needed. Wat can throw here? For sure CUDA and nvImageCodec API will not throw

thank you for the feedback @mkepa-nv -CUDA and nvImageCodec APIs don't throw - they return error codes. However, the catch blocks might still necessary because:

we throw std::runtime_error for error conditions (e.g., GPU availability check)

fmt::format() - can throw on formatting errors or memory allocation failures

-string operations, vector resizing, etc. can throw std::bad_alloc

mkepa-nv · 2025-11-26T11:44:54Z

cpp/plugins/cucim.kit.cuslide2/src/cuslide/nvimgcodec/nvimgcodec_tiff_parser.cpp

+    {
+        switch (kind)
+        {
+            case 1:  // NVIMGCODEC_METADATA_KIND_MED_APERIO


I would use enum named values from nvImageCodec header

already addressed @mkepa-nv -I've replaced all hardcoded integer values with proper enum constants

jantonguirao · 2025-11-25T12:18:26Z

cpp/plugins/cucim.kit.cuslide/src/cuslide/tiff/ifd.cpp

            {
-                loader->enqueue(std::move(decode_func),
-                                cucim::loader::TileInfo{ location_index, index, tiledata_offset, tiledata_size });
+                fmt::print("🔍 Executing decode_func directly (FORCED SINGLE-THREADED)\n");


Are all those new prints temporary?

thank you for the feedback @jantonguirao - i wrapped debug prints in #ifdef DEBUG guards in this commit

jantonguirao · 2025-11-25T12:21:46Z

cpp/plugins/cucim.kit.cuslide2/CMakeLists.txt

+if (TARGET CUDA::nvjpeg_static)
+        target_link_libraries(${CUCIM_PLUGIN_NAME}
+            PRIVATE
+                # Add nvjpeg before cudart so that nvjpeg.h in static library takes precedence.
+                CUDA::nvjpeg_static
+                # Add CUDA::culibos to link necessary methods for 'deps::nvjpeg_static'
+                CUDA::culibos
+                CUDA::cudart
+        )
+else()
+        target_link_libraries(${CUCIM_PLUGIN_NAME}
+            PRIVATE
+                CUDA::nvjpeg
+                CUDA::cudart
+        )
+endif()


nvImageCodec uses dlopen to load the nvjpeg shared-object library from system. Are we using nvjpeg directly anywhere?

thank you for the feedback @jantonguirao - this is a good obeservation - currently cuslide2 is currently reusing ifd.cpp from the original cuslide plugin, which uses NvJpegProcessor for batch JPEG decoding. So the nvjpeg linking is needed -we couuld refactor ifd.cpp to use pure nvImageCodec and remove the nvjpeg dependency

jantonguirao · 2025-11-25T12:32:09Z

cpp/plugins/cucim.kit.cuslide2/benchmarks/main.cpp

@@ -0,0 +1,256 @@
+/*
+ * Copyright (c) 2020-2022, NVIDIA CORPORATION.


nit:

Suggested change

* Copyright (c) 2020-2022, NVIDIA CORPORATION.

* Copyright (c) 2020-2025, NVIDIA CORPORATION.

thank you @jantonguirao - updated the year in the header in this commit

jantonguirao · 2025-11-25T13:37:06Z

cpp/plugins/cucim.kit.cuslide2/src/cuslide/nvimgcodec/nvimgcodec_decoder.h

+                                  uint32_t width, uint32_t height,
+                                  uint8_t** output_buffer,
+                                  const cucim::io::Device& out_device);
+#endif // CUCIM_HAS_NVIMGCODEC


You defined a dummy version of decode_ifd_region_nvimgcodec for the case that CUCIM_HAS_NVIMGCODEC is false (nvimgcodec_decoder.cpp:379) . Shouldn't you expose this declaration unconditionally then?

thank you for the feedback @jantonguirao - good catch , I addressed this feedback inthis commit

jantonguirao · 2025-11-25T14:18:23Z

cpp/plugins/cucim.kit.cuslide2/tests/main.cpp

@@ -0,0 +1,95 @@
+/*
+ * Copyright (c) 2021, NVIDIA CORPORATION.


thank you for the feedback @jantonguirao - I update the year in the copyright header in this commit

jantonguirao · 2025-11-25T14:20:34Z

cpp/src/loader/thread_batch_data_loader.cpp


 uint32_t ThreadBatchDataLoader::request(uint32_t load_size)
 {
+#ifdef DEBUG


If those debug messages are meant to stay, I'd recommend a preprocessor macro that would wrap fmt::print with the ifdef DEBUG condition. This way we avoid cluttering the codebase with many #ifdef DEBUG statements.

this is a good point @jantonguirao - I made note to address this in a subsequent PR as this might affect the first cuslide plugin and we need to properly test this

jantonguirao · 2025-11-25T14:22:53Z

scripts/test_aperio_svs.py

+    plugin_lib = repo_root / "cpp/plugins/cucim.kit.cuslide2/build-release/lib"
+
+    if not plugin_lib.exists():
+        plugin_lib = repo_root / "install/lib"


nit:

Suggested change

plugin_lib = repo_root / "install/lib"

plugin_lib = repo_root / "install" / "lib"

thank you @jantonguirao - good suggestion , addressed inthis commit

jantonguirao · 2025-11-25T14:44:27Z

cpp/plugins/cucim.kit.cuslide2/tests/test_read_rawtiff.cpp

+    }
+}
+
+TEST_CASE("Verify raw tiff read", "[test_read_rawtiff.cpp]")


I would like to see some comment explaining what this test does.

thank you @jantonguirao - i added a comment in [this(https://github.com//pull/978/commits/bab2aaadaa32032d42c8c9327efb838b489d6c00). @gigony FYI

…ility

cdinea · 2025-11-29T01:02:31Z

/ok to test bab2aaa

cdinea · 2025-12-01T19:59:12Z

/ok to test 9a0e8f5

cdinea and others added 30 commits November 19, 2025 16:01

refactor: Improve nvImageCodec integration and conda configuration

36e975c

feat: Consolidate nvImageCodec dependencies in dependencies.yaml

b75ba5f

build: Update CMake and build configurations for cuslide2

fc1986b

- Fix libjpeg-turbo cmake configuration for both cuslide and cuslide2 - Update nvimgcodec cmake dependency configuration - Update examples CMakeLists - Update build scripts and documentation

chore: Add PR_DESCRIPTION.md to .gitignore

4c688d8

feat: Add comprehensive nvImageCodec installation verification script

d3e322e

refactor: Remove nvImageCodec plugin implementation references from C…

13670f7

…MakeLists

Add CMake utility modules for cuslide2 plugin

678c40a

fix: Comment out cuslide2_nvimagecodec_test example for infrastructur…

fec4ee5

…e-only PR

build: Add missing CMake dependency files for cuslide2 plugin

e164af8

revert: Remove cuslide2 example from general examples directory

f737c12

revert: Restore original notebook README and clean .gitignore

67a37fe

fix: Disable tests/benchmarks and add missing config file for cuslide2

4486620

Add libnvimgcodec dependencies to dependencies.yaml and regenerate en…

b439403

…vironment files

Use yasm instead of nasm for libjpeg-turbo assembly compilatio

68fc1e2

Add nvImageCodec installation verification script

119de8c

Update pyproject.toml with nvimgcodec dependencies and remove old scr…

405c5fd

…ipt location

Add comprehensive cuslide2 infrastructure verification script

28a2ef9

Refactor verification script and fix build configuration

b0d9634

Add nvImageCodec infrastructure setup for cuslide2

ad0d955

style: Apply SPDX license headers and formatting fixes

60385e0

fixed with ruff per Greg

5f8a8cd

Signed-off-by: cdinea <[email protected]>

added comment detail

543d629

Signed-off-by: cdinea <[email protected]>

fix: Update copyright headers to SPDX format and fix typo

17148e2

Signed-off-by: cdinea <[email protected]>

added comment detail

4514538

Signed-off-by: cdinea <[email protected]>

Update dependencies.yaml

8c594d6

Co-authored-by: jakirkham <[email protected]>

Update dependencies.yaml per john's suuggestion

dd7d107

Co-authored-by: jakirkham <[email protected]>

Update dependencies.yaml per john's suggestion

85dce8e

Co-authored-by: jakirkham <[email protected]>

Update dependencies.yaml per john's suggestion

6076d49

Co-authored-by: jakirkham <[email protected]>

build: Regenerate conda environment files from dependencies.yaml

7bd8d9b

Signed-off-by: cdinea <[email protected]>

build: Fix spacing in pyproject.toml nvidia-nvimgcodec dependency

27acb53

Signed-off-by: cdinea <[email protected]>

cdinea added 4 commits November 25, 2025 09:32

Remove redundant initializer for initialized_ member

59543e5

Fix ROI decoding to use CPU-only backend when appropriate

bee35c9

Remove dead GPU-to-CPU copy code

61f8ab2

Refactor nvimgcodec_decoder to use RAII pattern for resource management

b6b5b30

cdinea added non-breaking Introduces a non-breaking change and removed breaking Introduces a breaking change labels Nov 25, 2025

cdinea added 4 commits November 25, 2025 10:49

Use pinned memory for host buffers for faster GPU-to-host transfers

8a4a6fc

Check return value of nvimgcodecFutureGetProcessingStatus

ec838c3

Consolidate DEBUG print statements into single ifdef block

4775952

Add consistent empty() check for image_description across all metadat…

4d85bea

…a kinds

cdinea added this to the v25.12.00 milestone Nov 25, 2025

mkepa-nv reviewed Nov 26, 2025

View reviewed changes

jantonguirao reviewed Nov 28, 2025

View reviewed changes

cdinea added 11 commits November 28, 2025 14:11

Add runtime version detection for nvImageCodec v0.6.0/v0.7.0 compatib…

2e138d1

…ility

Eliminate unnecessary metadata copy in extract_ifd_metadata

5f0db51

Replace hardcoded metadata kind values with nvImageCodec enum constants

85e98a6

Fix GPU availability check: error instead of silent fallback to CPU

a137953

Fix build: use only nvImageCodec v0.6.0 enum constants

9ca0ea6

Wrap debug prints in #ifdef DEBUG guards

e82293d

Address reviewer feedback: wrap debug prints and update copyright

ea3b269

Fix: expose decode_ifd_region_nvimgcodec declaration unconditionally

ec87ee8

Fix pathlib usage: use separate path segments

a0807e9

Add documentation comment for raw TIFF I/O benchmark test

bab2aaa

cdinea and others added 2 commits December 1, 2025 11:57

Fix code style per pre-commit hooks

742b066

Merge branch 'release/25.12' into feature/nvimagecodec-v0.6.0

9a0e8f5

cdinea mentioned this pull request Dec 1, 2025

[REVIEW] Feat: PR #1 nvImageCodec infrastructure only #966

Open

		@@ -0,0 +1,256 @@
		/*
		* Copyright (c) 2020-2022, NVIDIA CORPORATION.

	* Copyright (c) 2020-2022, NVIDIA CORPORATION.
	* Copyright (c) 2020-2025, NVIDIA CORPORATION.

	plugin_lib = repo_root / "install/lib"
	plugin_lib = repo_root / "install" / "lib"

[REVIEW] Feat: PR #2 decoding with nvImageCodec v0.6.0 #978

Are you sure you want to change the base?

[REVIEW] Feat: PR #2 decoding with nvImageCodec v0.6.0 #978

Uh oh!

Conversation

cdinea commented Nov 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

cuslide2: GPU-Accelerated decoding via nvImageCodec v0.6.0

Overview

Key Features

Build Instructions

Testing

Unit Tests (26 tests)

Aperio SVS Validation

`📥 Downloading Aperio SVS test file... ✅ Test file already exists: /tmp/CMU-1-Small-Region.svs ✅ Plugin configuration: /tmp/.cucim_aperio_test.json 🔬 Testing cuslide2 plugin with Aperio SVS

Philips TIFF Validation

`✅ Plugin configuration: /tmp/.cucim_philips_test.json

🔬 Testing Philips TIFF with cuslide2

Technical Implementation

Known Limitations

Migration Guide

Uh oh!

mkepa-nv Nov 26, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mkepa-nv Nov 26, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cdinea commented Nov 29, 2025

Uh oh!

cdinea commented Dec 1, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

cdinea commented Nov 21, 2025 •

edited

Loading

`📥 Downloading Aperio SVS test file...
✅ Test file already exists: /tmp/CMU-1-Small-Region.svs
✅ Plugin configuration: /tmp/.cucim_aperio_test.json
🔬 Testing cuslide2 plugin with Aperio SVS

mkepa-nv Nov 26, 2025 •

edited

Loading

mkepa-nv Nov 26, 2025 •

edited

Loading