Docker Image Availability: The CUDA 13.0.1 image reference in your Dockerfile (nvidia/cuda:13.0.1-cudnn-devel-ubuntu24.04) should be verified on Docker Hub. CUDA 13.0 was released in August 2025, and image availability may vary.

ROCm 7.1 Preview: AMD has released ROCm 7.1 preview (October 2025). You may want to consider testing with this version for early access to new features.
Breaking Changes:
CUDA 13.0 dropped support for Maxwell/Pascal/Volta architectures
HIP 7.0 introduced behavioral changes (see AMD documentation)

Author: coketaste

.gitignore

@@ -162,4 +162,7 @@ pmc_perf_*
 *.pid
 
 # Memory dumps
-*.dmp
+*.dmp
+
+.cursor/
+AGENTS.md

README.md

@@ -20,7 +20,7 @@ After completing this module, you will be able to:
 
 ### Prerequisites
 - NVIDIA GPU with CUDA support OR AMD GPU with ROCm support
-- CUDA Toolkit 13.0+ or ROCm 7.0+ (Docker images provide CUDA 13.0.1 and ROCm 7.0.1)
+- CUDA Toolkit 13.0+ or ROCm 7.0+ (Docker images provide CUDA 13.0.1 and ROCm 7.0)
 - C/C++ compiler (GCC, Clang, or MSVC)
 
 Tip: You can skip native installs by using our Docker environment (recommended):

modules/module1/README.md

@@ -43,13 +43,13 @@ This directory contains practical examples that accompany Module 1 of the GPU Pr
 
 ### For CUDA Examples
 - NVIDIA GPU with compute capability 5.0+
-- NVIDIA drivers 550+ recommended
-- CUDA Toolkit 12.0+ (Docker uses CUDA 12.9.1)
+- NVIDIA drivers 580+ recommended
+- CUDA Toolkit 13.0+ (Docker uses CUDA 13.0.1)
 - GCC/Clang compiler
 
 ### For HIP Examples
 - AMD GPU with ROCm support OR NVIDIA GPU
-- ROCm 6.0+ (for AMD) or CUDA 12.0+ (for NVIDIA backend)
+- ROCm 7.0+ (for AMD) or CUDA 13.0+ (for NVIDIA backend)
 - HIP compiler (hipcc)
 
 ## Quick Start

modules/module1/examples/README.md

@@ -19,7 +19,7 @@ After completing this module, you will be able to:
 
 ### Prerequisites
 - NVIDIA GPU with CUDA support OR AMD GPU with ROCm support
-- CUDA Toolkit 13.0+ or ROCm 7.0+ (Docker images provide CUDA 13.0.1 and ROCm 7.0.1)
+- CUDA Toolkit 13.0+ or ROCm 7.0+ (Docker images provide CUDA 13.0.1 and ROCm 7.0)
 - C/C++ compiler (GCC, Clang, or MSVC)
 
 Recommended: use our Docker dev environment

modules/module2/README.md

@@ -61,8 +61,8 @@ Comprehensive memory bandwidth optimization techniques:
 ## Building and Running Examples
 
 ### Prerequisites
-- CUDA Toolkit 12.0+ (for CUDA examples)
-- ROCm 6.0+ (for HIP examples)
+- CUDA Toolkit 13.0+ (for CUDA examples)
+- ROCm 7.0+ (for HIP examples)
 - Compatible GPU (NVIDIA or AMD)
 - C++17 compatible compiler
 
@@ -108,4 +108,4 @@ rocprof --stats ./build/02_memory_coalescing_hip
 
 ## Notes
 
-These examples are designed to be educational and performance-oriented. Use the provided Docker environment for consistent toolchains (CUDA 12.9.1, ROCm latest). Binaries are emitted to the `build/` directory by the Makefile.
+These examples are designed to be educational and performance-oriented. Use the provided Docker environment for consistent toolchains (CUDA 13.0.1, ROCm 7.0). Binaries are emitted to the `build/` directory by the Makefile.

modules/module2/examples/README.md

@@ -529,7 +529,7 @@ help_extended:
 	@echo ""
 	@echo "Requirements:"
 	@echo "  - CUDA Toolkit 10.0+ (for CUDA examples)"
-	@echo "  - HIP/ROCm 4.0+ (for HIP examples)" 
+	@echo "  - HIP/ROCm 7.0+ (for HIP examples)" 
 	@echo "  - Compute Capability 5.0+ (3.5+ for dynamic parallelism)"
 	@echo "  - Multi-GPU system recommended for full testing"
 	@echo "  - OpenMP support for parallel host code"

modules/module4/examples/Makefile

@@ -191,7 +191,7 @@ rocprof --version  # AMD ROCm Profiler
 ```
 
 **Minimum Requirements:**
-- CUDA Toolkit 12.0+ or HIP/ROCm 6.0+
+- CUDA Toolkit 13.0+ or HIP/ROCm 7.0+
 - Compute Capability 6.0+ (recommended for full feature support)
 - Profiling tools installed and properly configured
 - Sufficient GPU memory for performance testing (4GB+ recommended)

modules/module5/README.md

@@ -17,7 +17,7 @@ By completing this module, you will:
 ## Prerequisites
 
 **Recommended Requirements:**
-- CUDA Toolkit 12.0+ or ROCm 6.0+
+- CUDA Toolkit 13.0+ or ROCm 7.0+
 
 ### Core Content
 - **content.md** - Comprehensive guide covering all fundamental parallel algorithm patterns
@@ -169,7 +169,7 @@ make system_info
 ```
 
 **Minimum Requirements:**
-- CUDA Toolkit 11.0+ or ROCm 5.0+
+- CUDA Toolkit 13.0+ or ROCm 7.0+
 - Compute Capability 6.0+ recommended
 - 8GB+ GPU memory for large dataset examples
 - C++14 compatible compiler

modules/module6/README.md

@@ -170,7 +170,7 @@ rocm-smi --showproductname
 ```
 
 **Recommended Requirements:**
-- CUDA Toolkit 12.0+ or ROCm 6.0+
+- CUDA Toolkit 13.0+ or ROCm 7.0+
 - Compute Capability 7.0+ (Tensor Cores for applicable algorithms)
 - 16GB+ GPU memory for large-scale problems
 - Multi-GPU setup recommended for distributed algorithms