docs: fix various links

- CMake build recipes
- examples
- tooling
- jit

docs/guide/2-user-guide/1-installation/index.md

@@ -79,7 +79,7 @@ export PKG_CONFIG_PATH="${PKG_CONFIG_PATH}:${HOME}/libxsmm/lib"
 
 ### CMake Build Recipes
 
-For build recipes on different platforms, make sure to also read the [CMake Build Recipes](./2-cmake-build-recipes.html).
+For build recipes on different platforms, make sure to also read the [CMake Build Recipes](./1-cmake-build-recipes.html).
 
 ### Using Python in a virtual environment
 

docs/guide/2-user-guide/3-examples/index.md

@@ -4,10 +4,10 @@ title: Examples
 
 - [[dbcsr_example_1(program)]] : how to create a dbcsr matrix (fortran)
 - [[dbcsr_example_2(program)]] : how to set a dbcsr matrix (fortran)
-- dbcsr_example_3: how to multiply two dbcsr matrices (in fortran: [[dbcsr_example_3(program)]]) and in c++: [dbcsr_example_3](../../../../examples/dbcsr_example_3.cpp))
+- dbcsr_example_3: how to multiply two dbcsr matrices (in fortran: [[dbcsr_example_3(program)]]) and in c++: [dbcsr_example_3](https://github.com/cp2k/dbcsr/blob/develop/examples/dbcsr_example_3.cpp))
 - [[dbcsr_tensor_example_1(program)]] : how to create a dbcsr matrix (fortran)
     - the example can be run with different parameters, controlling block size, sparsity, verbosity and more
-- [dbcsr_tensor_example_2](../../../../examples/dbcsr_tensor_example_2.cpp): tensor contraction example (cpp)
+- [dbcsr_tensor_example_2](https://github.com/cp2k/dbcsr/blob/develop/examples/dbcsr_tensor_example_2.cpp): tensor contraction example (cpp)
     - tensor1 x tensor2 = tensor3, (13|2)x(54|21)=(3|45)
 
 ## Build

docs/guide/3-developer-guide/1-tooling/index.md

@@ -2,8 +2,8 @@ title: Tooling
 
 # Build System
 
-We support CMake for compilation. See [here](https://cp2k.github.io/dbcsr/page/2-user-guide/1-installation/1-install.html) on how to compile and 
-[here](https://cp2k.github.io/dbcsr/page/2-user-guide/1-installation/2-cmake-build-recipes.html) for more CMake details.
+We support CMake for compilation. See [here](../../2-user-guide/1-installation/index.html) on how to compile and 
+[here](../../2-user-guide/1-installation/1-cmake-build-recipes.html) for more CMake details.
 
 Compilations is based on [Fypp](https://github.com/aradi/fypp) meta-progamming package, which is available as submodule.
 

docs/guide/3-developer-guide/4-performance/2-just-in-time-compilation.md

@@ -10,5 +10,5 @@ Instead, kernels are JIT-ed on the fly, at runtime, as they are requested by the
 On NVIDIA's P100, the overhead of JIT has been found to be around 400ms for one kernel - a negligible overhead for typical DBCSR (and CP2K) runs.
 On AMD GPUs however, the overhead has been found to be of several seconds, a real hinderance to performance.
 
-For performance debugging and in order to check how much time a program spends doing JIT, look for the functions `jit_kernel_multiply` and `jit_kernel_transpose` in the [timings report](./1-insights.md) at the end of the output file.
+For performance debugging and in order to check how much time a program spends doing JIT, look for the functions `jit_kernel_multiply` and `jit_kernel_transpose` in the [timings report](./1-insights.html) at the end of the output file.