diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index aeb8dc56..120febf0 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -94,20 +94,20 @@ jobs:
         name:
           - btreemap
           - btreeset
+          # - io_chunks TODO: re-enable after new folder is merged.
           - memory-manager
           - vec
-          - compare
         include:
           - name: btreemap
             project_dir: ./benchmarks/btreemap
           - name: btreeset
             project_dir: ./benchmarks/btreeset
+          # - name: io_chunks
+          #   project_dir: ./benchmarks/io_chunks
           - name: memory-manager
             project_dir: ./benchmarks/memory_manager
           - name: vec
             project_dir: ./benchmarks/vec
-          - name: compare
-            project_dir: ./benchmarks/compare
 
     env:
       PROJECT_DIR: ${{ matrix.project_dir }}
diff --git a/benchmarks/Cargo.toml b/benchmarks/Cargo.toml
index 753b1972..cdc9f771 100644
--- a/benchmarks/Cargo.toml
+++ b/benchmarks/Cargo.toml
@@ -31,8 +31,8 @@ name = "vec"
 path = "vec/src/main.rs"
 
 [[bin]]
-name = "compare"
-path = "compare/src/main.rs"
+name = "io_chunks"
+path = "io_chunks/src/main.rs"
 
 [[bin]]
 name = "btreeset"
diff --git a/benchmarks/compare/canbench.yml b/benchmarks/io_chunks/canbench.yml
similarity index 100%
rename from benchmarks/compare/canbench.yml
rename to benchmarks/io_chunks/canbench.yml
diff --git a/benchmarks/compare/canbench_results.yml b/benchmarks/io_chunks/canbench_results.yml
similarity index 100%
rename from benchmarks/compare/canbench_results.yml
rename to benchmarks/io_chunks/canbench_results.yml
diff --git a/benchmarks/compare/src/main.rs b/benchmarks/io_chunks/src/main.rs
similarity index 87%
rename from benchmarks/compare/src/main.rs
rename to benchmarks/io_chunks/src/main.rs
index 04148e0f..cfe6ba73 100644
--- a/benchmarks/compare/src/main.rs
+++ b/benchmarks/io_chunks/src/main.rs
@@ -1,3 +1,21 @@
+//! # I/O Chunks Benchmark
+//!
+//! Measures read/write performance across different stable storage approaches
+//! when handling 100 MiB of data split into varying chunk sizes.
+//!
+//! Tests approaches:
+//! - **Stable Memory**: Direct memory operations (baseline)
+//! - **BTreeMap**: Key-value storage with chunk index as key
+//! - **StableVec**: Sequential storage pushing chunks
+//!
+//! Each approach is tested with:
+//! - 1 chunk (100 MiB)
+//! - 1K chunks (~100 KiB each)  
+//! - 1M chunks (~100 B each)
+//!
+//! **Goal**: Evaluate how chunk size affects performance and compare
+//! the overhead of stable structures versus raw memory operations.
+
 use benchmarks::vec::BoundedVecN;
 use canbench_rs::{bench, bench_fn, BenchResult};
 use ic_cdk::api::stable::WASM_PAGE_SIZE_IN_BYTES;