refactor: enable macos build workflow by 1a1a11a · Pull Request #217 · 1a1a11a/libCacheSim

1a1a11a · 2025-06-22T16:34:21Z

update macOS job configuration and upgrade checkout action to v4

trunk-io · 2025-06-22T16:34:25Z

Merging to develop in this repository is managed by Trunk.

To merge this pull request, check the box to the left or comment /trunk merge below.

…to v4

cursor

Bug: Missing CMake Configuration in macOS Job

The macOS job is missing the essential CMake configuration step. It attempts to build the project without first running cmake -B to initialize the build directory, which will cause the build to fail. This configuration step is correctly included in the ubuntu and selfhosted jobs.

.github/workflows/build.yml#L12-L27

libCacheSim/.github/workflows/build.yml

Lines 12 to 27 in 15f3241

    
           jobs: 
        
             macos: 
        
               name: macos / clang 
        
               runs-on: macos-latest 
        
               env: 
        
                 CC: clang 
        
               steps: 
        
                 - uses: actions/checkout@v4 
        
                 - name: Prepare 
        
                   run: bash scripts/install_dependency.sh 
        
                 - name: Build 
        
                   run: cmake --build ${{github.workspace}}/build --config ${{env.BUILD_TYPE}} 
        
                 - name: Test 
        
                   working-directory: ${{github.workspace}}/build 
        
                   run: | 
        
                     ctest -C ${{env.BUILD_TYPE}} --output-on-failure

Fix in Cursor

Was this report helpful? Give feedback by reacting with 👍 or 👎

Copilot

Pull Request Overview

This PR refactors the macOS build workflow and updates the checkout action to version 4 across all jobs.

Uncomments and updates the macOS job with a new dependency installation script and switches from macos-10.15 to macos-latest.
Upgrades the actions/checkout step from v2 to v4 in macOS, Ubuntu, and self-hosted jobs.

Comments suppressed due to low confidence (3)

.github/workflows/build.yml:17

Upgrading to actions/checkout@v4 is a good update; please verify that any changes in default behaviors do not affect downstream steps.

      - uses: actions/checkout@v4

.github/workflows/build.yml:32

Upgrading to actions/checkout@v4 is a good update; please verify that any changes in default behaviors do not affect downstream steps.

      - uses: actions/checkout@v4

.github/workflows/build.yml:52

Upgrading to actions/checkout@v4 is a good update; please verify that any changes in default behaviors do not affect downstream steps.

      - uses: actions/checkout@v4

Copilot · 2025-06-22T20:49:52Z

-  #       run: make test
+  macos:
+    name: macos / clang
+    runs-on: macos-latest


Using 'macos-latest' may lead to varying build environments over time; consider pinning to a specific macOS version to ensure consistent builds if required.

Suggested change

runs-on: macos-latest

runs-on: macos-12

1a1a11a · 2025-06-22T21:00:27Z

Not sure why the test results are not correct.

laustam · 2025-06-23T13:12:01Z

Might the test be failing due to a different RNG implementation on macOS?

1a1a11a · 2025-06-23T15:03:43Z

Might the test be failing due to a different RNG implementation on macOS?

I thought about this, but LRU does not use RNG, or is there a hidden one that I missed?

laustam · 2025-06-23T15:20:21Z

ERROR:/Users/runner/work/libCacheSim/libCacheSim/test/test_evictionAlgo.c:17:_verify_profiler_results: assertion failed (miss_cnt_true[i] == res[i].n_miss): (89776 == 89582)
Cacheus uint64_t cache_size[] = {134217728, 268435456, 402653184, 536870[91](https://github.com/1a1a11a/libCacheSim/actions/runs/15810660836/job/44561516073?pr=217#step:6:92)2, 671088640, 805306368, 939524096, 1073741824};
uint64_t miss_cnt_true[] = {89582, 82491, 80140, 72970, 69329, 69147, 68022, 66158};
uint64_t miss_byte_true[] = {4046596096, 3724201472, 3557762560, 3182379008, 2981198336, 2964742144, 28678[92](https://github.com/1a1a11a/libCacheSim/actions/runs/15810660836/job/44561516073?pr=217#step:6:93)736, 2791617536};
not ok /libCacheSim/cacheAlgo_Cacheus - ERROR:/Users/runner/work/libCacheSim/libCacheSim/test/test_evictionAlgo.c:17:_verify_profiler_results: assertion failed (miss_cnt_true[i] == res[i].n_miss): (89776 == 89582)
Bail out!

It seems that the LRU test passes, and that Cacheus is the problematic test (we can see this from the print_results debug prints). I briefly checked the Cacheus implementation, and there is some use of RNG there

haochengxia · 2025-06-23T15:21:46Z

ERROR:/Users/runner/work/libCacheSim/libCacheSim/test/test_evictionAlgo.c:17:_verify_profiler_results: assertion failed (miss_cnt_true[i] == res[i].n_miss): (89776 == 89582)
Cacheus uint64_t cache_size[] = {134217728, 268435456, 402653184, 536870[91](https://github.com/1a1a11a/libCacheSim/actions/runs/15810660836/job/44561516073?pr=217#step:6:92)2, 671088640, 805306368, 939524096, 1073741824};
uint64_t miss_cnt_true[] = {89582, 82491, 80140, 72970, 69329, 69147, 68022, 66158};
uint64_t miss_byte_true[] = {4046596096, 3724201472, 3557762560, 3182379008, 2981198336, 2964742144, 28678[92](https://github.com/1a1a11a/libCacheSim/actions/runs/15810660836/job/44561516073?pr=217#step:6:93)736, 2791617536};
not ok /libCacheSim/cacheAlgo_Cacheus - ERROR:/Users/runner/work/libCacheSim/libCacheSim/test/test_evictionAlgo.c:17:_verify_profiler_results: assertion failed (miss_cnt_true[i] == res[i].n_miss): (89776 == 89582)
Bail out!

It seems that the LRU test passes, and that Cacheus is the problematic test (we can see this from the print_results debug prints). I briefly checked the Cacheus implementation, and there is some use of RNG there

While LeCaR uses RNG as well, it passed the test.

haochengxia · 2025-06-23T16:38:12Z

Confirmed the difference is from RNG which makes lr different

# Mac
[Cacheus][macOS] Init: ghost_list_factor=1.00, update_interval=1073741824, lr=0.374000, w_lru=0.50, w_lfu=0.50, cache_size=1073741824, hashpower=20
[Cacheus][macOS] LRU=0x105808a00, LFU=0x130018600, LRU_g=0x130019400, LFU_g=0x13001a200
[Cacheus][macOS] sizeof(Cacheus_params_t)=112

# Ubuntu
[Cacheus][Linux] Init: ghost_list_factor=1.00, update_interval=1073741824, lr=0.276000, w_lru=0.50, w_lfu=0.50, cache_size=1073741824, hashpower=20
[Cacheus][Linux] LRU=0x55f84403ad00, LFU=0x55f84403ee00, LRU_g=0x55f84403e100, LFU_g=0x55f845bb6000
[Cacheus][Linux] sizeof(Cacheus_params_t)=112

dump rand() in test_evictionAlgo.c

int main(int argc, char *argv[]) {
  g_test_init(&argc, &argv, NULL);
  srand(0);  // for reproducibility
  printf("rand() = %d\n", rand());

### Mac
# random seed: R02Scbb0a3dd5ce5dd47fd2d2ff21f064356
rand() = 520932930

### Ubuntu
# random seed: R02Sf907d52631b1852990966e465debb2f8
rand() = 1804289383

1a1a11a · 2025-06-23T19:09:45Z

Confirmed the difference is from RNG which makes lr different

# Mac
[Cacheus][macOS] Init: ghost_list_factor=1.00, update_interval=1073741824, lr=0.374000, w_lru=0.50, w_lfu=0.50, cache_size=1073741824, hashpower=20
[Cacheus][macOS] LRU=0x105808a00, LFU=0x130018600, LRU_g=0x130019400, LFU_g=0x13001a200
[Cacheus][macOS] sizeof(Cacheus_params_t)=112

# Ubuntu
[Cacheus][Linux] Init: ghost_list_factor=1.00, update_interval=1073741824, lr=0.276000, w_lru=0.50, w_lfu=0.50, cache_size=1073741824, hashpower=20
[Cacheus][Linux] LRU=0x55f84403ad00, LFU=0x55f84403ee00, LRU_g=0x55f84403e100, LFU_g=0x55f845bb6000
[Cacheus][Linux] sizeof(Cacheus_params_t)=112

dump rand() in test_evictionAlgo.c

int main(int argc, char *argv[]) {
  g_test_init(&argc, &argv, NULL);
  srand(0);  // for reproducibility
  printf("rand() = %d\n", rand());

### Mac
# random seed: R02Scbb0a3dd5ce5dd47fd2d2ff21f064356
rand() = 520932930

### Ubuntu
# random seed: R02Sf907d52631b1852990966e465debb2f8
rand() = 1804289383

ERROR:/Users/runner/work/libCacheSim/libCacheSim/test/test_evictionAlgo.c:17:_verify_profiler_results: assertion failed (miss_cnt_true[i] == res[i].n_miss): (89776 == 89582)
Cacheus uint64_t cache_size[] = {134217728, 268435456, 402653184, 536870[91](https://github.com/1a1a11a/libCacheSim/actions/runs/15810660836/job/44561516073?pr=217#step:6:92)2, 671088640, 805306368, 939524096, 1073741824};
uint64_t miss_cnt_true[] = {89582, 82491, 80140, 72970, 69329, 69147, 68022, 66158};
uint64_t miss_byte_true[] = {4046596096, 3724201472, 3557762560, 3182379008, 2981198336, 2964742144, 28678[92](https://github.com/1a1a11a/libCacheSim/actions/runs/15810660836/job/44561516073?pr=217#step:6:93)736, 2791617536};
not ok /libCacheSim/cacheAlgo_Cacheus - ERROR:/Users/runner/work/libCacheSim/libCacheSim/test/test_evictionAlgo.c:17:_verify_profiler_results: assertion failed (miss_cnt_true[i] == res[i].n_miss): (89776 == 89582)
Bail out!

It seems that the LRU test passes, and that Cacheus is the problematic test (we can see this from the print_results debug prints). I briefly checked the Cacheus implementation, and there is some use of RNG there

Oh, I read it wrong. Yes, it looks like RNG is the reason. We have an internal fast and deterministic RNG, do you want to submit a PR to fix it?

1a1a11a · 2025-06-23T19:11:35Z

Confirmed the difference is from RNG which makes lr different

# Mac
[Cacheus][macOS] Init: ghost_list_factor=1.00, update_interval=1073741824, lr=0.374000, w_lru=0.50, w_lfu=0.50, cache_size=1073741824, hashpower=20
[Cacheus][macOS] LRU=0x105808a00, LFU=0x130018600, LRU_g=0x130019400, LFU_g=0x13001a200
[Cacheus][macOS] sizeof(Cacheus_params_t)=112

# Ubuntu
[Cacheus][Linux] Init: ghost_list_factor=1.00, update_interval=1073741824, lr=0.276000, w_lru=0.50, w_lfu=0.50, cache_size=1073741824, hashpower=20
[Cacheus][Linux] LRU=0x55f84403ad00, LFU=0x55f84403ee00, LRU_g=0x55f84403e100, LFU_g=0x55f845bb6000
[Cacheus][Linux] sizeof(Cacheus_params_t)=112

dump rand() in test_evictionAlgo.c

int main(int argc, char *argv[]) {
  g_test_init(&argc, &argv, NULL);
  srand(0);  // for reproducibility
  printf("rand() = %d\n", rand());

### Mac
# random seed: R02Scbb0a3dd5ce5dd47fd2d2ff21f064356
rand() = 520932930

### Ubuntu
# random seed: R02Sf907d52631b1852990966e465debb2f8
rand() = 1804289383

Thank you for confirming this!

1a1a11a · 2025-06-23T19:13:06Z

Confirmed the difference is from RNG which makes lr different

# Mac
[Cacheus][macOS] Init: ghost_list_factor=1.00, update_interval=1073741824, lr=0.374000, w_lru=0.50, w_lfu=0.50, cache_size=1073741824, hashpower=20
[Cacheus][macOS] LRU=0x105808a00, LFU=0x130018600, LRU_g=0x130019400, LFU_g=0x13001a200
[Cacheus][macOS] sizeof(Cacheus_params_t)=112

# Ubuntu
[Cacheus][Linux] Init: ghost_list_factor=1.00, update_interval=1073741824, lr=0.276000, w_lru=0.50, w_lfu=0.50, cache_size=1073741824, hashpower=20
[Cacheus][Linux] LRU=0x55f84403ad00, LFU=0x55f84403ee00, LRU_g=0x55f84403e100, LFU_g=0x55f845bb6000
[Cacheus][Linux] sizeof(Cacheus_params_t)=112

dump rand() in test_evictionAlgo.c

int main(int argc, char *argv[]) {
  g_test_init(&argc, &argv, NULL);
  srand(0);  // for reproducibility
  printf("rand() = %d\n", rand());

### Mac
# random seed: R02Scbb0a3dd5ce5dd47fd2d2ff21f064356
rand() = 520932930

### Ubuntu
# random seed: R02Sf907d52631b1852990966e465debb2f8
rand() = 1804289383

ERROR:/Users/runner/work/libCacheSim/libCacheSim/test/test_evictionAlgo.c:17:_verify_profiler_results: assertion failed (miss_cnt_true[i] == res[i].n_miss): (89776 == 89582)
Cacheus uint64_t cache_size[] = {134217728, 268435456, 402653184, 536870[91](https://github.com/1a1a11a/libCacheSim/actions/runs/15810660836/job/44561516073?pr=217#step:6:92)2, 671088640, 805306368, 939524096, 1073741824};
uint64_t miss_cnt_true[] = {89582, 82491, 80140, 72970, 69329, 69147, 68022, 66158};
uint64_t miss_byte_true[] = {4046596096, 3724201472, 3557762560, 3182379008, 2981198336, 2964742144, 28678[92](https://github.com/1a1a11a/libCacheSim/actions/runs/15810660836/job/44561516073?pr=217#step:6:93)736, 2791617536};
not ok /libCacheSim/cacheAlgo_Cacheus - ERROR:/Users/runner/work/libCacheSim/libCacheSim/test/test_evictionAlgo.c:17:_verify_profiler_results: assertion failed (miss_cnt_true[i] == res[i].n_miss): (89776 == 89582)
Bail out!

It seems that the LRU test passes, and that Cacheus is the problematic test (we can see this from the print_results debug prints). I briefly checked the Cacheus implementation, and there is some use of RNG there

Oh, I read it wrong. Yes, it looks like RNG is the reason. We have an internal fast and deterministic RNG, do you want to submit a PR to fix it?

At first glance, CACHEUS uses the internal RNG, so probably need to dig deeper.

haochengxia · 2025-06-23T19:54:31Z

Pin the seed for internal RNG
Change print_result since the output is simulated results instead of fixed reference results

1a1a11a · 2025-06-23T19:56:21Z

Pin the seed for internal RNG

Change print_result since the output is simulated results instead of fixed reference results

Thank you! Does it only affect CACHEUS?

haochengxia · 2025-06-23T19:58:50Z

Pin the seed for internal RNG

Change print_result since the output is simulated results instead of fixed reference results

Thank you! Does it only affect CACHEUS?

Yes. Only CacheUs is affected by the initial state of RNG

1a1a11a requested a review from haochengxia as a code owner June 22, 2025 16:34

1a1a11a changed the title ~~refactor: update macOS job configuration and upgrade checkout action …~~ refactor: enable macos build workflow Jun 22, 2025

refactor: update macOS job configuration and upgrade checkout action …

fb464b0

…to v4

1a1a11a force-pushed the 1a1a11a/macos_build branch from 6d9a903 to fb464b0 Compare June 22, 2025 16:37

1a1a11a mentioned this pull request Jun 22, 2025

Fix macOS libcachesim compilation #183

Closed

Merge branch 'develop' into 1a1a11a/macos_build

e697276

This comment was marked as outdated.

Sign in to view

fix workflow

15f3241

1a1a11a marked this pull request as draft June 22, 2025 20:44

1a1a11a requested a review from Copilot June 22, 2025 20:44

This comment was marked as outdated.

Sign in to view

cursor Bot reviewed Jun 22, 2025

View reviewed changes

1a1a11a added 2 commits June 22, 2025 20:45

change compiler to default

33f1371

fix

258454a

1a1a11a requested a review from Copilot June 22, 2025 20:49

1a1a11a self-assigned this Jun 22, 2025

Copilot AI reviewed Jun 22, 2025

View reviewed changes

Pin seed for cross-plat test

eca1dd5

Merge branch 'develop' into 1a1a11a/macos_build

88f6310

1a1a11a marked this pull request as ready for review June 23, 2025 19:56

1a1a11a merged commit 17fc4ec into develop Jun 23, 2025
9 checks passed

haochengxia mentioned this pull request Jun 24, 2025

[BUG]: macOS eviction tests fail #224

Closed

1a1a11a deleted the 1a1a11a/macos_build branch June 25, 2025 03:36

	jobs:
	macos:
	name: macos / clang
	runs-on: macos-latest
	env:
	CC: clang
	steps:
	- uses: actions/checkout@v4
	- name: Prepare
	run: bash scripts/install_dependency.sh
	- name: Build
	run: cmake --build ${{github.workspace}}/build --config ${{env.BUILD_TYPE}}
	- name: Test
	working-directory: ${{github.workspace}}/build
	run: \|
	ctest -C ${{env.BUILD_TYPE}} --output-on-failure

Conversation

1a1a11a commented Jun 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

trunk-io Bot commented Jun 22, 2025

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

cursor Bot left a comment

Choose a reason for hiding this comment

Bug: Missing CMake Configuration in macOS Job

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Uh oh!

Copilot AI Jun 22, 2025

Choose a reason for hiding this comment

Uh oh!

1a1a11a commented Jun 22, 2025

Uh oh!

laustam commented Jun 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

1a1a11a commented Jun 23, 2025

Uh oh!

laustam commented Jun 23, 2025

Uh oh!

haochengxia commented Jun 23, 2025

Uh oh!

haochengxia commented Jun 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

1a1a11a commented Jun 23, 2025

Uh oh!

1a1a11a commented Jun 23, 2025

Uh oh!

1a1a11a commented Jun 23, 2025

Uh oh!

haochengxia commented Jun 23, 2025

Uh oh!

1a1a11a commented Jun 23, 2025

Uh oh!

Uh oh!

haochengxia commented Jun 23, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

1a1a11a commented Jun 22, 2025 •

edited

Loading

laustam commented Jun 23, 2025 •

edited

Loading

haochengxia commented Jun 23, 2025 •

edited

Loading