Repository navigation
test(cuda): add the #3320 retry-study per-run logs #3336
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: master
Are you sure you want to change the base?
Changes from all commits
1c49829
0ba2acf
1714d44
d3da530
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,47 @@ | ||
| # oom-r20 — #3320 landed test: 12-round rerun + probe A/B + reclamation latency | ||
|
|
||
| Date: 2026-11-07 (07:15-08:25 UTC). Build: master `b38635f63`, #3320 test as landed | ||
| (unchanged except the stated arm edits), MinGW64 gcc 16.2.0 / nvcc 13.4, | ||
| Windows 11, driver 621.01. Devices: CUDA_VISIBLE_DEVICES 0 = RTX 5080, | ||
| 1 = RTX 4060 Laptop (nvidia-smi -i order is the opposite: 0=4060, 1=5080). | ||
| Raw logs: `armA/ armB/ armC/ armD/ ilv/` (per-run .err with rc appended, | ||
| timeline.log with nvidia-smi brackets, armD/sampler.log 5 s cadence). | ||
|
|
||
| ## Arms | ||
|
|
||
| | arm | binary | post-release probe | rounds | | ||
| | --- | --- | --- | --- | | ||
| | A | as-landed | one-shot cuMemAlloc/cuMemFree 2 MiB | 12 x 2 devices | | ||
| | B | no-probe | none (only change) | 12 x 2 | | ||
| | C | bounded poll | 20 ms poll, 60 s bound, timed | 12 x 2 | | ||
| | D | silent ladder | single probes after silent 1/5/15/45/105 s | 6 x 2 | | ||
| | ilv | interleaved A/B | as-landed vs no-probe, pairs back-to-back | 12 pairs 4060, 6 pairs 5080 | | ||
|
|
||
| ## Results | ||
|
|
||
| - **A**: 5080 12/12 and 4060 12/12 fail at "device memory not yet returned after | ||
| the hog was freed" (`memory back 0`); retry assertion never fired. Slow window. | ||
| - **ilv**: 5080 6/6 pairs both arms fail (as-landed @not-yet-returned, no-probe | ||
| @retry); 4060 12 pairs: as-landed 11/12 pass + 1 transient refusal, no-probe | ||
| 12/12 pass. **No systematic probe perturbation.** | ||
| - **C**: 4060 12/12 `memory back 1 after 0.0 ms (0 probes)`; 5080 6/6 bound hit | ||
| (`after 60000 ms (1921-1933 probes)`) then 6/6 instant after mid-campaign | ||
| window flip. | ||
| - **D**: 4060 6/6 probe OK at the 1 s checkpoint, retry rc 0; 5080 6/6 refused at | ||
| every checkpoint (1/5/15/45/105 s), retry rc -12. | ||
| - **armD/sampler.log** (5 s): during 5080 slow-window waits, used ~15.8 GiB / | ||
| free ~150 MiB stays flat -> the driver holds the freed hog until process exit. | ||
| Fresh processes re-take the same 30.8 GiB hog immediately (every run, all arms). | ||
|
|
||
| ## Conclusion | ||
|
|
||
| Windows slow window = the driver holds the released hog's memory for the | ||
| lifetime of the freeing process (>= 105 s observed, returned at exit); fast | ||
| window = instant return. The in-process one-shot probe is not a perturbation | ||
| source; a bounded wait + skip-with-reason is the only design that maps to the | ||
|
Comment on lines
+40
to
+41
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Qualify the probe conclusion. The 5080 as-landed runs stop before the retry, so those pairs do not compare retry outcomes with and without the probe. The 4060 results also include one as-landed refusal. Report that this sample found no systematic effect, rather than ruling out probe perturbation.
|
||
| observed behavior. Hog sizes 30.8/22.9 GiB (WDDM oversubscription) vs 710 MiB | ||
| on the maintainer's Linux 4090. | ||
|
|
||
| Arm edits used (local-only, reverted after): armB/C/D variants of | ||
| `check_oom_and_memory_return()` in core/test/test_cuda_oom_pictures_released.c; | ||
|
Comment on lines
+45
to
+46
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Preserve the Arm C and D edits. These edits were reverted, and
|
||
| binaries saved as build/test/{aslanded,noprobe}.exe. | ||
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30720 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,5 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/cuda/picture_cuda.c:356: CUDA_ERROR_OUT_OF_MEMORY (2) in cuMemAllocPitch((CUdeviceptr *)&pic->data[i], (size_t *)&pic->stride[i], (size_t)pic->w[i] * ((pic->bpc + 7) / 8), pic->h[i], 8 << hbd) | ||
| libvmaf ERROR problem during prepare_ring_buffer | ||
| [hog 22954 MiB, luma probe under hog 0, first rc -12, memory back 0] [first frame with no memory left: -12] [31mfail[0m | ||
| [31mdevice memory not yet returned after the hog was freed | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,97 @@ | ||
| === armA dev0 run01 start 2026-11-07T07:25:31Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run01 end 2026-11-07T07:25:41Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run02 start 2026-11-07T07:25:41Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run02 end 2026-11-07T07:25:48Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run03 start 2026-11-07T07:25:48Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run03 end 2026-11-07T07:25:52Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run04 start 2026-11-07T07:25:52Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run04 end 2026-11-07T07:25:57Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run05 start 2026-11-07T07:25:57Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run05 end 2026-11-07T07:26:02Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run06 start 2026-11-07T07:26:02Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run06 end 2026-11-07T07:26:07Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run07 start 2026-11-07T07:26:07Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run07 end 2026-11-07T07:26:13Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run08 start 2026-11-07T07:26:13Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run08 end 2026-11-07T07:26:18Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run09 start 2026-11-07T07:26:18Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run09 end 2026-11-07T07:26:23Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run10 start 2026-11-07T07:26:23Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run10 end 2026-11-07T07:26:27Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run11 start 2026-11-07T07:26:27Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run11 end 2026-11-07T07:26:32Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run12 start 2026-11-07T07:26:32Z | ||
| 0 MiB, 0 % | ||
| === armA dev0 run12 end 2026-11-07T07:26:37Z | ||
| 0 MiB, 0 % | ||
| === armA dev1 run01 start 2026-11-07T07:26:37Z | ||
| 567 MiB, 24 % | ||
| === armA dev1 run01 end 2026-11-07T07:26:41Z | ||
| 644 MiB, 90 % | ||
| === armA dev1 run02 start 2026-11-07T07:26:41Z | ||
| 644 MiB, 63 % | ||
| === armA dev1 run02 end 2026-11-07T07:26:44Z | ||
| 802 MiB, 35 % | ||
| === armA dev1 run03 start 2026-11-07T07:26:44Z | ||
| 802 MiB, 35 % | ||
| === armA dev1 run03 end 2026-11-07T07:26:48Z | ||
| 789 MiB, 98 % | ||
| === armA dev1 run04 start 2026-11-07T07:26:48Z | ||
| 789 MiB, 98 % | ||
| === armA dev1 run04 end 2026-11-07T07:26:51Z | ||
| 789 MiB, 20 % | ||
| === armA dev1 run05 start 2026-11-07T07:26:51Z | ||
| 789 MiB, 20 % | ||
| === armA dev1 run05 end 2026-11-07T07:26:55Z | ||
| 789 MiB, 21 % | ||
| === armA dev1 run06 start 2026-11-07T07:26:55Z | ||
| 789 MiB, 21 % | ||
| === armA dev1 run06 end 2026-11-07T07:26:58Z | ||
| 789 MiB, 67 % | ||
| === armA dev1 run07 start 2026-11-07T07:26:58Z | ||
| 789 MiB, 67 % | ||
| === armA dev1 run07 end 2026-11-07T07:27:01Z | ||
| 789 MiB, 19 % | ||
| === armA dev1 run08 start 2026-11-07T07:27:01Z | ||
| 789 MiB, 19 % | ||
| === armA dev1 run08 end 2026-11-07T07:27:04Z | ||
| 789 MiB, 20 % | ||
| === armA dev1 run09 start 2026-11-07T07:27:04Z | ||
| 789 MiB, 20 % | ||
| === armA dev1 run09 end 2026-11-07T07:27:07Z | ||
| 789 MiB, 20 % | ||
| === armA dev1 run10 start 2026-11-07T07:27:08Z | ||
| 789 MiB, 91 % | ||
| === armA dev1 run10 end 2026-11-07T07:27:11Z | ||
| 789 MiB, 20 % | ||
| === armA dev1 run11 start 2026-11-07T07:27:11Z | ||
| 789 MiB, 20 % | ||
| === armA dev1 run11 end 2026-11-07T07:27:14Z | ||
| 789 MiB, 21 % | ||
| === armA dev1 run12 start 2026-11-07T07:27:14Z | ||
| 789 MiB, 21 % | ||
| === armA dev1 run12 end 2026-11-07T07:27:17Z | ||
| 794 MiB, 21 % | ||
| ARM-A-DONE 2026-11-07T07:27:17Z |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30792 MiB, luma probe under hog 0, first rc -12, memory back n/a] [first frame with no memory left: -12] CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [retry rc -12] [31mfail[0m | ||
| [31mthe same index does not score once memory is back | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back n/a] [first frame with no memory left: -12] CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [retry rc -12] [31mfail[0m | ||
| [31mthe same index does not score once memory is back | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30720 MiB, luma probe under hog 0, first rc -12, memory back n/a] [first frame with no memory left: -12] CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [retry rc -12] [31mfail[0m | ||
| [31mthe same index does not score once memory is back | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back n/a] [first frame with no memory left: -12] CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [retry rc -12] [31mfail[0m | ||
| [31mthe same index does not score once memory is back | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back n/a] [first frame with no memory left: -12] CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [retry rc -12] [31mfail[0m | ||
| [31mthe same index does not score once memory is back | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| test_out_of_memory_returns_an_error_and_keeps_the_pool: CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [hog 30836 MiB, luma probe under hog 0, first rc -12, memory back n/a] [first frame with no memory left: -12] CUDA error at ../core/src/libvmaf.c:650: CUDA_ERROR_OUT_OF_MEMORY (2) in cuEventSynchronize(priv->cuda.ready) | ||
| [retry rc -12] [31mfail[0m | ||
| [31mthe same index does not score once memory is back | ||
| 1 tests run, 1 failed[0m | ||
| rc=1 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Separate the observation from the retention hypothesis.
The sampler reports device-wide usage, not ownership of the hog. The 2 MiB probe also cannot establish when the whole hog becomes available. In addition,
hog_release()discardscuMemFreeresults. The data therefore do not establish that successfully freed allocations remain held until process exit. Record the release results and allocation ownership, or state this mechanism as a hypothesis. (raw.githubusercontent.com)