manunicholasjacob dbce318e4f fix(xsysinfo): cap the default thread count by the container CPU limit ...

When --threads / LOCALAI_THREADS is unset, LocalAI uses the host's
physical core count. Inside a container with a CPU limit (docker --cpus,
a Kubernetes limits.cpu) that starts one thread per host core inside a
smaller CFS quota, so the backend is throttled for most of each period.

Since Go 1.25 the default GOMAXPROCS already follows the affinity mask and
the cgroup CPU limit. Use the smaller of the physical core count and
GOMAXPROCS as the default. Hosts without a limit are unchanged.

Assisted-by: Claude:claude-opus-5-5
Signed-off-by: manunicholasjacob <manunicholasjacob@gmail.com>
(cherry picked from commit aca800d003)

2026-10-25 08:43:05 +00:00

..

amd_rocm_meminfo_test.go

fix(xsysinfo): include AMD GTT in APU VRAM detection (#12094)

2026-09-18 21:48:44 +00:00

clinfo_test.go

fix(intel): VRAM detection (#9944)

2026-05-25 09:29:00 +02:00

clinfo.go

refactor: use slices.Contains to simplify code (#10702)

2026-07-06 19:33:28 +02:00

computecap_internal_test.go

feat(config): hardware-tuned defaults — Blackwell batch + VRAM-scaled concurrency (#10411)

2026-06-20 14:45:59 +02:00

cpu_test.go

fix(xsysinfo): cap the default thread count by the container CPU limit

2026-10-25 08:43:05 +00:00

cpu.go

fix(xsysinfo): cap the default thread count by the container CPU limit

2026-10-25 08:43:05 +00:00

disk_test.go

fix(nodes): never schedule a model onto a node that cannot store it (#11054)

2026-07-23 00:03:21 +02:00

disk.go

fix(nodes): never schedule a model onto a node that cannot store it (#11054)

2026-07-23 00:03:21 +02:00

drmfdinfo_test.go

fix(intel): VRAM detection (#9944)

2026-05-25 09:29:00 +02:00

drmfdinfo.go

fix(intel): VRAM detection (#9944)

2026-05-25 09:29:00 +02:00

gpu.go

fix(xsysinfo): report system RAM in /api/resources when a GPU is present

2026-10-25 08:43:05 +00:00

gpuvendor_internal_test.go

fix(gpu): detect GPUs via sysfs when no pci.ids database is present (#10966)

2026-07-20 00:37:06 +02:00

hasgpu_internal_test.go

fix(gpu): detect GPUs via sysfs when no pci.ids database is present (#10966)

2026-07-20 00:37:06 +02:00

intel_gputop_internal_test.go

fix(xsysinfo): avoid startup hang on intel_gpu_top (#12206)

2026-09-23 12:18:57 +02:00

memory_total_test.go

fix(xsysinfo): container-aware total RAM detection (cgroup/lxcfs) (#8059) (#10288)

2026-06-13 18:13:06 +02:00

memory_total.go

refactor: replace Split in loops with more efficient SplitSeq (#10879)

2026-07-17 22:07:16 +02:00

memory.go

fix(xsysinfo): container-aware total RAM detection (cgroup/lxcfs) (#8059) (#10288)

2026-06-13 18:13:06 +02:00

minvram_internal_test.go

fix(config): per-device VRAM headroom for Blackwell defaults (#10485) (#10494)

2026-06-25 00:07:48 +02:00

process_vram_linux_test.go

fix(xsysinfo): read process VRAM without proc children (#12482)

2026-10-05 01:37:12 +02:00

process_vram_linux.go

fix(xsysinfo): read process VRAM without proc children (#12482)

2026-10-05 01:37:12 +02:00

process_vram_other.go

feat(system): report per-model DRM VRAM (#12026)

2026-09-27 21:18:34 +02:00

sysfsgpu_internal_test.go

fix(gpu): detect GPUs via sysfs when no pci.ids database is present (#10966)

2026-07-20 00:37:06 +02:00

sysfsgpu.go

fix(gpu): detect GPUs via sysfs when no pci.ids database is present (#10966)

2026-07-20 00:37:06 +02:00

vrambudget_boundary_test.go

feat(vram): per-node VRAM allocation budget (LOCALAI_VRAM_BUDGET) (#10833)

2026-07-15 09:58:45 +02:00

xsysinfo_suite_test.go

fix(intel): VRAM detection (#9944)

2026-05-25 09:29:00 +02:00