ollama - 💡(How to fix) Fix Windows 11 + Intel GPU + Vulkan segfaults [1 comments, 2 participants]

Exception 0xc0000005 0x8 0x7ffd5e46cb12 0x7ffd5e46cb12 PC=0x7ffd5e46cb12 signal arrived during external code execution

runtime.cgocall(0x7ff7260f2f30, 0xc000045da0) runtime/cgocall.go:167 +0x3e fp=0xc000045d78 sp=0xc000045d10 pc=0x7ff724fa243e github.com/ollama/ollama/x/imagegen/mlx._Cfunc_mlx_random_key(0xc0000722e0, 0x19da0e2f514) _cgo_gotypes.go:1978 +0x50 fp=0xc000045da0 sp=0xc000045d78 pc=0x7ff7256370d0 github.com/ollama/ollama/x/imagegen/mlx.RandomKey.func1(...) github.com/ollama/ollama/x/imagegen/mlx/mlx.go:1870 github.com/ollama/ollama/x/imagegen/mlx.RandomKey(0x19da0e2f514) github.com/ollama/ollama/x/imagegen/mlx/mlx.go:1870 +0x5d fp=0xc000045dd8 sp=0xc000045da0 pc=0x7ff72563fffd github.com/ollama/ollama/x/imagegen/mlx.init.0() github.com/ollama/ollama/x/imagegen/mlx/mlx.go:1848 +0xa9 fp=0xc000045e28 sp=0xc000045dd8 pc=0x7ff72563fe09 runtime.doInit1(0x7ff72749b2d0) runtime/proc.go:7350 +0xdd fp=0xc000045f50 sp=0xc000045e28 pc=0x7ff724f8343d runtime.doInit(...) runtime/proc.go:7317 runtime.main() runtime/proc.go:254 +0x325 fp=0xc000045fe0 sp=0xc000045f50 pc=0x7ff724f74e85 runtime.goexit({}) runtime/asm_amd64.s:1700 +0x1 fp=0xc000045fe8 sp=0xc000045fe0 pc=0x7ff724fadb21

goroutine 2 gp=0xc0000028c0 m=nil [force gc (idle)]: runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?) runtime/proc.go:435 +0xce fp=0xc00006bfa8 sp=0xc00006bf88 pc=0x7ff724fa598e runtime.goparkunlock(...) runtime/proc.go:441 runtime.forcegchelper() runtime/proc.go:348 +0xb8 fp=0xc00006bfe0 sp=0xc00006bfa8 pc=0x7ff724f750f8 runtime.goexit({}) runtime/asm_amd64.s:1700 +0x1 fp=0xc00006bfe8 sp=0xc00006bfe0 pc=0x7ff724fadb21 created by runtime.init.7 in goroutine 1 runtime/proc.go:336 +0x1a

goroutine 3 gp=0xc000002c40 m=nil [GC sweep wait]: runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?) runtime/proc.go:435 +0xce fp=0xc00006df80 sp=0xc00006df60 pc=0x7ff724fa598e runtime.goparkunlock(...) runtime/proc.go:441 runtime.bgsweep(0xc00007a000) runtime/mgcsweep.go:276 +0x94 fp=0xc00006dfc8 sp=0xc00006df80 pc=0x7ff724f5de74 runtime.gcenable.gowrap1() runtime/mgc.go:204 +0x25 fp=0xc00006dfe0 sp=0xc00006dfc8 pc=0x7ff724f52285 runtime.goexit({}) runtime/asm_amd64.s:1700 +0x1 fp=0xc00006dfe8 sp=0xc00006dfe0 pc=0x7ff724fadb21 created by runtime.gcenable in goroutine 1 runtime/mgc.go:204 +0x66

goroutine 4 gp=0xc000002e00 m=nil [GC scavenge wait]: runtime.gopark(0xc00007a000?, 0x7ff7269d6560?, 0x1?, 0x0?, 0xc000002e00?) runtime/proc.go:435 +0xce fp=0xc000085f78 sp=0xc000085f58 pc=0x7ff724fa598e runtime.goparkunlock(...) runtime/proc.go:441 runtime.(*scavengerState).park(0x7ff7275b6280) runtime/mgcscavenge.go:425 +0x49 fp=0xc000085fa8 sp=0xc000085f78 pc=0x7ff724f5b909 runtime.bgscavenge(0xc00007a000) runtime/mgcscavenge.go:653 +0x3c fp=0xc000085fc8 sp=0xc000085fa8 pc=0x7ff724f5be7c runtime.gcenable.gowrap2() runtime/mgc.go:205 +0x25 fp=0xc000085fe0 sp=0xc000085fc8 pc=0x7ff724f52225 runtime.goexit({}) runtime/asm_amd64.s:1700 +0x1 fp=0xc000085fe8 sp=0xc000085fe0 pc=0x7ff724fadb21 created by runtime.gcenable in goroutine 1 runtime/mgc.go:205 +0xa5

goroutine 5 gp=0xc000003340 m=nil [finalizer wait]: runtime.gopark(0x0?, 0x0?, 0x0?, 0x0?, 0x0?) runtime/proc.go:435 +0xce fp=0xc000087e30 sp=0xc000087e10 pc=0x7ff724fa598e runtime.runfinq() runtime/mfinal.go:196 +0x107 fp=0xc000087fe0 sp=0xc000087e30 pc=0x7ff724f51207 runtime.goexit({}) runtime/asm_amd64.s:1700 +0x1 fp=0xc000087fe8 sp=0xc000087fe0 pc=0x7ff724fadb21 created by runtime.createfing in goroutine 1 runtime/mfinal.go:166 +0x3d rax 0x64 rbx 0x7ffd5e53098c rcx 0x1e3b59087deb0000 rdx 0x2d1648c0000 rdi 0x2d1a4130860 rsi 0x0 rbp 0xfdb29cf309 rsp 0xfdb29ceda0 r8 0x7ffffffffffffffc r9 0xfdb2a71000 r10 0x8101010101010100 r11 0x2d164a974b0 r12 0xffffffffffffffff r13 0x5d r14 0x2d164a24a58 r15 0x0 rip 0x7ffd5e46cb12 rflags 0x10202 cs 0x33 fs 0x53 gs 0x2b

TL;DR

The issue with ollama not interacting when using Vulkan support on an Intel GPU may be related to a compatibility or configuration problem, and trying a different configuration or updating drivers could help.

Guidance

• Check the Intel GPU drivers for updates, as outdated drivers might cause compatibility issues with Vulkan support.
• Verify that the Vulkan SDK is properly installed and configured on the system.
• Consider testing ollama with a different GPU or without Vulkan support to isolate the issue.
• Look for any specific configuration options in ollama that might need to be adjusted for Intel GPUs or Vulkan support.

Example

No specific code example is applicable in this case, as the issue seems related to configuration or compatibility rather than code.

Notes

The provided information does not specify the exact nature of the interaction issue with ollama, which could be related to various factors including but not limited to driver support, Vulkan configuration, or ollama's compatibility with the Intel GPU on Windows.

Recommendation

Apply workaround: Try updating the Intel GPU drivers and verify the Vulkan SDK installation to see if it resolves the interaction issue with ollama. This approach is recommended because it directly addresses potential compatibility and configuration problems without requiring code changes or specific version updates that might not be applicable.