Assembler Error

[Don-Swanson]

While trying to build on an aarch64 VPS, it errors out:

 => ERROR [serge-api builder 4/4] RUN cd llama.cpp &&     make &&     mv main llama                                                                                                            5.4s
 => [serge-web 6/6] COPY . .                                                                                                                                                                   0.1s
------
 > [serge-api builder 4/4] RUN cd llama.cpp &&     make &&     mv main llama:
#0 0.422 I llama.cpp build info: 
#0 0.423 I UNAME_S:  Linux
#0 0.423 I UNAME_P:  unknown
#0 0.423 I UNAME_M:  aarch64
#0 0.423 I CFLAGS:   -I.              -O3 -DNDEBUG -std=c11   -fPIC -pthread -mcpu=native
#0 0.423 I CXXFLAGS: -I. -I./examples -O3 -DNDEBUG -std=c++11 -fPIC -pthread -mcpu=native
#0 0.423 I LDFLAGS:  
#0 0.423 I CC:       cc (GCC) 10.2.0
#0 0.423 I CXX:      g++ (GCC) 10.2.0
#0 0.423 
#0 0.424 cc  -I.              -O3 -DNDEBUG -std=c11   -fPIC -pthread -mcpu=native   -c ggml.c -o ggml.o
#0 4.559 Assembler messages:
#0 4.559 Error: unknown architectural extension `sb+ssbs'
#0 4.559 Error: unrecognized option -march=armv8.2-a+crypto+fp16+rcpc+dotprod+sb+ssbs
#0 5.212 make: *** [Makefile:221: ggml.o] Error 1
------
failed to solve: process "/bin/sh -c cd llama.cpp &&     make &&     mv main llama" did not complete successfully: exit code: 2```

[hamx01]

I have the same thing on my ARM64 VPS

[gaby]

@TheEagle13 @hamx01 is this still happening with the latest release? The official image is built for arm64, can you try that?

[hamx01]

@gaby i've done docker run -d -v weights:/usr/src/app/weights -v datadb:/data/db/ -p 8008:8008 ghcr.io/nsarrazin/serge:latest on my ARM64 server, and now it is working. By the way, great work with this download manager for models.

[gaby]

@hamx01 Awesome! 🙌