mirror of
https://github.com/containers/ramalama.git
synced 2026-02-05 15:47:26 +01:00
334c641042
Testing Farm provides instances with K80 and V100 gpus, but recent cuda versions no longer support K80. Only run the no-rpms tests manually, triggered by a "/packit test" comment on the PR. This avoids frequent failures/timeouts when gpus are not available in Testing Farm. Signed-off-by: Mike Bonnet <mikeb@redhat.com>
37 lines
603 B
Plaintext
37 lines
603 B
Plaintext
discover:
|
|
how: fmf
|
|
|
|
execute:
|
|
how: tmt
|
|
|
|
prepare:
|
|
how: feature
|
|
epel: enabled
|
|
|
|
provision:
|
|
how: artemis
|
|
hardware:
|
|
gpu:
|
|
device-name: GV100 (Tesla V100)
|
|
vendor-name: NVIDIA
|
|
disk:
|
|
- size: ">= 512 GiB"
|
|
memory: ">= 16 GB"
|
|
|
|
/gpu_info:
|
|
summary: Display GPU info
|
|
discover+:
|
|
test: /test/tmt/no-rpm/gpu_info
|
|
|
|
/validate:
|
|
discover+:
|
|
test: /test/tmt/no-rpm/validate
|
|
|
|
/unit:
|
|
discover+:
|
|
test: /test/tmt/no-rpm/unit
|
|
|
|
/bats-nocontainer:
|
|
discover+:
|
|
test: /test/tmt/no-rpm/bats-nocontainer
|