GitLab

* Fixed minor bug with DistributedConfigsHelper prematurely crashing the test for insufficient GPUs before @pytest.skip condition.
Signed-off-by: Alp Dener <adener@nvidia.com>

* Update tests/jax/distributed_configs_helper.py
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

* Debug PyTest errors when running on single-GPU system
Signed-off-by: Tim Moon <tmoon@nvidia.com>

---------
Signed-off-by: Alp Dener <adener@nvidia.com>
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Co-authored-by: Tim Moon <tmoon@nvidia.com>

[JAX] Regression tests for custom ops sharding with both xmap and custom_partitioning.

Coverage:
- layernorm: fwd/grad, zero_centered_gamma, DP, TP_COL, DP_TP_COL
- rmsnorm: fwd/grad, DP, TP_COL, DP_TP_COL
- softmax: fwd/grad, SCALED, SCALED_MASKED, SCALED_UPPER_TRIANG_MASKED, DP, TP_COL, TP_ROW, DP_TP_COL, DP_TP_ROW
- self_fused_attn: fwd/grad, NO_BIAS, PRE_SCALE_BIAS, POST_SCALE_BIAS, NO_MASK, CAUSAL_MASK, PADDING_MASK, DP, TP_COL, DP_TP_COL
- cross_fused_attn: fwd/grad, NO_BIAS, NO_MASK, PADDING_MASK, DP, TP_COL, DP_TP_COL
Signed-off-by: Alp Dener <adener@nvidia.com>