Disable FP8 in Mcore integration test on older GPUs (#1357)

Debug Mcore integration test Avoid FP8 on Ampere and older. Generate synthetic data instead of depending on external data. Signed-off-by: Tim Moon <[email protected]>
NVIDIA · Dec 6, 2024 · d8b13cb · d8b13cb
1 parent d978e80
commit d8b13cb
Show file tree

Hide file tree

Showing 3 changed files with 21 additions and 4 deletions.
diff --git a/qa/L1_pytorch_mcore_integration/.gitignore b/qa/L1_pytorch_mcore_integration/.gitignore
@@ -0,0 +1,2 @@
+Megatron-LM
+vocab.json
diff --git a/qa/L1_pytorch_mcore_integration/merges.txt b/qa/L1_pytorch_mcore_integration/merges.txt
@@ -0,0 +1 @@
+#version: 0.2
diff --git a/qa/L1_pytorch_mcore_integration/test.sh b/qa/L1_pytorch_mcore_integration/test.sh
@@ -8,13 +8,27 @@ set -e
 : ${TE_PATH:=/opt/transformerengine}
 : ${MCORE_PATH:=${TE_PATH}/qa/L1_pytorch_mcore_integration/Megatron-LM}
 
+# Check whether FP8 is supported
+DEVICE_ARCH=$(nvidia-smi --query-gpu=compute_cap --format=csv,noheader | head -n 1 | sed 's/[^0-9]//g')
+if [[ ${DEVICE_ARCH} -ge 89 ]]; then
+    WITH_FP8=1
+fi
+
 # Download Megatron-LM if needed
 if [ ! -d "${MCORE_PATH}" ]; then
     pushd $(dirname ${MCORE_PATH})
     git clone -b core_r0.9.0 https://github.com/NVIDIA/Megatron-LM.git Megatron-LM
     popd
 fi
 
+# Create mock vocab
+VOCAB_FILE=${TE_PATH}/qa/L1_pytorch_mcore_integration/vocab.json
+printf "" > ${VOCAB_FILE}
+printf "{" >> ${VOCAB_FILE}
+printf "\"<|endoftext|>\": 0" >> ${VOCAB_FILE}
+seq 1 4095 | awk '{ printf(", \"%d\": %d", $1, $1) }' >> ${VOCAB_FILE}
+printf "}" >> ${VOCAB_FILE}
+
 # Megatron-LM invocation
 COMMAND="
 NVTE_TORCH_COMPILE=0
@@ -40,17 +54,17 @@ ${MCORE_PATH}/pretrain_gpt.py
 --hidden-size 128
 --num-attention-heads 8
 --seq-length 128
---max-position-embeddings 2048
+--max-position-embeddings 128
 --micro-batch-size 1
 --global-batch-size 8
 --train-iters 10
 --eval-iters 10
 --lr 1e-4
 --mock-data
---vocab-file /data/gpt3/pile-cc1-cc2-shuf/bpe/gpt2-vocab.json
---merge-file /data/gpt3/pile-cc1-cc2-shuf/bpe/gpt2-merges.txt
+--vocab-file ${VOCAB_FILE}
+--merge-file ${TE_PATH}/qa/L1_pytorch_mcore_integration/merges.txt
 --transformer-impl transformer_engine
---fp8-format hybrid
+${WITH_FP8:+--fp8-format hybrid}
 "
 COMMAND=$(echo "${COMMAND}" | tr '\n' ' ')