CLASSIC
Resources
Grok1 314B 25.02 (DGXC Benchmarking)

Grok1 314B 25.02 (DGXC Benchmarking)

Description
This recipe contains information and scripts to produce performance results for the Grok1 314B training workload.
Publisher
NVIDIA
Latest Version
25.02
Modified
April 17, 2025
Compressed Size
12.06 KB

Overview

This recipe contains information and scripts to produce performance results for the Grok 1 training workload. The scripts help perform environment setup and launch benchmark jobs. This variant of the workload is best-suited for GPU clusters with

  • At least 8 GPUs with at least 80 GB memory each. Training of this 314-billion parameter variant of the workload will not fit on fewer GPUs with less memory.
  • H100 or GH200 GPUs. This workload runs with FP8 and BF16 precision.
GPUs SeqLen Layers TP PP CP EP DP VP MBS GBS GA
8 4096 2 4 1 1 2 1 NA 1 1024 128
16 4096 4 4 1 1 4 1 NA 1 1024 128
32 4096 4 4 1 1 8 1 NA 1 1024 128
64 8192 8 4 1 2 8 1 NA 1 1024 128
128 8192 16 4 2 2 8 1 8 1 1024 128
256 8192 32 4 4 2 8 1 8 1 1024 128
512 8192 64 4 8 2 8 1 8 1 1024 128
1024 8192 64 4 8 2 8 2 8 1 2048 128
2048 8192 64 4 8 2 8 4 8 1 4096 128

Expected Performance

Performance for Grok 1 training is measured by seconds per iteration, or in other words seconds per training step. This metric is logged for every training step in a .out file which is generated inside of the $STAGE_PATH/results/$GSW_VERSION/$DTYPE/314b/$JOB_TOTAL_GPUS folder.

Since the performance fluctuates significantly at the beginning, we are using the last training step timing to obtain throughput value.

grep train_step_timing *.out
Epoch 0: : 100%|██████████| 50/50 [23:22<00:00, reduced_train_loss=0.186, global_step=49.00, consumed_samples=51200.0, train_step_timing in s=24.00]

To obtain throughput as a tokens per second measurement, follow this formula: 

(sequence length) * (global batch size) / (training_step_timing) = (throughput in tokens per second)

E.g. 8192 * 4096 / 24 = 1398101

To calculate time to train with 1T tokens estimate:

(total tokens) / (throughput in tokens per second) / (number of seconds in a day) = (time to train in days)

E.g. 1e12 / 1398101 / 86400 = 8.28 days

To calculate the model flops utilization (MFU). Calculation shown here.

MFU = (global batch size) * (model flops) / (training step time) / (number of GPUs) /peak GPU FLOPS)

The peak theoretical throughput for H100 FP8 is 1979 TFLOPS and for H100 BF16 is 989 TFLOPS.

The model flops for Grok 1 for GBS=1 per GPU for 2048 GPUs is 4.27E+15.

E.g. Grok 1 BF16 on 2048x H100 GPUs (GBS=4096)

peak FLOPS for H100 = 989 TFLOPS
training step time = 24
model flops = 4.27E+15

MFU = 4096 * 4.27E+15 / 24 / 2048 / 989E+12 = 36.0%
Grok 1 314b BF16 Throughput on 8x H100 GPUs Throughput on 16x H100 GPUs Throughput on 32x H100 GPUs Throughput on 64x H100 GPUs Throughput on 128x H100 GPUs Throughput on 256 H100 GPUs Throughput on 512 H100 GPUs Throughput on 1024 H100 GPUs Throughput on 2048 H100 GPUs
Training step time (seconds per step) 24.00 21.71 10.70 21.42 20.73 20.83 21.01 21.20 21.29
Throughput in tokens per second 174762 193223 391991 391606 404699 402756 399191 791378 1576065
Model flops utilization 45.4% 44.5% 45.1% 42.9% 42.8% 41.8% 41.09% 40.73% 40.56%
Time to train 1T tokens in days NA NA NA NA NA NA 28.99 14.63 7.34
Grok 1 314b FP8 Throughput on 8x H100 GPUs Throughput on 16x H100 GPUs Throughput on 32x H100 GPUs Throughput on 64x H100 GPUs Throughput on 128x H100 GPUs Throughput on 256 H100 GPUs Throughput on 512 H100 GPUs Throughput on 1024 H100 GPUs Throughput on 2048 H100 GPUs
Training step time (seconds per step) 19.71 17.80 8.70 17.53 16.70 16.75 16.90 17.01 17.30
Throughput in tokens per second 212757 235635 482159 478583 502311 500752 496367 981295 1939562
Model flops utilization 27.7% 27.1% 27.7% 26.2% 26.6% 26.0% 25.53% 25.24% 24.94%
Time to train 1T tokens in days NA NA NA NA NA NA 23.32 11.73 5.97

For proxy configs (<512 GPUs scales) we don't provide time to train estimates to avoid misleading conclusions. Proxy configs are not realistic and were created to allow fit of Grok model to smaller number of GPUs than intended.

Request Access

No special pre-requisites or access required for this recipe.

Prepare Environment

Slurm

We reference a number of Slurm commands and parameters in this document. A brief summary is included below. It's important to note these are a guide and might not be applicable to all environments. Please consult with your system administrator for the parameters that are specific to your system.

Common parameters:

  • SBATCH_PARTITION or -p - Partition (or queue) to use.
  • SBATCH_ACCOUNT or -A - Slurm account to associate with your job, different from your user. Meant for accounting purposes.
  • SBATCH_GPUS_PER_NODE or --gres=gpu:<num gpus> - If your cluster is configured with GRES this should be set to all GPUs in a node. Ignore if not configured.
    • Encountering errors such as 'GPUs not found' or 'Cannot submit to this partition without GPU resources' means this setting is required.

These parameters can be set either by exporting the environment variable or using the corresponding sbatch flag.

Workload Setup

Create a staging area by running the attached setup.sh. The script converts the docker image from nvcr.io/nvidia/nemo:24.12 to the nvidia+nemo+24.12.sqsh file under the $STAGE_PATH folder and copies NeMo Launcher code from the container.

# Set the path where all artifacts will be downloaded
export STAGE_PATH=<path to your shared file system folder> (e.g. /lustre/myproject/nemo)

# Run the setup
sbatch -A ${SBATCH_ACCOUNT} -p ${SBATCH_PARTITION} -N 1 ./setup.sh

Check the corresponding slurm-<job_id>.out file for status information.

Important: STAGE_PATH used in this step must be used when running the workload.

Dataset

Grok 1 uses synthetic data. A dataset does not need to be downloaded.

Run Training

Once the environment has been prepared, it is time to train a model. This section will demonstrate how to initiate training the model.

The training will run for the first 50 steps and will stop afterwards. Log files and results will be located under the $STAGE_PATH/results/$GSW_VERSION/$DTYPE/314b/$JOB_TOTAL_GPUS folder.

Below is a command template for launching Grok 1 model training.

DTYPE=<fp8/bf16> sbatch -A ${SBATCH_ACCOUNT} -p ${SBATCH_PARTITION} -N ${NUM_NODES} ./launch.sh

Where:

  • DTYPE is a required environment variable.
    • DTYPE can be either fp8 or bf16.
  • NUM_NODES can be calculated by N_GPUS / N_GPUS_PER_NODE, N_GPUS_PER_NODE is 8 for DGX H100, therefore for 256 GPUs scale, NUM_NODES should be 256 / 8 = 32.
  • Slurm Settings for more information on Slurm parameters.

Profiling

We have two profiling methods supported: Nsight, and NCCL Trace.

Due to overhead while profiling: the results generated with these settings is not valid for comparison. 'Performance' and 'Profiling' runs should be done separately.

Note: Profiling and NCCL Trace are currently mutually exclusive.

Run Nsight Profiling

Nsight Systems is included in our containers. To enable profiling with Nsight Systems set variable ENABLE_PROFILE=true when submitting your job.

In order to view the resulting profiles, ensure you have the latest version of Nsight Systems installed. For more information visit: Nsight Systems

Default Profiling Settings:

  • MPI Ranks: 0-8
  • Job Steps: 20-30
  • Output Location: .nsys-rep files are saved in the nsys folder within the existing results directory.
  • Filename format: ${MODEL}-${MODEL_SIZE}-${DTYPE}_${NUM_GPUS}g_${SLURM_JOB_ID}_${SLURM_NODEID}_${SLURM_LOCALID}.nsys-rep

Example command:

ENABLE_PROFILE=true DTYPE=<fp8/bf16> sbatch -A ${SBATCH_ACCOUNT} -p ${SBATCH_PARTITION} -N ${NUM_NODES} ./launch.sh

Customizing profiling behavior:

  • Specify job steps to profile:
    • RUN_CONF_PROFILE_START_STEP: start profiling on this job step. Default: 20
    • RUN_CONF_PROFILE_STOP_STEP: stop profiling on this job step. Default: 30
  • Select MPI ranks to profile:
    • RUN_CONF_PROFILE_RANKS: Comma-separated list of MPI ranks to profile. Example: "0,1,2,3" Default: "0,1,2,3,4,5,6,7"
  • Enable GPU device metrics capture:
    • RUN_CONF_PROFILE_GPU_METRICS: boolean, set to 'true' to capture device metrics.
    • Default: false
    • Note: Additional system configuration is required for GPU device metrics to work.
  • Enable CPU metrics capture:
    • RUN_CONF_PROFILE_CPU: boolean, set to 'true' to capture CPU metrics.
    • Default: false

Example customized profiling command:

ENABLE_PROFILE=true RUN_CONF_PROFILE_GPU_METRICS=true RUN_CONF_PROFILE_RANKS="0" DTYPE=<fp8/bf16> sbatch -A ${SBATCH_ACCOUNT} -p ${SBATCH_PARTITION} -N ${NUM_NODES} ./launch.sh

Troubleshooting:

If you encounter issues, try the defaults ENABLE_PROFILE=true first as these should be broadly applicable to most systems.

Viewing results

In order to view the profile traces (*.nsys-rep files) interactively:

  • Install the latest Nsight Systems client on your preferred system
  • Copy the generated .nsys-rep files to a folder on your preferred system. E.g., /home/nsight-traces/
  • Open Nsight Systems client, then click "File | Open" and select one or more .nsys-rep files from /home/nsight-systems folder. For more details, see Reading Your Report in GUI guide.
  • Once loaded you can analyze the workload behavior to learn about any performance bottlenecks associated with the job run.

Since most of the benchmarking jobs run on multiple GPUs, there will be multiple .nsys-rep files generated for each run. Multi-Report Analysis Guide will be very helpful to automate the analysis and get to results quicker by using Nsight recipes.

See these tutorials to get a quick start if you are new to Nsight profiling.

Run NCCL Trace

NCCL traces provide a breakdown of the communication pattern outlining both the type of NCCL calls being made and their message sizes.

To collect NCCL Trace information, set variable ENABLE_NCCL_TRACE=true when submitting your job.

Defaults:

  • File Size: NCCL Trace generates large files, therefore profiling is limited to the first 10 steps.
  • Output Location: Trace files are saved to a separate directory with nccl-trace appended to the version string.
  • Output Directory: $STAGE_PATH/results/$GSW_VERSION-nccl-trace/$DTYPE/${MODEL_SIZE}/$JOB_TOTAL_GPUS

Example command:

ENABLE_NCCL_TRACE=true DTYPE=<fp8/bf16> sbatch -A ${SBATCH_ACCOUNT} -p ${SBATCH_PARTITION} -N ${NUM_NODES} ./launch.sh

Notes

model flops = (sequence length) * ((attention flops) + (mlp flops) + (embedding flops))

model flops breakdown:
    attention flops = 12 * (number of layers) * (hidden size)^2 * (1 + (number of query groups)/(number of attention heads) + (sequence length)/(hidden size)/2)
    mlp flops = 18 * (number of layers) * (FFN size) * (hidden size) * (top K)
    embedding flops = 6 * (vocab size) * (hidden size)

Grok1 314b calculation:
    sequence length = 8192
    attention flops = 12 * 64 * 6144^2 * (1 + 8/48 + 8192/6144/2) = 53,150,220,288
    mlp flops = 18 * 64 * 32768 * 6144 * 2 = 463,856,467,968
    embedding flops = 6 * 128256 * 6144 = 4,728,029,184

    model flops = 8192 * (53,150,220,288 + 463,856,467,968 + 4,728,029,184) = 4.27E15