Git Product home page Git Product logo

llm-analysis's Issues

latency [BUG]

The latency i am getting here and the actual time when i am inferencing are not same. And also there is a huge difference between these two. So could be the problem?

How to get the analysis of model Qwen1.5-0.5B

@mvpatel2000 @cli99 @weimingzha0 @digger-yu @BhAem I want to get the analysis info Time to first token (s) 、Time for completion (s) and Tokens/second about the model Qwen1.5-0.5B , so do I just need to run the following command :

HF_ENDPOINT=https://hf-mirror.com
gpu_name='a100-sxm-80gb'
dtype_name="w16a16e16"
output_dir='outputs_infer'
model_name=Qwen/Qwen1.5-0.5B
batch_size_per_gpu=1
tp_size=2
output_file_suffix="-bs${batch_size_per_gpu}"
cost_per_gpu_hour=2.21
seq_len=128
num_tokens_to_generate=242
flops_efficiency=0.7
hbm_memory_efficiency=0.9
achieved_tflops=200                # will overwrite the flops_efficiency above
achieved_memory_bandwidth_GBs=1200 # will overwrite the hbm_memory_efficiency above

if [[ ! -e $output_dir ]]; then
    mkdir $output_dir
elif [[ ! -d $output_dir ]]; then
    echo "$output_dir already exists but is not a directory" 1>&2
fi

HF_ENDPOINT=$HF_ENDPOINT CUDA_VISIBLE_DEVICES=3 python -m llm_analysis.analysis infer --model_name=${model_name} --gpu_name=${gpu_name} --dtype_name=${dtype_name} -output_dir=${output_dir} --output-file-suffix=${output_file_suffix} \
    --seq_len=${seq_len} --num_tokens_to_generate=${num_tokens_to_generate} --batch_size_per_gpu=${batch_size_per_gpu} \
    --tp_size=${tp_size} \
    --cost_per_gpu_hour=${cost_per_gpu_hour} \
    --flops_efficiency=${flops_efficiency} --hbm_memory_efficiency=${hbm_memory_efficiency} --log_level DEBUG

mistral and mixtral inference[BUG]

Describe the bugMistral and Mixtral models not able to infer
When i give the name of the model as i do for other models in case of mistral there is a key error from the configuration_auto.py file in llm_analysis module. This is because there is no key with mistral in the config_map.

So could you also add all the models from hugging face which are not yet defined!!

[BUG]Is it possible that hbm_memory_efficiency is not working in the code?

Describe the bug
I tried with two unused hbm_memory_efficiency, 1 and 0.6 but ended up with the same value for (weight+op_state+grad+act)_memory_per_gpu. Is it possible that hbm_memory_efficiency is not working in the code?

To Reproduce
Steps to reproduce the behavior:

  1. python -m llm_analysis.analysis train --model_name /hdd/echozhou/llm-analysis/examples/llama --gpu_name a100-pcie-40gb --activation_recomputation 1 --tp_size 1 --pp_size 3 --sp_size 1 --dp_size 1 --gradient_accumulation_steps 4 -b 16 --seq_len 1400 --total_num_gpus 3 --total_num_tokens 1e12 --activation_recomputation 2 --flops_efficiency 1 --hbm_memory_efficiency 0.6 --output_dir /hdd/echozhou/llm-analysis/examples/llama/test
  2. python -m llm_analysis.analysis train --model_name /hdd/echozhou/llm-analysis/examples/llama --gpu_name a100-pcie-40gb --activation_recomputation 1 --tp_size 1 --pp_size 3 --sp_size 1 --dp_size 1 --gradient_accumulation_steps 4 -b 16 --seq_len 1400 --total_num_gpus 3 --total_num_tokens 1e12 --activation_recomputation 2 --flops_efficiency 1 --hbm_memory_efficiency 1 --output_dir /hdd/echozhou/llm-analysis/examples/llama/test

Expected behavior
The final memory consumption you get is all (weight+op_state+grad+act)_memory_per_gpu: 20.14 GB

Screenshots
image

image

question about the memory calculation

Hello @cli99, Thank you very much for open-sourcing your library for analyzing large language models. This is very helpful for us to understand various optimization algorithms and parallel configuration strategies. After going through the code, I have encountered a few questions.

In the paper "Reducing Activation Recomputation in Large Transformer Models", there are two LayerNorms in one transformer layer. But in the code:

        weight_memory_per_layer = (
            weight_memory_attn_per_layer
            + weight_memory_mlp_per_layer
            + weight_memory_layernorm_per_layer
        )

only one "weight_memory_layernorm_per_layer" is added.

Also in this paper, the blocks which can be parallelized using tensor parallelism are attention and mlp. But in the code below, LayerNorm can also be parallelized when tensor parallelism is applied.

        weight_memory_layernorm_per_layer = (
            self.get_num_params_per_layer_layernorm()
            * self.dtype_config.weight_bits
            / BITS_PER_BYTE
            / self.parallelism_config.tp_size
            / sharded_dp_size
        )

When I print the summary_dict in the provied python script in llama2 folder, it given me the following result:

{'batch_size_per_gpu': 1, 'seq_len': 512, 'tp_size': 2, 'ep_size': 1, 'pp_size': 1, 'num_tokens_to_generate': 32, 'flops_efficiency': 0.6, 'hbm_memory_efficiency': 0.6, 'layernorm_dtype_bytes': 2, 'use_kv_cache': True, 'kv_cache_latency': 0.00014570698054601933, 'kv_cache_memory_per_gpu': 89128960.0, 'weight_memory_per_gpu': 55292731392.0, 'weight_memory_embedding_per_gpu': 262144000.0, 'prefill_activation_memory_per_gpu': 16777216.0, 'prefill_max_batch_size_per_gpu': 1824, 'prefill_num_flops_fwd_total': 57305601146880.0, 'decode_activation_memory_per_gpu': 32768.0, 'decode_max_batch_size_per_gpu': 343, 'decode_num_flops_fwd_total': 110585446400.0, 'prefill_latency': 0.15908735621271564, 'prefill_latency_fwd_attn': 0.03487366607863248, 'prefill_latency_fwd_mlp': 0.11746919100170941, 'prefill_latency_fwd_layernorm': 0.001097087853522969, 'prefill_latency_fwd_tp_comm': 0.004473924266666667, 'prefill_latency_fwd_sharded_dp_comm': 0.0, 'prefill_latency_fwd_input_embedding': 0.00045651196944907646, 'prefill_latency_fwd_output_embedding_loss': 0.0007169750427350427, 'decode_latency': 0.04685015182136376, 'decode_latency_fwd_attn': 0.009875397743992155, 'decode_latency_fwd_mlp': 0.03510895406244892, 'decode_latency_fwd_layernorm': 2.1427497139120487e-06, 'decode_latency_fwd_tp_comm': 0.0012799999999999999, 'decode_latency_fwd_sharded_dp_comm': 0.0, 'decode_latency_fwd_input_embedding': 0.00043654994278240976, 'decode_latency_fwd_output_embedding_loss': 1.4003418803418803e-06, 'total_decode_latency': 1.4992048582836404, 'total_latency': 1.658292214496356, 'total_per_token_latency': 0.05182163170301113, 'prefill_tokens_per_sec': 3218.3575878613824, 'decode_tokens_per_sec': 21.344648013370964, 'total_tokens_per_sec': 19.296960885581193, 'prefill_cost_per_1k_tokens': 0.00038149203258474565, 'decode_cost_per_1k_tokens': 0.057521575291785504, 'total_cost_per_1k_tokens': 0.06362544781314144}

weight_memory_per_gpu is 55292731392.0, which is less 70B * 2 / 2 = 70B

So can you provide more information about this? Look forward to your response, thank you once again.

[REQUEST] How to get other GPU config

Is your feature request related to a problem? Please describe.
I recently wanted to test on T4, but I don't know how to measure intra_node information.

Describe the solution you'd like
The following is the T4 information I checked, including intra_node_bandwidth_in_GB_per_sec intra_node_min_message_latency inter_node_bandwidth_in_GB_per_sec. I don’t know how to obtain it.

{
    "name": "T4-pcie-16gb",
    "mem_per_GPU_in_GB": 16,
    "hbm_bandwidth_in_GB_per_sec": 320,
    "intra_node_bandwidth_in_GB_per_sec": XXX,
    "intra_node_min_message_latency": XXX,
    "peak_fp16_TFLOPS": 65,
    "peak_i8_TFLOPS": 130,
    "peak_i4_TFLOPS": 260,
    "inter_node_bandwidth_in_GB_per_sec": XXX
}

A question about layernorm activation memory.

Hi,

The function get_memory_activation_per_layer_layernorm() will return a value of seq_len * batch_size * hidden_dim / sp_size * dtype_bytes, which in fp16 will be 2sbh/s.

However, I find the paper Reducing Activation Recomputation in Large Transformer Models mentions that the activation memory of LayerNorm is 4sbh.

Unfortunately, I'm not so familiar with LLM memory consumption. Since other activation memory result fits the paper, I wonder if there exists a mistake inside the paper or there is a bug in this function?

Thanks,
Esar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    πŸ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. πŸ“ŠπŸ“ˆπŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❀️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.