# Token per second calculations
**URL:**
**Author:**  [@almersawi](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Falmersawi)%5C
**Post date:** [November 10, 2023, 12:31pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ftoken-per-second-calculations%2F61801%2F1 "2023-11-10T12:31:54Z")
Hi! I’m trying to calculate the number of token per second that I expect to get from “llama 7b” model deployed on A10G (31.52 TFLOPS for FP16).
I know that the number of tokens = (TFLOPS / (2 \* number of model parameters))
When I do the calculations I found that
no\_of\_tokens = (31.52 \* 10e12) / (2 \* 7 \* 10e9) = 2251.4285714285716 tokens / second
but what I get from the model is approximately 32 token / second.
Am I missing something?
---
**Author:**  [@Diogo](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FDiogo)%5C
**Post date:** [November 8, 2024, 9:27pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ftoken-per-second-calculations%2F61801%2F2 "2024-11-08T21:27:44Z")
Dude I am at the same spot… Did you found out what were you missing?
---
**Author:**  [@wenxi](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fwenxi)%5C
**Post date:** [April 20, 2025, 5:57am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ftoken-per-second-calculations%2F61801%2F3 "2025-04-20T05:57:33Z")
It could be due to Memory Bandwidth Limitations.
A simple operation for the 7B model might require loading all 14GB of parameters. At 600 GB/s memory bandwidth, just loading all parameters once would take: 14GB ÷ 600GB/s = 0.023 seconds. This translates to a theoretical maximum of ~43 operations per second involving all parameters