# Titan V white paper and int8 tops

**URL:** <https://forums.developer.nvidia.com/t/titan-v-white-paper-and-int8-tops/72776>\
**Category:** CUDA Programming and Performance\
**Created:** [April 9, 2019, 9:35pm UTC](https://forums.developer.nvidia.com/t/titan-v-white-paper-and-int8-tops/72776 "2019-04-09T21:35:07Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![haoran.wang0gqpx](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@haoran.wang0gqpx](https://forums.developer.nvidia.com/u/haoran.wang0gqpx)\
**Post date:** [April 9, 2019, 9:35pm UTC](https://forums.developer.nvidia.com/t/titan-v-white-paper-and-int8-tops/72776/1 "2019-04-09T21:35:07Z")

</div>

Hi,

What’s the int8 TOPS of titanV? Is it 55 TOPS?  
Do you know where I can download the white paper of Titan V?

Thanks very much

---

<div class="post-metadata">

**Author:** ![Robert\_Crovella](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/robert_crovella/32/14043_2.png) [@Robert\_Crovella](https://forums.developer.nvidia.com/u/Robert_Crovella)\
**Post date:** [April 9, 2019, 9:37pm UTC](https://forums.developer.nvidia.com/t/titan-v-white-paper-and-int8-tops/72776/2 "2019-04-09T21:37:38Z")

</div>

I’m not aware of a TitanV whitepaper published by NVIDIA.

The peak theoretical int8 TOPs of TitanV should be four times the peak theoretical FP32 rate. This is true for most GPUs that have int8 support, excepting TensorCore usage on compute capability 7.5 (Turing).

---

<div class="post-metadata">

**Author:** ![njuffa](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@njuffa](https://forums.developer.nvidia.com/u/njuffa)\
**Post date:** [April 9, 2019, 9:49pm UTC](https://forums.developer.nvidia.com/t/titan-v-white-paper-and-int8-tops/72776/3 "2019-04-09T21:49:32Z")

</div>

Anandtech states “(INT8) 55.2 TOPS” for the Titan V:

[https://www.anandtech.com/show/12673/titan-v-deep-learning-deep-dive](https://www.anandtech.com/show/12673/titan-v-deep-learning-deep-dive)

I cannot find that number confirmed anywhere on NVIDIA’s website. I don’t see a Titan V whitepaper on the NVIDIA site either, just a “NVIDIA TESLA V100 GPU ARCHITECTURE” whitepaper, which says:

[https://images.nvidia.com/content/volta-architecture/pdf/volta-architecture-whitepaper.pdf](https://images.nvidia.com/content/volta-architecture/pdf/volta-architecture-whitepaper.pdf)

> [@](#):
>
> GPUs such as the NVIDIA Tesla P40 that support INT8 computation deliver nearly 48 INT8 TOPS of performance

So at least the number is plausible.