mli / transformers-benchmarks Goto Github PK
View Code? Open in Web Editor NEWreal Transformer TeraFLOPS on various GPUs
License: Apache License 2.0
transformers-benchmarks's People
Contributors
Stargazers
Watchers
Forkers
myfzju kingerfzb gary-code techthiyanes valleysprings forks-learning chenhaomingbob korsh84 zhenxianglin tuozhanjun skywalkerwang98 iamfaith lzy17 shawyuan97 jackie-wx barneyqiao iskybear bingwen-hu zyxcambridge zmt123sdu kong-learning mars-1990 perfectqing rockyuan comaniac grant-rk chhzh123 jirigesi chrisraynoor ouyangyu cnxqscn nhsjgczryf yangyang8599 aisensiy zjjsl1 linuer ma-chengcheng flyingdog-huang xinzhang-ops brandfucker fanghostt howard-hou geometrylearner 921011 798253405 wangpeiyi9979 uqsirius sinboyxx wpf19911118 readme2gh qyang0211 niejinliang read2me2gh qhduan eeyorelee develop-productivity wiegehts572 kuangjunwei1 ii0 crouchggj shjwudp xyang2013 hotbaby ailearnwjf t-atlas tianhonyu olokshyn huliangdream frankinwi zizhazhu mars-wei zhoutianzi666 qiuzeyu virt-lab ericmaszhu chenxichen95 sabrinalameiras lpylpy0514 fanjinglian yuliu-eric nagatoyuki0943 kennyng-19 peixuewu 2201957 tjxj whu-pzhang hey-cu sphish tims-ml qustfmy techwithray gengxin 1999ericma yeatrytransformers-benchmarks's Issues
About the theoretical value of the GPU
请问沐神：
- 在notebook中指向的wiki里，3090ti的理论值40是从表中的core boosted value（39.997）得到的吗？
- 我在自己的很多块3090上，用CUDA11.7和nvidia-driver 525跑出来的TFLOPS都只有24，距离base（29.3）和boost（35.6）的理论值都有一定的差距。请问notebook中用3090ti跑的TFLOPS是经过超频的吗？要想达到接近理论值的FLOPS需要做怎样的设置呢？
3090ti 的理论数据出自哪里？
参考techpowerup数据， 3090ti 的单精度/半精度数据与表格里展示的不太一样
Exceptionally high memory bandwidth
我在 RTX 4090 工作站上运行 banchmark 程序，取得了异常高的总线带宽数据：
Pytorch version : 1.14.0a0+44dac51
CUDA version : 12.0
GPU : NVIDIA GeForce RTX 4090
Matrix Multiplication:
n=128 n=512 n=2048 n=8192
torch.float32 1.048 29.653 82.788 86.676
torch.float16 1.304 46.890 167.112 158.596
Memory Bandwidth:
65536 262144 1048576 4194304
TFLOPS 0.025 0.099 0.324 0.484
GB/s 196.343 792.595 2590.594 3868.374可以看到显存带宽为 3868 GB/s，而我查到的 4090 理论显存带宽为 1000 GB/s 左右。
而我在 A800 服务器上运行 banchmark 程序的结果是正常的：
Pytorch version : 2.0.0a0+1767026
CUDA version : 12.1
GPU : NVIDIA A800 80GB PCIe
Matrix Multiplication:
n=128 n=512 n=2048 n=8192
torch.float32 0.464 25.947 82.386 105.973
torch.float16 0.343 31.456 192.540 215.333
Memory Bandwidth:
65536 262144 1048576 4194304
TFLOPS 0.009 0.036 0.143 0.216
GB/s 72.026 288.159 1143.973 1727.486显存带宽 1727 GB/s 低于理论上限 1935 GB/s
这导致 4090 的显存带宽远高于 A800，在我的实际训练中 4090也取得了更快的训练速度。
请问 4090 这样高的带宽是正常的吗？如果不正常的话有什么可能的原因？
AttributeError: 'NoneType' object has no attribute 'samples_per_sec'
I try to run your "transformers.ipynb" with python. But it had an error. How can I solve it?
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
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.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.