Vignesh Kothapalli

fix the optimizer iteration count when gradient accumulation is enabled

6

## Checklist before submitting - [x] Did you read the [contributor guide](https://github.com/horovod/horovod/blob/master/CONTRIBUTING.md)? - [x] Did you update the docs? NA - [x] Did you write any tests to validate this...

feat: support uint8 and int8 allreduce in pytorch

11

## Checklist before submitting - [x] Did you read the [contributor guide](https://github.com/horovod/horovod/blob/master/CONTRIBUTING.md)? - [x] Did you update the docs? - [x] Did you write any tests to validate this change?...

**Is your feature request related to a problem? Please describe.** Currently, allreduce with torch is registered to support allreduce on `IntTensor, LongTensor, HalfTensor, FloatTensor, DoubleTensor` types but not on `ByteTensor,...

CUPTI symbols are undefined after libkineto build

1

Hi, I have followed the documentation [here](https://github.com/pytorch/kineto/tree/main/libkineto#building-libkineto) and built libkineto. However, the sample programs [build.sh](https://github.com/pytorch/kineto/blob/main/libkineto/sample_programs/build.sh) throws an error as follows: ``` /usr/bin/ld: /usr/local/lib/libkineto.a(init.cpp.o): in function `libkineto_init': init.cpp:(.text+0x3f2a): undefined reference to...