You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@tvm.apache.org by GitBox <gi...@apache.org> on 2022/04/26 03:00:47 UTC

[GitHub] [tvm] jinfagang opened a new issue, #11122: [Bug] tvm fp16 much more slower than float32

jinfagang opened a new issue, #11122:
URL: https://github.com/apache/tvm/issues/11122

   I have a simple MLP model which have 3 layer fcs, I using te to build it manually and specified the dtype in each layer. the exported lib inference much more slower than float32.....
   
   and the exported .so size is twice bigger than float32. How could this be possible?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [tvm] masahi commented on issue #11122: [Bug] tvm fp16 much more slower than float32

Posted by GitBox <gi...@apache.org>.
masahi commented on issue #11122:
URL: https://github.com/apache/tvm/issues/11122#issuecomment-1109381441

   You need to tell TVM that your target has native fp16, otherwise we can't generate fp16 code. See my forum post answer.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [tvm] masahi closed issue #11122: [Bug] tvm fp16 much more slower than float32

Posted by GitBox <gi...@apache.org>.
masahi closed issue #11122: [Bug] tvm fp16 much more slower than float32
URL: https://github.com/apache/tvm/issues/11122


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [tvm] jinfagang commented on issue #11122: [Bug] tvm fp16 much more slower than float32

Posted by GitBox <gi...@apache.org>.
jinfagang commented on issue #11122:
URL: https://github.com/apache/tvm/issues/11122#issuecomment-1109399626

   @masahi thanks, I have gened the lib with llvm config:
   
   ```
   llvm arm64 -mtriple=iphoneos-apple-darwin
   ```
   
   And all te build operations using dtype = fp16, the inference memory actually shrined down, but the .a size and speed are not expected. So it actually gened fp16 .so, but somehow the .a is bigger which I can not figure out, it doesn't make sense.
   
   And the speed slower.
   
   Does there any examples on how to exactly gen fp16 .so or .a on iPhone 12? 


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [tvm] jinfagang commented on issue #11122: [Bug] tvm fp16 much more slower than float32

Posted by GitBox <gi...@apache.org>.
jinfagang commented on issue #11122:
URL: https://github.com/apache/tvm/issues/11122#issuecomment-1109377663

   @masahi thanks. I am using iPhone12 realphone, which obisouly support fp16. But the generated .a size, bigger than float32, speed slower.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [tvm] masahi commented on issue #11122: [Bug] tvm fp16 much more slower than float32

Posted by GitBox <gi...@apache.org>.
masahi commented on issue #11122:
URL: https://github.com/apache/tvm/issues/11122#issuecomment-1109313992

   Without HW support for fp16, fp16 compute will be emulated via software. So it is not surprising that the generated code is bigger than fp32.
   
   Please post basic usage questions to the forum. 


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org