i try bigger channel img and weight like below.
`img_shape = ( 1,512,224,224 ) , w_shape = (256,512,3,3 )`
shape format is a NCHW.
and the result is
direct => 50.641ms
winograd => 604.84ms
the performance not good than direct conv2d...
Should I use more channels?
---
[Visit
To
try bigger number of channels. Winograd is slow for small channels.
---
[Visit
Topic](https://discuss.tvm.ai/t/topi-winograd-convolution-performance-is-too-slow/6161/2)
to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, [click
her
Hello!
Currently, I am testing to compare the performance of direct conv2d and
winograd conv2d using TOPI.
However, as a result of experiments, conv2d using winograd algorithm is too
much worse than direct.
The code below is the code I experimented with.
## data shape
data_shape = (1,
Awesome, thanks a lot @trevor-m. One more quick question before I try it out -
what data type is DLTensor->data? The `codegen_c` base casts it to the type
that the argument to the function is (in my case, input is a `float*` and
input_mask is an `int*`).
---
[Visit
Topic](https://discuss
Hi @jonso, when I do relay.build with target="cuda", the data inputs supplied
to my runtime module are already placed on the GPU by the graph runtime.The
DLTensor->data will be a device pointer to the data in GPU memory and you can
pass this directly to CUDA libraries.
If you need to get the
@jonso if you can get into the `GetFunction` in external module, it means there
is no problem for runtime symbol lookup. Can you check if the input data is
correct? For example, the data you have in the external runtime should be from
here:
https://github.com/apache/incubator-tvm/blob/master/
Ah I see. One reason might be an empty host module in this case. I'd call out
@trevor-m since he has the experience to offload subgraphs to TRT while keeping
thre rest on CUDA.
---
[Visit
Topic](https://discuss.tvm.ai/t/external-codegen-with-cuda-target/6159/4) to
respond.
You are recei
Sorry about that, I think I misspoke. I already have the annotation pass set up
properly and my codegen is being called. However, when I try to print out one
of my inputs from my codegen, the program crashes.
I have a feeling that since the target is “cuda”, the data isn’t being moved
from G
No that's a different flow. TVM itself has the cuBLAS and cuDNN support already
([example](https://github.com/apache/incubator-tvm/blob/master/python/tvm/contrib/cudnn.py)).
If you set the target with `-libs`, it's using the TVM builtin one instead of
your codegen. To use your codegen, now you
Thank you @comaniac and @matt-arm. I will see if MergeCompilerRegions is
something that will work for us, otherwise will keep an eye out for your PR.
---
[Visit
Topic](https://discuss.tvm.ai/t/external-codegen-status-of-annotating-composite-functions/6150/4)
to respond.
You are receiving
Hey @zhiics and @comaniac,
I am working on an external codegen that will run on GPU. My external codegen
module is a CSourceModule. The code generated in this module will call some
CUDA APIs.
If I go through the external codegen workflow and set the target to `cuda
-libs=cublas,cudnn`, will
AnnotateTarget doesn't support composite functions at the moment. I intend to
send a PR to resolve this very soon (hopefully this week). You can use
MergeCompilerRegions if you like, but this will only be applicable if you also
support conv2d, bias and relu individually as well as merged.
Hello!
I am currently making a network using Winograd algorithm and problem is
performance is less than direct conv2d.
According to the paper [Fast Algorithms for Convolutional Neural
Networks](https://arxiv.org/pdf/1509.09308.pdf), performance should be higher
than direct implementation as
In xgb cost model, it extracts feature some kind of methods(itervar, curve,
knobs)
But, its size depends on kind of knobs and # of knobs.
So any other method, I extract the constant size of feature from any schs and
tasks
---
[Visit Topic](https://discuss.tvm.ai/t/extract-constant-size-fe
14 matches
Mail list logo
