accelergy-project / timeloop-accelergy-tutorial Goto Github PK

I am getting the below errors when I am trying to build a docker image. Any thoughts?

My command is "docker build -t tutorial_withvi .". I am trying to add VI to the image. Below are errors in text.

=> ERROR [stage-1 3/7] COPY src/ /usr/local/src/ 0.0s
=> ERROR [stage-1 4/7] COPY ./bin/refresh-exercises /usr/local/bin 0.0s
=> CACHED [stage-1 5/7] RUN apt-get install vim 0.0s
=> ERROR [stage-1 6/7] COPY docker-entrypoint.sh /usr/local/bin

Hello,

I am testing the exercises in the tutorial on Timeloop and Accelergy. All the exercises worked well, but 2020.ispass exercise is only based on a simple technology model.

So can I change it into a more accurate model? I've tried to use accelergy-aladdin-plug-in and added a new list item in accelergy_config.yaml, but there's still no use.

Thank you in advance!

Hi. First of all, thank you for your great works.
I have a question about your works.
One of the contributions of the Accelergy paper is to support energy for fine-grained actions. But I think Timeloop doesn't define fine-grained behaviors like gated_MAC or repeated/random access.
Can you tell me how the energy of fine-grained actions is reflected when I run Accelergy and Timeloop together?

Thanks

Hi

I want to locally install timeloop-accelergy (without sudo access). First, I installed all required libraries locally and then tried to install timeloop-accelergy using virtual environment. But I got an error when I run the command scon -j4 --accelergy --static.

/usr/local/bin/ld: cannot find -lyaml-cpp
/usr/local/bin/ld: cannot find -lncurses
/usr/local/bin/ld: cannot find -ltinfo
/usr/local/bin/ld: cannot find -lgpm
collect2: error: ld returned 1 exit status
scons: *** [build/timeloop-design-space] Error 1

I followed this link to edit the SConstruct file in timeloop-accelergy/accelergy-timeloop-infrastructure/src/timeloop/SConstruct and set my local directory. But I got the same error! I didn't have any problem with installing timeloop-accelergy with sudo access in standard location. Do you have any idea for this error? Thanks!

Hi, I am using docker to run the timeloop & accelergy in the diretory "/exercises/timeloop+accelergy" and when I am using the command (timeloop-mapper arch/eyeriss_like-int16.arch.yaml
arch/components/.yaml
prob/prob.yaml
mapper/mapper.yaml
constraints/.yaml), I got the following errors:

Problem configuration complete.
ERROR: key not found: arithmetic, at line: 0

Do you know what seems to be the problem?

Thanks.
Travis

Dear developers,

There is an error when I run the command scon -j4 --accelergy --static, and the error is like this:

/usr/include/boost/multiprecision/cpp_int.hpp:175:4: error: right operand of shift expression ‘(1 << 63)’ is >= than the precision of the left operand [-fpermissive]
BOOST_STATIC_CONSTANT(limb_type, sign_bit_mask = 1u << (limb_bits - 1));

Do you have any solutions for this error? Thank you!

Does Timeloop support RNNs?

Hello, developers,

I was trying to install natively timeloop-accelergy on my centos following the official instructions, but an error came up when I run "scons -j4 --accelergy --static", the error is :" g++: error: unrecognized command line option '-std=c++14' ". Could you please help me to fix this issue? Thank you very much!

Hello,

I am testing the exercises in the tutorial on Timeloop and Accelergy. All the exercises worked well. To better understand how Timeloop+Accelergy works, I am trying to modify the exercises.

To begin with, I tried to replace the arithmetic unit in the exercises:

	- name: mac      |> - name: add
	  class: intmac  |>   class: intadder

But I get the following errors:

• For the exercise of timeloop+accelergy
  what(): vector::_M_range_check: __n (which is 40577184) >= this->size() (which is 6) Aborted (core dumped)

• For the exercise of accelergy: 04_eyeriss_like
  ERROR: Cannot find corresponding entry for component eyeriss_like.PE[0].mac

For the moment I don't understand why I can't replace the mac unit by an adder.
The class: intadder is defined in the list of "primitive_component_libs.yaml" and there is no constraint to use the mac class.

Do you have any idea what I did wrong?

Thank you in advance