accelergy-project / accelergy-timeloop-infrastructure Goto Github PK
View Code? Open in Web Editor NEWLinux docker for the DNN accelerator exploration infrastructure composed of Accelergy and Timeloop
License: MIT License
Linux docker for the DNN accelerator exploration infrastructure composed of Accelergy and Timeloop
License: MIT License
Hi,
First of all, thanks for your great workshop (Sparse Tensor @ISCA2021).
I encountered the following error in compiling the Timeloop @ 2d4a244.
src/loop-analysis/nest-analysis-tile-info.cpp:50:15: error: ‘accumulate’ is not a member of ‘std’
50 | return std::accumulate(accesses.begin(), accesses.end(), static_cast<std::uint64_t>(0));
| ^~~~~~~~~~
scons: *** [build/loop-analysis/nest-analysis-tile-info.o] Error 1
scons: building terminated because of errors.
And the solution is adding #include <numeric>
to nest-analysis-tile-info.cpp
I compiled Timeloop in the following distribution:
Ubuntu 20.04 LTS (focal)
Hi,
Just wanted to check in if there's a way for us to load the weight of pre-trained models onto one of the yaml files and use them to perform the necessary analysis on the architecture and mapping?
If yes, could you please redirect me to the relevant documents?
Thanks!
There are two minor typos in the git command of README.md file:
The first one is in:
% git clone --recurse-submodules https://github.com/Accelergy-Project/accelergy-timeloop-infrastucture.git
and the second is in:
% git clone --recurse-submodules https://github.com/Accelergy-Project/accelergy-timeloop-infrastucture.git
For both of them the correct git command is:
% git clone --recurse-submodules https://github.com/Accelergy-Project/accelergy-timeloop-infrastructure.git
Hi,
I am following the native install steps from this link: https://timeloop.csail.mit.edu/v4/installation#to-do-need-an-install-script-for-pytimeloop . But when I do "make install_timeloop" I am getting below error:
/usr/bin/ld: cannot find -lboost_log_setup: No such file or directory
/usr/bin/ld: cannot find -lboost_filesystem: No such file or directory
/usr/bin/ld: cannot find -lboost_log: No such file or directory
/usr/bin/ld: cannot find -lboost_thread: No such file or directory
collect2: error: ld returned 1 exit status
scons: *** [build/timeloop-compound-config-test] Error 1
scons: building terminated because of errors.
make: *** [Makefile:121: install_timeloop] Error 2
Thanks!
Hello -
I am trying to build timeloop from source using these instructions - https://accelergy.mit.edu/infra_instructions.html
I am able to get through every instruction until this one:
scons -j4 --accelergy --static
This throws an error that reads:
In file included from src/isl-wrapper/ctx-manager.cpp:1:
include/isl-wrapper/ctx-manager.hpp:4:10: fatal error: isl/cpp.h: No such file or directory
4 | #include <isl/cpp.h>
| ^~~~~~~~~~~
compilation terminated.
scons: *** [build/isl-wrapper/ctx-manager.o] Error 1
In file included from include/loop-analysis/isl-analysis/isl-nest-analysis.hpp:3,
from src/loop-analysis/isl-analysis/isl-nest-analysis.cpp:1:
include/loop-analysis/isl-ir.hpp:37:10: fatal error: isl/cpp.h: No such file or directory
37 | #include <isl/cpp.h>
| ^~~~~~~~~~~
compilation terminated.
scons: *** [build/loop-analysis/isl-analysis/isl-nest-analysis.o] Error 1
In file included from include/loop-analysis/isl-analysis/isl-nest-analysis.hpp:3,
from include/loop-analysis/isl-analysis/isl-to-legacy-adaptor.hpp:5,
from src/loop-analysis/isl-analysis/isl-to-legacy-adaptor.cpp:1:
include/loop-analysis/isl-ir.hpp:37:10: fatal error: isl/cpp.h: No such file or directory
37 | #include <isl/cpp.h>
| ^~~~~~~~~~~
compilation terminated.
scons: *** [build/loop-analysis/isl-analysis/isl-to-legacy-adaptor.o] Error 1
In file included from src/isl-wrapper/isl-functions.cpp:6:
include/isl-wrapper/ctx-manager.hpp:4:10: fatal error: isl/cpp.h: No such file or directory
4 | #include <isl/cpp.h>
| ^~~~~~~~~~~
compilation terminated.
I have the ISL library installed. I do not see a cpp.h header in the ISL header directory.
(These are all the headers I see in the ISL include directory (apologies for pasting the entire long list):
aff.h
aff_type.h
arg.h
ast.h
ast_build.h
ast_type.h
band.h
constraint.h
ctx.h
deprecated
flow.h
hash.h
hmap.h
hmap_templ.c
id.h
id_to_ast_expr.h
id_to_id.h
id_to_pw_aff.h
ilp.h
list.h
local_space.h
lp.h
map.h
map_to_basic_set.h
map_type.h
mat.h
maybe.h
maybe_ast_expr.h
maybe_basic_set.h
maybe_id.h
maybe_pw_aff.h
maybe_templ.h
multi.h
obj.h
options.h
point.h
polynomial.h
polynomial_type.h
printer.h
printer_type.h
schedule.h
schedule_node.h
schedule_type.h
set.h
set_type.h
space.h
stdint.h
stream.h
union_map.h
union_map_type.h
union_set.h
union_set_type.h
val.h
val_gmp.h
vec.h
version.h
vertices.h
The deprecated/ directory contains:
aff_int.h
ast_int.h
constraint_int.h
ilp_int.h
int.h
map_int.h
mat_int.h
point_int.h
polynomial_int.h
set_int.h
union_map_int.h
val_int.h
vec_int.h)
What am I missing here? Should I be using an older commit of the isl library?
Thanks,
Priyanka.
There was an error in compiling timeloop that has been recently fixed:
Hi
I am following the native install steps from this link: https://timeloop.csail.mit.edu/v4/installation#to-do-need-an-install-script-for-pytimeloop . But when I do "make install_timeloop", I get the below error:
Successfully built accelergy accelergy-adc-plug-in accelergy-aladdin-plug-in accelergy-cacti-plug-in library accelergy-neurosim-plug-in accelergy-table-based-plug-ins
Installing collected packages: accelergy, accelergy-adc-plug-in, accelergy-aladdin-plug-in, accelergy-cacti-plug-in, library, accelergy-neurosim-plug-in, accelergy-table-based-plug-ins
Attempting uninstall: accelergy
Found existing installation: accelergy 0.4
Uninstalling accelergy-0.4:
Successfully uninstalled accelergy-0.4
Attempting uninstall: accelergy-adc-plug-in
Found existing installation: accelergy-adc-plug-in 0.1
Uninstalling accelergy-adc-plug-in-0.1:
Successfully uninstalled accelergy-adc-plug-in-0.1
Attempting uninstall: accelergy-aladdin-plug-in
Found existing installation: accelergy-aladdin-plug-in 0.1
Uninstalling accelergy-aladdin-plug-in-0.1:
Successfully uninstalled accelergy-aladdin-plug-in-0.1
Attempting uninstall: accelergy-cacti-plug-in
Found existing installation: accelergy-cacti-plug-in 0.1
Uninstalling accelergy-cacti-plug-in-0.1:
Successfully uninstalled accelergy-cacti-plug-in-0.1
Rolling back uninstall of accelergy-cacti-plug-in
Moving to /usr/local/lib/python3.8/dist-packages/accelergy_cacti_plug_in-0.1.dist-info/
from /usr/local/lib/python3.8/dist-packages/~ccelergy_cacti_plug_in-0.1.dist-info
Moving to /usr/local/share/accelergy/estimation_plug_ins/accelergy-cacti-plug-in/__pycache__/
from /usr/local/share/accelergy/estimation_plug_ins/accelergy-cacti-plug-in/~_pycache__
Moving to /usr/local/share/accelergy/estimation_plug_ins/accelergy-cacti-plug-in/cacti.estimator.yaml
from /tmp/pip-uninstall-tbbosost/cacti.estimator.yaml
Moving to /usr/local/share/accelergy/estimation_plug_ins/accelergy-cacti-plug-in/cacti_wrapper.py
from /tmp/pip-uninstall-tbbosost/cacti_wrapper.py
Moving to /usr/local/share/accelergy/estimation_plug_ins/accelergy-cacti-plug-in/default_SRAM.cfg
from /tmp/pip-uninstall-tbbosost/default_SRAM.cfg
Moving to /usr/local/share/accelergy/estimation_plug_ins/accelergy-cacti-plug-in/default_cache.cfg
from /tmp/pip-uninstall-tbbosost/default_cache.cfg
ERROR: Could not install packages due to an EnvironmentError: [Errno 21] Is a directory: '/usr/local/share/accelergy/estimation_plug_ins/accelergy-cacti-plug-in/cacti'
make: *** [Makefile:113: install_accelergy] Error 1
thanks!
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.